CDISC SDTM Implementation Guide (Version 3.1.

1)

Study Data Tabulation Model Implementation Guide:

Human Clinical Trials Prepared by the CDISC Submission Data Standards Team

Notes to Readers
This is the implementation guide for Version 1.1 of the CDISC Study Data Tabulation Model, posted for comment. This Implementation Guide comprises version 3.1.1 of the CDISC Submission Data Standards and Domain Models. See CDISC notes and assumptions regarding use of --OCCUR variable in CM, SU, AE, MH domains in Section 6.

Revision History
Date 2004-07-14 2005-08-26 Version 3.1 3.1.1 Final Summary of Changes
Released version reflecting all changes and corrections identified during comment periods. Released version reflecting all changes and corrections identified during comment period.

Note: Please see Appendix 10.7 for Representations and Warranties; Limitations of Liability, and Disclaimers.

CDISC, 2005. All rights reserved FINAL

Page 1 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

CONTENTS
1 1.1 1.2 1.3 1.4 1.5 2 INTRODUCTION ..................................................................................................................................................5 PURPOSE.....................................................................................................................................................5 ORGANIZATION OF THIS DOCUMENT...........................................................................................................6 RELATIONSHIP TO PRIOR CDISC DOCUMENTS .......................................................................................6 HOW TO READ THIS IMPLEMENTATION GUIDE ...........................................................................................8 SUBMITTING COMMENTS .......................................................................................................................8

FUNDAMENTALS OF THE SDTM........................................................................................................................9 2.1 OBSERVATIONS AND VARIABLES ...........................................................................................................9 2.2 DATASETS AND DOMAINS .........................................................................................................................10 2.3 SPECIAL-PURPOSE DOMAINS ........................................................................................................... 11 2.4 THE GENERAL DOMAIN CLASSES .......................................................................................................... 11 2.5 THE CDISC STANDARD DOMAIN MODELS...............................................................................................12 2.6 CREATING A NEW DOMAIN .......................................................................................................................13 2.6.1 How to Include Variables in New Domains (Steps 4 through 7).....................................................15 SUBMITTING DATA IN STANDARD FORMAT .....................................................................................................16 3.1 STANDARD METADATA FOR DATASET CONTENTS AND ATTRIBUTES..........................................................16 3.2 USING THE CDISC DOMAIN MODELS IN REGULATORY SUBMISSIONS ......................................................17 3.2.1 CDISC Submission Dataset Definition Metadata ...........................................................................17 3.2.2 CDISC Submission Value-Level Metadata .....................................................................................19 3.2.3 Conformance...................................................................................................................................20 ASSUMPTIONS FOR DOMAIN MODELS ............................................................................................................21 4.1 GENERAL ASSUMPTIONS FOR ALL DOMAINS ............................................................................................21 4.1.1 General Dataset Assumptions .........................................................................................................21 4.1.2 General Variable Assumptions ........................................................................................................22 4.1.3 Coding and Controlled Terminology Assumptions .........................................................................23 4.1.4 Actual and Relative Time Assumptions ..........................................................................................25 4.1.5 Other Assumptions..........................................................................................................................31 MODELS FOR SPECIAL PURPOSE DOMAINS ....................................................................................................35 5.1.1 Demographics Domain Model DM............................................................................................35 5.1.1.1 Assumptions for Demographics (DM) domain model ....................................................................37 5.1.2 Comments Domain Model CO...................................................................................................38 5.1.2.1 Assumptions for Comments (CO) domain model...........................................................................39 DOMAIN MODELS BASED ON THE GENERAL CLASSES ...................................................................................40 6.1 INTERVENTIONS ........................................................................................................................................40 6.1.1 Concomitant Medications CM ...................................................................................................40 6.1.1.1 Assumptions for Concomitant Medications (CM) domain model ..................................................43 6.1.2 Exposure EX ..............................................................................................................................45 6.1.2.1 Assumptions for Exposure (EX) domain model .............................................................................47 6.1.3 Substance Use - SU.........................................................................................................................49 6.1.3.1 Assumptions for Substance Use (SU) domain model .....................................................................52 6.2 EVENTS ....................................................................................................................................................53 6.2.1 Adverse Events AE ....................................................................................................................53 6.2.1.1 Assumptions for Adverse Events (AE) domain model....................................................................56 6.2.2 Disposition DS ...........................................................................................................................59 6.2.2.1 Assumptions for Disposition (DS) domain model ..........................................................................60 6.2.3 Medical History MH ..................................................................................................................62 6.2.3.1 Assumptions for Medical History (MH) domain model .................................................................64

Page 2 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

6.3 FINDINGS ..................................................................................................................................................66 6.3.1 ECG Test Results EG .................................................................................................................66 6.3.1.1 Assumptions for ECG (EG) domain model ....................................................................................69 6.3.2 Inclusion/Exclusion Exceptions IE ............................................................................................70 6.3.2.1 Assumptions for Inclusion/Exclusion Exceptions (IE) domain model ...........................................71 6.3.3 Laboratory Test Results LB........................................................................................................72 6.3.3.1 Assumptions for Laboratory Test Results (LB) domain model.......................................................75 6.3.4 Physical Examinations PE..........................................................................................................76 6.3.4.1 Assumptions for Physical Examinations (PE) domain model.........................................................78 6.3.5 Questionnaires QS......................................................................................................................79 6.3.5.1 Assumptions for Questionnaire (QS) domain model ......................................................................81 6.3.6 Subject Characteristics SC .........................................................................................................82 6.3.6.1 Assumptions for Subject Characteristics (SC) domain model ........................................................83 6.3.7 Vital Signs VS............................................................................................................................84 7 TRIAL DESIGN DATASETS ................................................................................................................................87 7.1 INTRODUCTION ...................................................................................................................................87 7.2 PLANNED ELEMENTS, ARMS, AND VISITS ..........................................................................................88 7.2.1 Trial Elements .................................................................................................................................88 7.2.2 Trial Arms .......................................................................................................................................89 7.2.3 Trial Visits.......................................................................................................................................90 7.3 SUBJECT ELEMENTS AND VISITS ......................................................................................................91 7.3.1 Subject Elements.............................................................................................................................91 7.3.2 Subject Visits...................................................................................................................................92 7.4 TRIAL ARMS .............................................................................................................................................93 7.4.1 Identifying Trial Arms.....................................................................................................................93 7.4.2 Developing the Trial Arms Table ....................................................................................................94 7.4.3 Distinguishing between Branches and Transitions........................................................................101 7.4.4 Trial Epoch Concept .....................................................................................................................101 7.4.5 Rules concept ................................................................................................................................106 7.4.6 Recap of Trial Arms Variables ......................................................................................................106 7.4.7 Truncated Arms.............................................................................................................................107 7.5 TRIAL ELEMENTS ...................................................................................................................................107 7.5.1 Identifying Trial Elements ............................................................................................................109 7.5.2 Developing the Trial Elements Table ............................................................................................ 110 7.5.3 Recap of Trial Elements Variables ................................................................................................ 112 7.5.4 Distinguishing Elements from Epochs.......................................................................................... 112 7.6 TRIAL VISITS .......................................................................................................................................... 113 7.6.1 Identifying Trial Visits .................................................................................................................. 113 7.6.2 Developing the Trial Visits Table.................................................................................................. 113 7.6.3 Recap of Trial Visits Variables ...................................................................................................... 114 7.7 SUBJECT ELEMENTS ............................................................................................................................... 115 7.7.1 Identifying Subject Elements ........................................................................................................ 115 7.7.2 Unplanned Elements ..................................................................................................................... 115 7.7.3 Deriving SE End Date/Times ........................................................................................................ 115 7.7.4 Recap of Subject Elements Variables............................................................................................ 116 7.7.5 Using Subject Elements Data to Place Subject Data within an Element or Epoch ....................... 116 7.8 SUBJECT VISITS ...................................................................................................................................... 116 7.8.1 Identifying Subject Visits.............................................................................................................. 116 7.8.2 Recap of Subject Visits Variables.................................................................................................. 117 7.9 TRIAL INCLUSION/EXCLUSION CRITERIA................................................................................................ 117 7.10 TRIAL SUMMARY INFORMATION ......................................................................................................... 118 7.11 HOW TO MODEL THE DESIGN OF A CLINICAL TRIAL ................................................................................. 119 REPRESENTING DATA RELATIONSHIPS ....................................................................................................................120 8.1 RELATING GROUPS OF RECORDS WITHIN A DOMAIN .....................................................................121 8.1.1 --GRPID Example.........................................................................................................................121
Page 3 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

8.2 8.2.1 8.2.2 8.3 8.3.1 8.4 8.4.1 8.4.2 8.4.3 8.5 8.5.1 8.6 9

RELATING PEER RECORDS IN SEPARATE DATASETS ................................................................122 RELREC Dataset ..........................................................................................................................123 RELREC Dataset Examples..........................................................................................................123 RELATING DATASETS .......................................................................................................................124 RELREC Dataset Relationship Example ......................................................................................124 RELATING NON-STANDARD VARIABLE VALUES TO A PARENT DOMAIN .............................125 SUPPQUAL Dataset .....................................................................................................................126 Submitting Supplemental Qualifiers in Separate Datasets............................................................127 SUPPQUAL Examples .................................................................................................................127 RELATING COMMENTS TO A PARENT DOMAIN ..........................................................................128 COMMENTS Example.................................................................................................................128 HOW TO DETERMINE WHERE DATA BELONG IN THE SDTM...................................................................129

IMPLEMENTATION EXAMPLES ......................................................................................................................131 9.1 DEMOGRAPHICS EXAMPLE ............................................................................................................131 9.2 INTERVENTIONS EXAMPLES..........................................................................................................131 9.2.1 CM Example: Intermittent Use of Concomitant Medications......................................................131 9.2.2 EX Examples: ...............................................................................................................................133 9.2.3 SU Example: .................................................................................................................................135 9.3 EVENTS EXAMPLES ..........................................................................................................................139 9.3.1 AE Example ..................................................................................................................................139 9.3.2 DS Examples.................................................................................................................................140 9.3.3 MH Example: ............................................................................................................................................143 9.4 FINDINGS EXAMPLES.......................................................................................................................144 9.4.1 EG Examples ................................................................................................................................144 9.4.2 IE Example....................................................................................................................................147 9.4.3 LB Examples.................................................................................................................................148 9.4.4 PE Example...................................................................................................................................151 9.4.5 QS Examples.................................................................................................................................153 9.4.6 SC Example ..................................................................................................................................155 9.4.7 VS Example ..................................................................................................................................156 9.5 TRIAL DESIGN EXAMPLES ...............................................................................................................158 9.5.1 Defining Epochs - a How-to Example Using the Example Crossover Study................................158 9.5.2 Trial Summary Example ...............................................................................................................162 APPENDICES ...................................................................................................................................................163 10.1 CDISC SDS TEAM .................................................................................................................................163 10.2 GLOSSARY OF TERMS ............................................................................................................................164 10.3 STANDARDIZED AND RESERVED CODES .......................................................................................165 10.3.1 Reserved Domain Codes...............................................................................................................165 10.3.2 Electrocardiogram Test Codes (for measured or calculated parameters) ......................................167 10.3.3 Vital Signs Test Codes ..................................................................................................................167 10.3.4 Supplemental Qualifiers Name Codes ..........................................................................................168 10.3.5 Trial Summary Codes ...................................................................................................................169 10.4 CDISC VARIABLE-NAMING FRAGMENTS ...............................................................................................171 10.5 LESSONS LEARNED FROM THE PILOT ......................................................................................................173 10.6 REVISION HISTORY ........................................................................................................................175 . 10.7 REPRESENTATIONS AND WARRANTIES; LIMITATIONS OF LIABILITY, AND DISCLAIMERS ..........183

10

Page 4 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

1 Introduction
1.1 PURPOSE
This document describes the CDISC Version 3.1.1 (V3.1.1) Submission Data Standards (SDS), which have been prepared by the Submissions Data Standards team of the Clinical Data Interchange Standards Consortium (CDISC). Like its predecessors, V3.1.1 is intended to guide the organization, structure, and format of standard clinical trial tabulation datasets submitted to a regulatory authority such as the US Food and Drug Administration (FDA). V3.1.1 supersedes all prior versions of the CDISC Submission Data Standards. This document should be used in close concert with the CDISC Study Data Tabulation Model (SDTM) available at http://www.cdisc.org/models/sds/v3.1/index.html and describes how to implement the SDTM for use with data tabulations submitted for human clinical trials. Version 1.1 of the SDTM, which should be read before this Implementation Guide (SDTMIG), describes a general conceptual model for representing clinical study data that is submitted to regulatory authorities. SDTMIG V3.1.1 provides specific domain models, assumptions, business rules, and examples for preparing standard datasets that are based on the SDTM. Tabulation datasets, which are electronic listings of individual observations for a subject that comprise the essential data collected in a clinical trial, are one of four types of data currently submitted to the FDA along with patient profiles, listings, and analysis files. By submitting tabulations that conform to the standard structure, sponsors may benefit by no longer having to submit separate patient profiles or listings with a product marketing application. V3.1.1 is not currently intended to fully meet the needs supported by analysis datasets, which will continue to be submitted separately in addition to the tabulations. Since July 2004, the FDA has referenced use of the SDTM in the Study Data Specifications for the Electronic Common Technical Document, available at http://www.fda.gov/cder/regulatory/ersr/Studydata-v1.1.pdf. The availability of standard submission data will provide many benefits to regulatory reviewers. Reviewers can be trained in the principles of standardized datasets and the use of standard software tools, and thus be able to work with the data more effectively with less preparation time. Another benefit of the standardized datasets is to support the FDAs efforts to develop a repository for all submitted trial data and a suite of standard review tools to access, manipulate, and view the tabulations. Use of these data standards is also expected to benefit industry by streamlining the flow of data from collection through submissions, and facilitating data interchange between partners and providers. Note that the SDTM represents an interchange standard, rather than a presentation format -it is assumed that tabulation data will be transformed by software tools to better support viewing and analysis. This document is intended for companies and individuals involved in the collection, preparation, and analysis of clinical data that will be submitted to regulatory authorities. Audiences are also advised to read the CDISC Submission Metadata Model available at http://www.cdisc.org/pdf/SubmissionMetadataModelV2.pdf for additional historical background on how to provide metadata descriptions for submission data. The primary goal of the Metadata Model is to provide regulatory reviewers with a clear understanding of the datasets provided in a submission by communicating clear descriptions of the structure, purpose, attributes, and contents of each dataset and dataset variable. Guidance, specifications, and regulations for the application of this model will be provided separately by regulatory authorities. Audiences are advised to refer to these guidance documents for the most current recommendations for the submission of clinical data.

CDISC, 2005. All rights reserved FINAL

Page 5 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

1.2

ORGANIZATION OF THIS DOCUMENT

This document is organized into the following sections: Section 1, INTRODUCTION, provides an overall introduction to the V3.1.1 models and describes changes from prior versions. Section 2, FUNDAMENTALS OF THE SDTM, recaps the basic concepts of the SDTM, and describes how this implementation guide should be used in concert with the SDTM. Section 3, SUBMITTING DATA IN STANDARD FORMAT, explains how to describe metadata for regulatory submissions, and how to assess conformance with the standards. Section 4, ASSUMPTIONS FOR DOMAIN MODELS, describes basic concepts, business rules, and assumptions that should be taken into consideration before applying the domain models. Section 5, MODELS FOR SPECIAL PURPOSE DOMAINS, describes the Demographics and Comments special-purpose domains. Section 6, DOMAIN MODELS BASED ON THE GENERAL CLASSES, provides specific annotated metadata models based on the three general observation classes. These include the revised V3.1 models for the ten common domains previously modeled by CDISC in Version 2.0; the domains for Inclusion/Exclusion Exceptions, Subject Characteristics, and Substance Use introduced in Version 3.0; and the Questionnaires domain introduced with V 3.1. Section 7, TRIAL DESIGN DATASETS, describes implementation issues related to the use of the Trial Design Model described in the SDTM. Section 8, REPRESENTING DATA RELATIONSHIPS, describes how to represent relationships between separate domains, datasets, and/or records. Section 9, IMPLEMENTATION EXAMPLES, provides specific examples based on actual clinical-trial data. Several of these examples were prepared as a result of the V3 Pilot Project conducted in the Summer of 2003. Section 10, APPENDICES, provides additional background material and describes other supplemental material relevant to implementation.

1.3

RELATIONSHIP TO PRIOR CDISC DOCUMENTS

As stated above, this document, Version 3.1.1, together with the SDTM, represents the most recent version of the CDISC Submission Data Domain Models, previously known as V3.1. Since all updates to Version 3.1 are intended to be backward compatible with version 3.1, the term V3.x is used to refer to Version 3.1 and all subsequent versions. The most significant changes since V3.1 in this version include: The implementation guide has been expanded with new variables for consistency with Version 1.1 of the SDTM Sponsors may now include any valid variable from the same SDTM general class in a domain model. Previously, domain models were restricted to only those qualifier variables specified in the Implementation Guide A new Trial Summary (TS) dataset has been added to the Trial Design Model to describe summary characteristics of the study
CDISC, 2005. All rights reserved FINAL

Page 6 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

A preferred alternative means of submitting supplemental qualifiers in separate datasets per domain has been defined in Section 8. Substantial clarifications in Section 8 A new section on warranties and disclaimers has been added as Appendix 10.7 Numerous corrections have been applied to the text, assumptions, domain models and examples throughout.

A detailed list of changes between versions is provided in Appendix 10.6. Note that V3.1.1 continues to represent most data in a vertical structure. Standard horizontal listings of data, such as those described in the V2 horizontal representations of ECGs and Vitals by visit, will be produced by FDA standard review tools. A complete list of changes from V3.1 is included in the V3.1.1 Change Log (see Section 10.6.3). V3.1 was the first fully implementation-ready version of the CDISC Submission Data Standards that was directly referenced by the FDA for use on clinical studies involving human drug products. However, future improvements and enhancements such as V3.1.1 will continue to be made as sponsors develop more experience submitting data in this format. Therefore, CDISC will be preparing regular updates to the implementation guide to provide corrections, clarifications, additional domain models, examples, business rules, and conventions for using the standard domain models. CDISC will produce further documentation on controlled terminology as separate publications beginning in 2005, so sponsors are encouraged to check the CDISC website (www.cdisc.org/standards/) frequently for additional information.

CDISC, 2005. All rights reserved FINAL

Page 7 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

1.4

HOW TO READ THIS IMPLEMENTATION GUIDE

The SDS V3.1.1 Implementation Guide (SDTMIG) is best read online, so the reader can benefit from the many hyperlinks included to both internal and external references. The following guidelines may be helpful in reading this document: 1. 2. 3. 4. 5. 6. First read the SDTM to gain a general understanding of the conceptual model for the Submission Data Standards. Next, read Sections 1-3 of this document to review the key concepts for preparing domains and submitting data to regulatory authorities. Refer to the Glossary in Section 10.2 as necessary. Read the General Assumptions for all Domains in Section 4. Review Sections 5 and 6, referring back to Assumptions as directed (hyperlinks are provided). Review the Implementation examples for each domain in Section 9 to gain an understanding of how to apply the domain models for specific types of data. Read Section 7 to understand the fundamentals of the Trial Design Model and consider how to apply the concepts for typical protocols. Since the Trial Design model includes many new concepts, readers may choose to defer this step until after they have mastered the essentials of creating standardized domains. New extensions to the trial design model will be published separately on the CDISC web site. Review Section 8 to learn the advanced concepts of how to express relationships between datasets, records and additional variables not specifically defined in the models. Finally, review the additional appendices as appropriate.

7. 8.

1.5

SUBMITTING COMMENTS

Comments on this document can be submitted through the CDISC Discussion Board.

Page 8 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

2 Fundamentals of the SDTM

2.1 OBSERVATIONS AND VARIABLES
The V3.x submission data standards are based on the SDTM Version 1.1 document, which provides a general framework for organizing information collected during clinical trials that is to be submitted to the FDA. The SDTM is built around the concept of observations collected about subjects who participated in a clinical study. Each observation can be described by a series of variables, corresponding to a row in a dataset or table. Each variable can be classified according to its Role. A Role determines the type of information conveyed by the variable about each distinct observation and how it can be used. Variables can be classified into four major roles: Identifier variables, which identify the study, subject of the observation, the domain, and the sequence number of the record Topic variables, which specify the focus of the observation (such as the name of a lab test) Timing variables, which describe the timing of the observation (such as start date and end date) Qualifier variables, which include additional illustrative text, or numeric values that describe the results or additional traits of the observation (such as units or descriptive adjectives).

A fifth type of variable role, Rule, was introduced with V3.1 to express an algorithm or executable method to define start, end, or looping conditions in the Trial Design model. The set of Qualifier variables can be further categorized into five sub-classes: Grouping Qualifiers are used to group together a collection of observations within the same domain. Examples include --CAT and --SCAT. Result Qualifiers describe the specific results associated with the topic variable for a finding. It is the answer to the question raised by the topic variable. Examples include --ORRES, --STRESC, and --STRESN. Many of the values in the DM domain are also classified as Result Qualifiers. Synonym Qualifiers specify an alternative name for a particular variable in an observation. Examples include --MODIFY and --DECOD, which are equivalent terms for a --TRT or --TERM topic variable, --TEST and --LOINC which are equivalent terms for a --TESTCD. Record Qualifiers define additional attributes of the observation record as a whole (rather than describing a particular variable within a record). Examples include --REASND, AESLIFE, and all other SAE flag variables in the AE domain; and --BLFL, --POS and --LOC, --SPEC, --LOT, --NAM. Variable Qualifiers are used to further modify or describe a specific variable within an observation and is only meaningful in the context of the variable they qualify. Examples include --ORRESU, --ORNRHI, and --ORNRLO, all of which are variable qualifiers of --ORRES, and --DOSU and --DOSFRM, all of which are variable qualifiers of --DOSE.

For example, in the observation, 'Subject 101 had mild nausea starting on Study Day 6,' the Topic variable value is the term for the adverse event, 'NAUSEA'. The Identifier variable is the subject identifier, '101'. The Timing variable is the study day of the start of the event, which captures the information, 'starting on Study Day 6', while an example of a Record Qualifier is the severity, the value for which is 'MILD'. Additional Timing and Qualifier variables could be included to provide the necessary detail to adequately describe an observation.
CDISC, 2005. All rights reserved FINAL Page 9 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

2.2

DATASETS AND DOMAINS

Observations are normally collected for all subjects in a series of domains. A domain is defined as a collection of logically-related observations with a topic-specific commonality about the subjects in the trial. The logic of the relationship may relate to the scientific subject matter of the data, or to its role in the trial. Typically, each domain is represented by a dataset, but it is possible to have information relevant to the same topicality spread among multiple datasets (see Section 8.6). Each dataset is distinguished by a unique, two-character DOMAIN code that should be used consistently throughout the submission. This DOMAIN code is used in the dataset name, the value of the DOMAIN variable within that dataset, and as a prefix for most variable names in the dataset. The dataset structure for observations is a flat file representing a table with one or more rows and columns. Normally, one dataset is submitted for each domain. Each row of the dataset represents a single observation and each column represents one of the variables. Each dataset or table is accompanied by metadata definitions that provide information about the variables used in the dataset. The metadata are described in a data definition document named 'Define' that is submitted along with the data to regulatory authorities. (See the Case Report Tabulation Data Definition Specification (define.xml), available on the CDISC website, for information about an xml representation of the data definition document.) The CDISC Submission Metadata Model uses seven distinct metadata attributes to be defined for each dataset variable in the metadata definition document: The Variable Name (limited to 8-characters for compatibility with the SAS Transport format) A descriptive Variable Label, using up to 40 characters, which should be unique for each variable in the dataset The data Type (e.g., whether the variable value is a character or numeric) The set of controlled terminology for the value or the presentation format of the variable (Controlled Terms or Format) The Origin or source of each variable The Role of the variable, which determines how the variable is used in the dataset. For the V3.x domain models, roles are used to represent the categories of variables as Identifier, Topic, Timing, or the five types of Qualifiers. Since these roles are predefined for all domains that follow the general classes, they do not need to be specified by sponsors in their Define data definition document. Actual submission metadata may use additional role designations, and more than one role may be assigned per variable to meet different needs. Comments or other relevant information about the variable or its data.

Data stored in dataset variables include both raw (as originally collected) and derived values (e.g., converted into standard units, or computed on the basis of multiple values, such as an average). In the SDTM only the name, label, and type are listed with a set of CDISC guidelines that provide a general description for each variable used by a general observation class. The Domain models included in Sections 5 and 6 of this document provide additional information about Controlled Terms or Format and Origin, as well as notes on proper usage. Comments are included as necessary according to the needs of individual studies. The presence of an asterisk (*) in the 'Controlled Terms or Format' column indicates that a discrete set of values (controlled terminology) is expected to be made available for this variable. This set of values may be sponsor-defined in cases where standard vocabularies have not yet been defined (represented by a single *) or from an external published source such as MedDRA (represented by **). The CDISC controlled terminology group will be publishing additional guidance on use of controlled terminology separately in the future.

Page 10 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

2.3

SPECIAL-PURPOSE DOMAINS

The CDISC V3.x Submission Data Domain Models include special-purpose domains with a specific structure and cannot be extended with any additional qualifier or timing variables other than those specified. Two of these Demographics and Comments -- are described in Section 5 of this document. .Demographics includes a set of standard variables that describe each subject in a clinical study; Comments describes a fixed structure for recording free-text comments on a subject, or comments related to records or groups of records in other domains. Additional fixed structure, non-extensible special-purpose domains are discussed in the Trial Design model described in Section 7 and in Section 8, which discusses record relationships.

2.4

THE GENERAL DOMAIN CLASSES

Most observations collected during the study (other than those represented in special purpose domains) should be divided among three general observation classes: Interventions, Events, or Findings: The Interventions class captures investigational treatments, therapeutic treatments, and surgical procedures that are intentionally administered to the subject (with some actual or expected physiological effect) either as specified by the study protocol (e.g., exposure), coincident with the study assessment period (e.g., concomitant medications), or other substances self-administered by the subject (such as alcohol, tobacco, or caffeine) The Events class captures occurrences or incidents independent of planned study evaluations occurring during the trial (e.g., 'adverse events' or 'disposition') or prior to the trial (e.g., 'medical history'). The Findings class captures the observations resulting from planned evaluations to address specific questions such as observations made during a physical examination, laboratory tests, ECG testing, and sets of individual questions listed on questionnaires.

In most cases, the identification of the general class appropriate to a specific collection of data by topicality is straightforward. Often the Findings general class is the best choice for general observational data collected as measurements or responses to questions. In cases when the topicality may not be as clear, the choice of class may be based more on the scientific intent of the protocol or analysis plan or the data structure. Appendix 10.6 of this document proposes additional guidelines for choosing the appropriate general class and proposes a modeling approach where a findings structure should be used to represent data that may relate to interventions or events records. All datasets based on any of the general observation classes share a set of common Identifier variables and Timing variables, which are described in the SDTM. Three general rules apply when determining which variables to include in a domain: The same set of Identifier variables applies to all domains based on the general observation classes. An optional identifier can be used wherever appropriate. Any valid Timing variable is permissible for use in any submission dataset (such as to describe studies with more precise time points such as a Pharmacokinetics trial), but it should be used consistently where applicable for all domains. Any additional Qualifier variables from the same general class may be added to a domain model.

Assumptions for use of the Domain are described in Section 4 of this document.

CDISC, 2005. All rights reserved FINAL

Page 11 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

2.5

THE CDISC STANDARD DOMAIN MODELS

The following standard domains with their respective DOMAIN codes have been defined or referenced by CDISC in this document (domain models marked with an asterisk (*) are not included in this document and will be posted separately for comment): Special-Purpose Domains: Demographics - DM Interventions: Concomitant Medications - CM Substance Use - SU Events: Adverse Events - AE Medical History - MH Findings: *Drug Accountability - DA Inclusion/Exclusion Exceptions - IE *Microbiology Specimens - MB *Microbiology Susceptibility - MS *Pharmacokinetics Concentrations - PC *Pharmacokinetics Parameters - PP Trial Design Domains: Trial Elements - TE Trial Arms - TA Trial Visits - TV Subject Elements - SE Subject Visits - SV Trial Inclusion/Exclusion Criteria TI Trial Summary - TS Special-Purpose Relationship Datasets (defined in Section 8): Supplemental Qualifiers - SUPPQUAL Relate Records - RELREC ECG Tests - EG Laboratory Tests - LB Questionnaires - QS Physical Examinations - PE Subject Characteristics - SC Vital Signs - VS Disposition - DS *Protocol Deviations - DV Exposure - EX

Comments - CO

Note, a sponsor would only submit the data domains that are actually collected. Decisions on what data to collect should be based on the scientific objectives of the study. A list of standard domain codes for other commonly used domains is provided in Appendix 10.3.1. In the future, additional domain models based on the three general observation classes, and additional standard domain codes will be published by CDISC, and sponsors are encouraged to check the CDISC web site periodically for updates.

Page 12 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) When preparing submissions based on the domain models, sponsors should not add any variables other than additional relevant timing variables and qualifiers from the same general class to the V3.x models, since non-standard variables could compromise the FDAs abilities to populate the data repository and use standard tools. A sponsor is free to drop certain variables from the domain model, and the corresponding descriptions from the data definition document, but new variables (other than those that are from the same general class) must not be added, and existing variables should not be renamed, or modified for novel usage. When evaluating how to represent data according to the V3.x model, sponsors should keep in mind that the V3/V3.x paradigm is very different from that of previous versions. Data that historically would have been modeled horizontally into one record is now often converted into multiple records in V3.x. If additional variables remain that cannot be mapped to the variables defined for a standard domain model, Sponsors should use the Supplemental Qualifiers dataset described in Section 8.4 of this document. Note that any collected data in an analysis dataset must also appear in a tabulation dataset. If a sponsor wishes to propose that new variables be added to the Version 3.x models, they should provide the rationale and description of the added variable(s) to the CDISC SDS Team so the impact on V3.x models can be assessed. Specific examples are also requested to illustrate the context for the new variables. This information can be provided to the CDISC SDS Team by posting it on the CDISC Public Discussion Forum.

2.6

CREATING A NEW DOMAIN

This section describes the overall process of how to create a new CDISC SDTM domain. New domains should only be created if a published standard domain does not already exist for the data being modeled. When a new domain is created, one of the most important decisions is the choice of the proper general class. Deciding on the number of separate domains to use within a general class should be made based on the specific requirements of the study. See Section 9 of this document for specific examples of the application of the CDISC SDTM for Interventions, Events and Findings general domain models. When creating a new domain, sponsors should follow the predefined steps listed below: 1. Ensure that there is a definite need to create a new domain. For example, the Findings vertical structure can often accommodate many types of information that may have previously spanned separate domains in a sponsors internal database. In V3.x, such information can often be represented as separate records in the same submissions domain. Mechanisms for incorporating new Qualifier variables not described in the general observation classes and for defining relationships between separate datasets or records are described in Section 8 of this document. Once the need for a new domain is determined, verify that there are no existing domain models by reviewing the current models posted on the CDISC website. Choose the general observation class (Interventions, Events, or Findings) that best fits the data as follows: a. Identify the topic of the observation and determine which of the three general observation classes it most closely resembles. If the new domain shares both the same topicality and general observation class as an existing domain in the submission, the existing domain should be used. Look for other existing domain models that may serve as a relevant prototype (most domains will follow the Findings model). Determine if the chosen general observation class has most of the required and expected qualifiers for the new domain. Select the variables that apply to the new domain once you have selected the general observation class.

2. 3.

b. c. d. 4. 5.

Select and include the identifier variables (USUBJID, STUDYID, and --SEQ). Include the topic variable from the identified general observation class (e.g., --TESTCD for Findings).
Page 13 July 28, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) 6. 7. 8. Select and include the relevant Qualifier variables from the identified general observation class (e.g., Findings) omitting any extraneous variables from the general model that do not apply. Select and include the applicable timing variables from the SDTM. Check with the CDISC website for a previously identified two-character domain identifier or abbreviation. If one has not been assigned by CDISC, then the sponsor may select the unique twocharacter domain code to be used consistently throughout the submission. Apply the two-character domain code to the appropriate variables in the domain. a. Replace all variable prefixes shown in the models with two hyphens '--' with the domain code.

9.

10. Set the order of variables consistent with the order of the domain model most similar to the new domain. 11. Adjust the labels of the variables only as appropriate to properly convey the meaning in the context of the data being submitted in the newly-created domain. Use title case for all labels (title case means to capitalize the first letter of every word except for articles, prepositions, and conjunctions). 12. Compare for consistency with CDISC General Assumptions and other variables used in similar domain examples to ensure new variables are not being created when appropriate standard variables already exist. 13. Label the dataset within the metadata definition document as follows: <the domain name>, <model type: Findings, Events, Interventions>, CDISC SDTM Version <Version Number e.g., 3.1.1>, <dataset structure (e.g., one record per subject or one record per event) >. 14. Submit a request to add specific new variables necessary to represent additional information in the general observation classes through the CDISC Public Discussion Forum. As noted in Step 8 of the above process, each domain is distinguished by a unique, two-character DOMAIN code that should be used consistently throughout the submission. The two-character identifier should be controlled for consistency so that the FDA tools for receiving data can recognize it. The DOMAIN code should also be included in the name of the corresponding dataset files (e.g., ae.xpt) and also used as a prefix for variables to distinguish them in the event that merges are performed between datasets using SAS. CDISC will maintain a separate list of standard domain codes on the CDISC website and sponsors may post additional suggested codes to add to this list through the CDISC Public Discussion Forum.

Page 14 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

2.6.1

How to Include Variables in New Domains (Steps 4 through 7)

Superset of CDISC SDS Variables

Variables for General Domain Classes

Variables for Special Purpose Domains (Fixed & Study Design Domains)

Identifier Variables

Timing Variables

Topic & Qualifier Variables

Demography & Comment Domain Variables

Study Design Variables

Interventions Specific Variables

Events Specific Variables

Findings Specific Variables

New Domain

As illustrated in the figure above, the CDISC SDTM has an inclusive superset or a predefined list of variables. The SDTM variables are subdivided into two general variable classes: those for the 'general observation classes' and those for the 'special domains' (those with fixed structures and the trial design domains). For more information regarding these classifications, please see sections 2.3 and 2.4. From a data modeling perspective, once it has been determined that a new domain is needed (i.e., there is no published equivalent domain), the sponsor determines the domain's general observation class (either as an Intervention, Event, or Finding) as described in steps 1, 2 and 3 above. Next the sponsor must go through a process of selecting the variables to include in the new domain (Steps 4 through 7). All new domains will contain variables from the Identifier (e.g., subject of the observation, the domain) and Timing (e.g., visit designators, start and end dates) variable classes. The 'Identifier Variables' and 'Timing Variables' boxes being connected to the 'New Domain' box via a bolded solid black line illustrate this. The new domain, however, can only inherit the Topic and Qualifier variables from the specific general observation class that it was determined to fit. Therefore, if a new domain has been classified as a Findings domain, it cannot contain Qualifier variables that were defined for the Intervention or Events domain classes and vice versa. This principle is illustrated in the graphic by the 'Intervention', 'Events', and 'Findings' boxes being connected to the 'New Domain' box via dashed lines. In summary, all new domains include Identifier and Timing variables. However, the choice of Topic and Qualifier variables that can be included is dependent upon which general observation class the domain has been modeled as 'Interventions', 'Events', or 'Findings'.

CDISC, 2005. All rights reserved FINAL

Page 15 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

3 Submitting Data in Standard Format

3.1 STANDARD METADATA FOR DATASET CONTENTS AND ATTRIBUTES
The V3.x CDISC SDS domain models provide a standard depiction of some of the most commonly used data domains, using the metadata attributes originally described in the CDISC Submission Metadata Model. The descriptive metadata attributes that should be included in a submission dataset definition file as applied in the domain models are: The CDISC-standard variable name (standardized for all submissions, even though sponsors may be using other variable names internally in their operational database) The CDISC-standard variable label Data type of the variable (character or numeric, to conform to the data types recognized by SAS) Controlled terms and formats used by the sponsor (the CDISC domain models use a single asterisk (*) or a double asterisk (**) to indicate when controlled terminology applies) The origin or source of the data (e.g., CRF page, derived) The role of the variable in the dataset corresponding to the role in the SDTM (Since these roles are predefined for all standard domains that follow the general classes, they do not need to be specified by sponsors in their Define data definition document.)

In addition to these metadata attributes, the CDISC domain models include two other columns to assist sponsors in preparing their datasets one column for notes relevant to the use of each variable, and one to indicate how a variable is classified as a CDISC Core Variable. The concept of core variable is used both as a measure of compliance and to provide general guidance to sponsors. CDISC Core Variables fall into three categories: A Required variable is any variable that is basic to the identification of a data record (i.e., essential key variables and a topic variable) or is necessary to make the record meaningful. Required variables should always be included in the dataset and cannot be null for any record. An Expected variable is any variable necessary to make a record useful in the context of a specific domain. Columns for Expected variables are assumed to be present in each submitted dataset even if some values are null. A Permissible variable should be used in a domain as appropriate when collected or derived. All Timing variables (including those not explicitly included in a domain model) and any Qualifier variable specified in a domain model are permissible for use in that domain. Null values are allowed, but the Sponsor can decide whether a Permissible variable should be included as a column when all values for that variable are null.

The Core Variable column provides guidance on which variables are normally required, expected or permissible to include in an actual dataset based on a model (see Section 4.1.1.5). However, any decisions regarding the specific content that is to be included for any submission should always be discussed in advance with the regulatory agency.
Page 16 August 26, 2005 CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) The domain models illustrate how to apply the SDTM when creating a specific domain dataset. In particular, these models illustrate the selection of a subset of the variables offered in one of the general observation classes along with applicable timing variables. The models also show how a standard variable from a general observation class should be adjusted to meet the specific content needs of a particular domain, including making the label more meaningful, specifying controlled terminology, and creating domain-specific notes and examples. Thus the domain models demonstrate not only how to apply the model for the most common domains, but also gives insight on how to apply general model concepts to other domains not yet defined by CDISC.

3.2

USING THE CDISC DOMAIN MODELS IN REGULATORY SUBMISSIONS

New users will find the V3.x CDISC domain models useful as a template for preparing submission files; however, users should be aware that the set of descriptive attributes shown in the models is not precisely the same as that required in the dataset definition document that must accompany a submission. The specific differences are as follows: The column reserved for sponsor comments in the data definition document is not included in the CDISC domain models, since this needs to be created and populated by the sponsor for each dataset to explain any sponsor-specific rules and definitions The CDISC Notes column should not be submitted to the agency The CDISC Core column should not be submitted to the agency The CDISC References column should not be submitted to the agency.

The last three columns in the models that are not to be submitted are shaded so they can easily be distinguished. In addition to the variable notes, the CDISC models include a set of general assumptions applicable to all domains that describe many of the critical concepts behind the models (see Section 4). A set of specific assumptions applicable only to a particular domain is listed after the data definitions for that domain model. In summary, the main body of the Define data definition document submitted by a sponsor should display for each data domain the 'Variable Name', 'Variable Label', 'Type', 'Controlled Terms or Formats', 'Origin' and 'Comments' (Role is optional for standard domains). The dataset definition document that accompanies a submission should also describe each dataset that is included in the submission and describe the logical key structure of each dataset. While most studies will include DM and a set of safety domains based on the three general observation classes (typically including EX, CM, AE, DS, MH, IE, LB, and VS), the actual choice of which data to submit will depend on the protocol and the needs of the regulatory reviewer. Dataset definition metadata should include dataset filenames, descriptions, locations, structures, purpose, keys, and comments as described in the CDISC Submission Metadata Model and shown below in Section 3.2.1. Note that key variables should refer to the natural keys that determine the order of records within the dataset. The physical key structure that uniquely identifies each record in an SDTM dataset based on a general class is actually based on the STUDYID, USUBJID and --SEQ variable in each domain.

3.2.1
Dataset

CDISC Submission Dataset Definition Metadata

Description Location Structure Purpose Key Variables

DM CO

Demographics Comments

dm.xpt co.xpt

One record per subject One record per comment per subject One record per medication intervention episode per subject

Tabulation Tabulation

STUDYID, USUBJID STUDYID, USUBJID, COSEQ STUDYID, USUBJID, CMTRT, CMSTDTC

Page 17 July 28, 2005

CM

Concomitant Medications

cm.xpt

Tabulation

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Key Variables

Dataset

Description

Location

Structure

Purpose

EX

Exposure

ex.xpt

One record per constant dosing interval per subject

Tabulation

STUDYID, USUBJID, EXTRT, EXSTDTC STUDYID, USUBJID, VISITNUM, SUTRT STUDYID, USUBJID, AETERM, AESTDTC STUDYID, USUBJID, DSSTDTC STUDYID, USUBJID, MHTERM STUDYID, USUBJID, EGTESTCD, VISITNUM,T PTNUM, EGSEQ STUDYID, USUBJID, IETESTCD STUDYID, USUBJID, LBTESTCD VISITNUM, TPTNUM STUDYID, USUBJID, VISITNUM, PETESTCD, STUDYID, USUBJID, QSTESTCD, VISITNUM, TPTNUM, QSSEQ STUDYID, USUBJID, SCTESTCD STUDYID, USUBJID, VSTESTCD, VISITNUM, TPTNUM STUDYID, USUBJID, DVSEQ STUDYID, USUBJID, MBTESTCD, MBSEQ

SU

Substance Use

su.xpt

One record per substance type per visit per subject

Tabulation

AE

Adverse Events

ae.xpt

One record per adverse event per subject One record per disposition status or protocol milestone per subject One record per medical history event per subject One record per ECG observation per time point per visit per subject

Tabulation

DS MH

Disposition Medical History

ds.xpt mh.xpt

Tabulation Tabulation

EG

ECG

eg.xpt

Tabulation

IE

Inclusion/ Exclusion Exceptions Laboratory Tests

ie.xpt

One record per Inclusion/Exclusion criteria exception per subject One record per lab test per time point per visit per subject

Tabulation

LB

lb.xpt

Tabulation

PE

Physical Exam

pe.xpt

One record per body system per visit per subject

Tabulation

QS

Questionnaires

qs.xpt

One record per question per time point per visit per subject

Tabulation

SC

Subject Characteristics Vital Signs

sc.xpt

One record per characteristic per subject One record per vital sign measurement per time point per visit per subject

Tabulation

VS

vs.xpt

Tabulation

DV*

Protocol Deviations Microbiology

dv.xpt

One record per protocol deviation per subject One record per specimen test or susceptibility test

Tabulation

MB*

mb.xpt

Tabulation

Page 18 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Key Variables

Dataset

Description

Location

Structure

Purpose

MS*

Microbiology Susceptibility

ms.xpt

One record per susceptibility observations.

Tabulation

STUDYID, USUBJID, MSTESTCD, MSSEQ STUDYID, USUBJID, DATESTCD, DADTC, DASEQ STUDYID, USUBJID, PCTESTCD, VISITNUM, TPTNUM STUDYID, USUBJID, PPTESTCD, VISITNUM STUDYID, ETCD STUDYID, ARMCD, ETCD STUDYID, VISITNUM, ARMCD STUDYID, USUBJID, ETCD STUDYID, USUBJID, VISITNUM STUDYID, IETESTCD STUDYID, TSPARMCD, TSSEQ STUDYID, RDOMAIN, USUBJID, IDVAR, IDVARVAL, RELID STUDYID, RDOMAIN, USUBJID, IDVAR, IDVARVAL, QNAM

DA*

Drug Accountability

da.xpt

One record per accountability observation

Tabulation

PC*

PK Concentrations

pc.xpt

One record per concentration per analyte

Tabulation

PP*

PK Parameters

pp.xpt

One record per PK parameter per analyte

Tabulation

TE TA

Trial Elements Trial Arms

te.xpt ta.xpt

One record per element One record per planned element per arm One record per planned visit per arm One record per actual element per subject One record per subject per actual visit One record per I/E criterion One record per trial summary parameter. One record per relationship

Tabulation Tabulation

TV

Trial Visits

tv.xpt

Tabulation

SE

Subject Elements Subject Visits

se.xpt

Tabulation

SV

sv.xpt

Tabulation

TI TS

Trial Inclusion/ Exclusion Criteria Trial Summary

ti.xpt ts.xpt relrec.xpt

Tabulation Tabulation

RELREC

Related Records

Tabulation

SUPPQUAL

Supplemental Qualifiers

suppqual.xpt

One record per qualifier value

Tabulation

*Domain models to be published separately for comment

3.2.2

CDISC Submission Value-Level Metadata

In general, the CDISC Version 3.x data models are more closely related to normalized relational data models in a vertical structure. This structure requires every row to have a primary key since each row of a
CDISC, 2005. All rights reserved FINAL Page 19 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1) dataset represents a single observation and each column represents one of the variables contained within the dataset. Since the V3.x data structures are essentially static, much of the information that may have been communicated in columns in a more horizontal structure will now involve adding new records instead. The model also includes variables that contain values that can be used to convert the dataset from the 'vertical' format to a 'horizontal' representation (or more 'denormalized'). Therefore, for some domains there is a need to provide record-level metadata (referred to as value-level or hierarchical metadata). For example, the Vital Signs data domain could contain subject records related to diastolic and systolic blood pressure, height, weight, and body mass index (BMI). If these data were submitted in compliance with the CDISC standards, it would be provided in the more normalized structure of one row per vital signs measurement. This means that there could be five records per subject (one for each parameter), and the parameter names would be stored in the Test Code/Name variables, and the parameter value in a result variable. Since the unique Test Code/Names have different attributes (i.e., different origins, roles, and definitions) there is a need to provide metadata for this information. The hierarchical metadata could be provided as a separate section of the Define data definition document. The table should be similar in structure to the CDISC Submission Metadata Model described previously. This information, which historically has been submitted as a pdf document named 'Define.pdf' may also be submitted in an XML format. For details on the CDISC specification for submitting the Define data definition document in XML, see http://www.cdisc.org/models/sds/v3.1/index.html.

3.2.3

Conformance
Following the complete metadata structure for data domains and variables Following CDISC domain models wherever applicable Including all collected and relevant derived data in one of the standard domains, special-purpose datasets or general-observation-class structures Including all Required and Expected variables as columns in a domain Using CDISC-specified standard domain names and prefixes Using CDISC-specified standard variable names Using CDISC-specified variable labels for all standard domains (as described in the SDTMIG) Using CDISC-specified data types for all variables Following CDISC-specified controlled terminology and format guidelines for variables, when provided Ensuring that each record in a dataset includes a set of keys and a topic variable Conforming to all business rules described in the CDISC notes and general and domain-specific assumptions.

Conformance with the CDISC Domain Models is indicated by:

When creating new domains, sponsors should always review the CDISC-standard labels from comparable domains and adjust only if necessary to properly convey the meaning of the submitted data. Sponsors should also supply the correct origins, roles (optional for standard domains), and appropriate comments, plus any additional controlled terminology or format information required for the FDA reviewer to properly interpret the data. Since most regulatory submissions involve data that has been collected over many years, CDISC recognizes that full conformance with the SDS model may not be immediately achievable, but will improve over time.

Page 20 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

4 Assumptions for Domain Models

4.1
4.1.1
4.1.1.1

GENERAL ASSUMPTIONS FOR ALL DOMAINS

General Dataset Assumptions
Review the Study Data Tabulation Model as well as this Implementation Guide before attempting to apply any of the individual domain models. The CDISC Submission Metadata Model may also be reviewed for additional background information on how to represent metadata for general datasets. See the Case Report Tabulation Data Definition Specification (define.xml), available on the CDISC website, for information about an xml representation of the data definition document. Additional analysis variables should be added to analysis datasets if required by regulatory reviewers or if necessary to accommodate the scientific requirements of the submission. These should be named consistently with other variables in the model, described clearly, and presented in a format that is consistent across a submission. Specific assumptions regarding data types should also be applied. However, no new variables should be added to any tabulation dataset except through the Supplemental Qualifiers mechanism described in Section 8. Additional Timing variables from the general observation classes can be added as needed to a standard domain model based on the three general observation classes. The order of variables in the Define data definition document should reflect the order of data in the dataset. The current order of variables in the CDISC domain models has been chosen to facilitate the review of the models and application of the models. Sponsors may thus wish to reorder timing and qualifier variables in order to place more emphasis on the most important variables, but are encouraged to apply a consistent variable ordering scheme for all domains in a submission wherever possible. CDISC Core variables: In V3.0, CDISC identified all Core variables in domain models (identified by a 'Y' in the 'Core' column). In V3.x, three categories of variables are specified in the 'Core' column: A Required variable is any variable that is basic to the identification and meaning of a data record (i.e., essential key Identifiers and a Topic variable). Values cannot be null. An Expected variable is any variable necessary to make a record meaningful in the context of a specific domain. In most cases, some but not all values for expected variables may be null in a domain if unknown or not done. But when an expected variable has not been collected, a null column should still be included and a comment should be included in the define data definition document to state it was not collected. A Permissible variable may be used as appropriate, when collected or derived.
Page 21 July 28, 2005

4.1.1.2

4.1.1.3

4.1.1.4

4.1.1.5

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) All timing variables are considered permissible. Any qualifier variables from the same general class are permissible for that domain. Required and expected variables must be included in the dataset. Required variables cannot be null, but the expected variables may contain some null values. In studies where limited data collection has been agreed upon in advance with the agency, some expected variables might be submitted with all null values. Normally the sponsor should only include those variables designated as permissible that were collected or derived; however, the sponsor can choose whether permissible variables that are always null should be included. o o 4.1.1.6 Additional guidance on dataset naming will be provided in future versions of the CDISC SDTMIG.

4.1.2
4.1.2.1

General Variable Assumptions

Data variable names should be limited to 8 characters, and cannot start with a number, nor should they contain characters other than letters, numbers, or underscores. This is a SAS V5 Transport file limitation, and since use of SAS V5 is specified in the current guidance this limitation will be in effect until the use of other formats (such as XML) becomes acceptable to regulatory authorities. Data variable descriptive names (labels), up to 40 characters, should be provided as data variable labels. This is a correction to the 32 characters noted as a limitation in the FDA guidance document, Providing Regulatory Submissions in Electronic Format NDAs (IT-3, January 1999). Use of variable names (other than domain prefixes), formats, decodes, terminology, and data types for the same type of data should be consistent within and across studies within a submission. In general, sponsors should use the CDISC-standard labels in all standard domains.

4.1.2.2

In order to minimize the risk of difficulty when merging/joining domains for reporting purposes, the two-character Domain Identifier is used as a prefix in most variable names. The rule for applying the prefix is as follows: All variable names should be prefixed with the Domain Identifier in which they reside except: a. b. c. d. Required Identifiers (STUDYID, DOMAIN, USUBJID) Commonly used Merge Keys (VISIT, VISITNUM, VISITDY, and many of the variables in trial design such as ELEMENT and ARM) All Demography domain (DM) variables with the exception of the DM-specific Timing variables DMDTC, and DMDY All variables in RELREC and SUPPQUAL.

Required Identifiers are not prefixed because they are usually used as keys when merging/joining observations. The --SEQ and the optional Identifiers --GRPID and --REFID are prefixed because they are only used as keys when relating observations in special cases. 4.1.2.3 'Subject' should be used where applicable to generically refer to both 'patients' and 'healthy volunteers.' The term 'Subject' should be used consistently in all labels and comments. It is recommended that textual data be submitted in upper case text. Exceptions may include long text data (such as comment text), values of --TEST in Findings datasets (which may be more readable in mixed case if used as labels in transposed views) and certain controlled
CDISC, 2005. All rights reserved FINAL

4.1.2.4

Page 22 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1) terminology (such as MedDRA terms) that are already in mixed case. The Sponsors Define data definition document may indicate as a general note or assumption whether case sensitivity applies to text data for any or all variables in the dataset. If case sensitivity differs across variables, it should be documented at the variable level in the Define document. 4.1.2.5 If a special convention is used for missing values (or values that were not collected for certain records in a Tabulation), it should be described in the Define data definition document and should be used consistently in all datasets. Sponsors may assign categories (--CAT variable) and subcategories (--SCAT variable) within domains. The categories and subcategories will provide additional context for the Topic variable and should be used consistently across all records in the domain. For example, a lab record with LBTEST = 'SODIUM' might have LBCAT = 'CHEMISTRY' and LBSCAT = 'ELECTROLYTES'. The values for these variables may be collected on the CRF or derived. --CAT and --SCAT values should not be redundant with the domain or dictionary classification provided by --DECOD and --BODSYS. That is, they should provide a different means of defining or classifying records. For example, a sponsor may be interested in identifying all adverse events that the investigator feels might represent an infection, and thus will collect that categorization on the CRF. This categorization might differ from the categorization derived from the adverse-event coding dictionary. The intended relationship among the records grouped using --GRPID (see Section 8.1 for details) may appear to duplicate a relationship already identified by other variables such as VISIT, --CAT, --SCAT, or --CLAS. These latter variables have meaning across subjects, whereas the values of --GRPID apply to a specific domain and subject, and are intended solely to facilitate grouping a set of records for a subject. For example, in the Laboratory Test Results dataset, LBTEST values of HEMATOCRIT, HEMOGLOBIN, and RBC might be assigned a value of HEMATOLOGY for LBCAT. LBCAT would be populated for all records falling into that category across all subjects. However, for one subject, the investigator commented on these specific records noting that the subject had a long history of anemia. For this subject, the three hematology records would be assigned a common LBGRPID value. That value would be used in the COMMENTS special-purpose domain to link the comment to the appropriate lab records. 4.1.2.7 Sponsors sometimes collect free text information on the CRF. When the CRF includes a list of values including "Other, Specify", then the free text value should be placed in a SUPPQUAL dataset as described in Section 8.4. For example, when a description of "Other Medically Important Serious Adverse Events" category is collected on a CRF then the free text description should be stored in the SUPPQUAL dataset.

4.1.2.6

4.1.3

Coding and Controlled Terminology Assumptions

4.1.3.1 The Submission Data Standards Team has defined two types of Controlled Terminology (CT). The first consists of CT that has been published externally (e.g., MedDRA, LOINC or maintained by CDISC in Appendix 10.4). The second consists of CT that is defined by a sponsor, and is used consistently throughout a submission (e.g., a standard code list). Two asterisks (**) within the Controlled Terms or Format column indicate the respective variable should be populated with values from an external source. A single asterisk (*) within the Controlled Terms or Format column indicates the respective variable should be populated with a value from a sponsor-controlled list. The values that appear in the Controlled Terms or Format column are expected to be used in submissions. CDISC will be providing future guidance on the use of controlled terminology in separate documents available through the CDISC website. If applicable, additional information is provided within this column for common alpha-codes, decodes, and references.
CDISC, 2005. All rights reserved FINAL Page 23 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

4.1.3.2 It is recommended that controlled terminology be submitted in upper case text other than the two exceptions noted below. Some sponsors may choose to submit controlled terminology in mixed case text to enhance the appearance of reports. As noted in item 4.1.2.4 above, case sensitivity should be described in the Define document. a. b. If the external reference for the controlled terminology is not in upper case then the data should conform to the case prescribed in the external reference (e.g., MedDRA and LOINC). Units, which are considered symbols rather than abbreviated text (e.g., mg/dL).

4.1.3.3 The controlled terminology should be displayed for each applicable variable within the 'Controlled Terms or Format' column of the Define data definition document. Variables with numerous controlled terms may be placed within an appendix section of Define. 4.1.3.4 Controlled terminology or text should be used instead of, or in addition to, arbitrary number codes in order to reduce ambiguity for submission reviewers. For example, for concomitant medications, the verbatim term and/or dictionary term should be presented, rather than numeric codes. Separate coded values may be submitted as supplemental qualifiers or included in the define data definition document and may be necessary in analysis datasets. 4.1.3.5 Controlled terminology for domain topic variables should be stored as follows: For safety events such as AEs and Medical History, fill --DECOD with the dictionarys preferred term and fill --BODSYS with the preferred body system name. If a dictionary is multi-axial, the value in --BODSYS should represent the primary path. When using MedDRA, for example, --DECOD should contain the PT (Preferred Term), and --BODSYS should contain the primary SOC (System Organ Class). For concomitant medications, fill --DECOD with the drug's generic name, and fill --CLAS with the drug class only if the dictionary used codes drugs to a single class. When using WHODRUG, for example, --CLAS would not be filled since a drug may have multiple classes.

In either case, no other intermediate levels or relationships should be stored in the dataset. By knowing the dictionary and version used, the reviewer will be able to obtain intermediate levels in a hierarchy (as in MedDRA), or a drugs ATC codes (as in WHO Drug). The sponsor may be required to submit the dictionary if it is not already available to the reviewer. 4.1.3.6 The topic variable for many of the general domain models is often stored as verbatim text. For an Event domain, the topic variable is --TERM. For an Interventions domain, the topic variable is --TRT. For a Findings domain, the topic variable, --TESTCD, should use Controlled Terminology (e.g., SYSBP for systolic blood pressure). If a CDISC-standard controlled terminology exists it should be used, otherwise the sponsor should define its own controlled list of terms. If the verbatim term is modified to facilitate coding, the modified text is stored in --MODIFY. In most cases (other than PE), the dictionary coded text is derived into --DECOD. Since the PEORRES variable is modified instead of the topic variable for PE, the dictionaryderived text would be placed in PESTRESC. The items used in each of the defined domains are: AE: DS: CM: MH: PE:
Page 24 August 26, 2005

AETERM/AEMODIFY/AEDECOD DSTERM/DSDECOD CMTRT/CMMODIFY/CMDECOD MHTERM/MHMODIFY/MHDECOD PEORRES/PEMODIFY/PESTRESC

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

4.1.3.7

Variables where the response is 'Yes' or 'No' ('Y'/'N') should normally be populated for both 'Y' and 'N' responses. This eliminates confusion regarding whether a blank response indicates 'N' or is a missing value. However, some variables are collected or derived in a manner that allows only one response, such as when a single check box indicates 'Yes'. In situations such as these, where it is unambiguous to only populate the response of interest, it is permissible to only populate one value ('Y' or 'N') and leave the alternate value blank. An example of when it would be acceptable to use only a value of 'Y' would be for Baseline Flag (BLFL) variables, where 'N' is not necessary to indicate that a value is not a baseline value. Note: Permissible values for variables with controlled terms of 'Y' or 'N' may be extended to include 'U' if it is the sponsors practice to explicitly collect or derive values indicating 'Unknown' for that variable.

4.1.4
4.1.4.1

Actual and Relative Time Assumptions

Date/Time formats for --DTC Variables: The CDISC SDS V2 models included date variables and some separate time variables. In Version 3.0, to address the FDAs request to provide a uniform date/time representation, all date and time variables were replaced with a single date/time variable ending in 'DTM'. These date and time values were stored as SAS date/times, which in SAS are stored as the number of seconds since midnight on January 1, 1960. A SAS format was required to display the date/times in human readable form. This was changed in version 3.1. Dates and time of day are now stored according to the international standard ISO 8601 (ISO, URL: http://www.iso.ch/iso/en/ISOOnline.openerpage). The SDTMIG template uses ISO 8601 for calendar dates and times of day, which are expressed as follows: YYYY-MM-DDThh:mm:ss where: [YYYY] = four-digit year [MM] = two-digit representation of the month (01-12, 01=January, etc.) [DD] = two-digit day of the month (01 through 31) [T] = (time designator) indicates time information follows [hh] = two digits of hour (00 through 23) (am/pm is NOT allowed) [mm] = two digits of minute (00 through 59) [ss] = two digits of second (00 through 59) Other characters defined for use within the ISO 8601 standard are: [-] (hyphen): to separate the time elements "year" and "month", "month" and "day" [:] (colon): to separate the time elements "hour" and "minute", and "minute" and "second" [/] (solidus): to separate components in the representation of time intervals [P] (duration designator): precedes the components that represent the duration Key aspects of the IS0 8601 standard are as follows: ISO 8601 represents dates as a text string using the notation YYYY-MM-DD. ISO 8601 represents times as a text string using the notation hh:mm:ss.
Page 25 July 28, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) ISO 8601 does not require dashes to separate date parts nor colons to separate time parts; however, these are required for the SDTM and V3.x SDTMIG to improve readability. When a date is stored with a time in the same variable (as a date-time), the date is written in front of the time and the time is preceded with 'T' using the notation YYYY-MM-DDT hh:mm:ss (e.g. 2001-12-26T00:00:01).

Implementation of the ISO standard means that nominal date/time variables are no longer numeric data types but will now have to be character/text data types. The SDS fragment employed for date/time variables is --DTC, where '--' is the two letter domain code [with or without the ST or EN strings indicating start or stop dates]. 4.1.4.2 Date/Time Precision. In V3.0, a separate date/time precision variable ending in 'DTP' was paired with each date/time variable to address the FDA desire to easily detect any value that was partially or fully imputed. This has been eliminated in V3.x. According to ISO 8601, precision (also referred to by ISO 8601 as "completeness" or "representations with reduced accuracy"), is specified by the number of parts present in the date, time, or date-time value. Whatever is unknown is not included so if the date is completely missing then the date value should be null. For example, if month is unknown, then only year is included in the value string. If day is unknown, only year and month are included in the value string. If seconds are unknown, all parts of the date-time string are included except the last two numbers representing seconds. If day is known but month is unknown then neither day nor month is included in the date representation. Furthermore, one-digit numbers are always padded with a leading zero. Hence, the use of a separate precision variable, as defined in Version 3.0, is no longer required. The table below provides examples of ISO 8601 representation complete date and truncated date/time values using ISO 8601 "appropriate right truncations" of incomplete datetime representations. Date and Time as Originally Recorded December 15, 2003 13:14:17 December 15, 2003 13:14 December 15, 2003 13 December 15, 2003 December, 2003 2003 Interval of Uncertainty Complete date and time Unknown seconds Unknown minutes Unknown time Unknown day Unknown month Nominal Date/Time (--DTC) 2003-12-15T13:14:17 2003-12-15T13:14 2003-12-15T13 2003-12-15 2003-12 2003

1 2 3 4 5 6

This date and date/time model provides for nonstandard intervals of uncertainty, such as those commonly seen in Medical History. To represent these intervals while applying the ISO8601 standard, it is recommended that the sponsor concatenate the date/time values (using the most complete representation of the date/time known) that describe the beginning and the end of the intervals of uncertainty and separate them with a solidus as shown in the table below: Interval of Uncertainty 1 2 3 4 Between 10:00 and 10:30 on the Morning of December 15, 2003 After the first of this year (2003) until "now" (February 15, 2003, noon) Between the first and the tenth of December, 2003 Sometime in the first half of 2003 Nominal Date/Time (--DTC) 2003-12-15T10:00/2003-12-15T10:30 2003-01-01/2003-02-15T12:00 2003-12-01/2003-12-10 2003-01-01/2003-06-30

Page 26 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) Other uncertainty intervals may be represented by the omission of components of the date when these components are unknown. Examples of this method of omitted component representation are represented in the table below: Date and Time as Originally Recorded 1 2 3 4 5 December 15, 2003 13:15:17 December 15, 2003 __:15 December 15, 2003 13:__:17 The 15th of some month in 2003 December 15, but can't remember the year Interval of Uncertainty Complete date and time Unknown hour with known minutes Unknown minutes with known date, hours and seconds Unknown month and time with known year and day Unknown year with known month and day Nominal Date/Time (--DTC) 2003-12-15T13:15:17 2003-12-15T15T-:15 2003-12-15T13:-:17 2003---15 --12-15

Using a character-based data type to implement the ISO 8601 date/time standard will ensure that the date/time information will be machine and human readable without the need for further manipulation, and will be platform and software independent. NOTE FOR BACKWARD COMPATIBILITY WITH SDS V. 3.0: Because the --DT fragment was used in the SDS Version 2 standards, and the DTM fragment was used to represent numeric based date/time values in Version 3, the DTC fragment has been used to identify character-based date/time information in Version 3.1. This will allow sponsors to continue to use the --DT and --DTM fragments for variables used in operational databases or analysis files. The SAS format IS8601DT can be used to display SAS date/time information in the ISO 8601 format. Examples are shown below: Date and time as SAS Numeric Formatted SAS originally recorded Date/Time (--DTM)1 Date/Time2 1 29June, 1956 11:32:09 -110636871 1956-06-29T11:32:09 2 29June, 1956 11:32:00 -110636880 1956-06-29T11:32:00 3 29June, 1956 11:00:00 -110638800 1956-06-29T11:00:00 4 29June, 1956 00:00:00 -110678400 1956-06-29T00:00:00 3 5 June1956 Not calculated Not calculated3 1 DTM = input(DTC, datetime18.) 2 format of --DTM IS8601DT 3 SAS requires a full, valid date. Apply a standard convention to impute days. 4.1.4.3 Duration is frequently used during a review; however, the duration timing variable (--DUR) should generally be used in a domain if it was collected in lieu of a start date/time (--STDTC) and end date/time (--ENDTC). If both --STDTC and --ENDTC are collected, durations can be calculated by the difference in these two values, and need not be in the submission dataset. In V3.0, there were separate duration (--DUR) and duration unit (--DURU) variables. In V3.x, the --DURU variable is eliminated, as both duration and duration units can be provided in the single --DUR variable according to ISO 8601. The values provided in --DUR should follow one of the following ISO 8601 formats: PnYnMnDTnHnMnS or PnW The letter "P" precedes other values in the duration format. The 'n' preceding each letter represents the number of Years, Months, Days, Hours, Minutes, Seconds, or Weeks. As with the date/time format,
CDISC, 2005. All rights reserved FINAL Page 27 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1) 'T' is used to separate the date and time components. Note, weeks cannot be mixed with days in representations When duration is being measured after an event, the correct usage when you know the start date/time from which the duration is measured is: YYYY-MM-DDThh:mm:ss/PnYnMnDTnHnMnS For duration measured prior to an event (like 2 hours before last dose) the syntax is: PnYnMnDTnHnMnS/YYYY-MM-DDThh:mm:ss Note only the hyphen delimited extended format conforming to "Format with time-unit designators" (ISO 5.5.4.2.1 - second edition 2000-12-15) is considered valid for SDTM use. Usage Notes: o o o There should be no spaces in the representation. The 'T' designator must always precede Time components, but is not necessary if no time components are included. Missing components can be omitted.
--DUR Value P2Y P10W P3M14D P3D P6M17DT3H P14DT7H57M PT42M18S -PT2H

The table below provides some examples of ISO 8601 compliant durations:
Duration as originally recorded 2 Years 10 weeks 3 Months 14 days 3 Days 6 Months 17 Days 3 Hours 14 Days 7 Hours 57 Minutes 42 Minutes 18 Seconds 2 hours before a reference time point in RFTDTC

4.1.4.4

The Study Day variable (--DY) describes the relative day of the observation starting with the reference date as Day 1. It is determined by comparing the date portion of any DTC variable to the date portion of the Subject Reference Date (RFSTDTC from the Demography domain). It is considered a descriptive representation of a relative date within the study. The subject reference date is designated as Study Day 1. The Study Day value is incremented by 1 for each date following RFSTDTC. Dates prior to RFSTDTC are decremented by 1, with the date preceding RFSTDTC designated as Study Day -1 (there is no Study Day 0). This algorithm for determining Study Day is consistent with how people typically describe sequential days relative to a fixed reference point, but creates problems if used for mathematical calculations because it does not allow for a Day 0. As such, Study Day is not suited for use in subsequent numerical computations, such as calculating duration. The raw date values should be used rather than Study Day in those calculations. All Study Day values are integers. Thus, to calculate Study Day: --DY = (date portion of --DTC) - (date portion of RFSTDTC) + 1 if --DTC is on or after RFSTDTC --DY = (date portion of --DTC) - (date portion of RFSTDTC) if --DTC precedes RFSTDTC The algorithm for this calculation should be consistent across all domains.

4.1.4.5 Clinical encounters are described by the CDISC Visit variables. VISITNUM indicates the clinical encounter number, a numeric version of VISIT used for sorting. VISITDY indicates the planned study day of VISIT. A text description of the visit (VISIT) is recommended for intelligibility and consistency with the protocol and CRF.
Page 28 August 26, 2005 CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

VISITNUM is required when data are collected more than once per subject or at a discrete time point (e.g., Labs, ECG, Vital Signs or other domains with multiple assessment points). VISITNUM is not included in Subject Characteristics because the data in this domain is collected only once per subject. It is permissible in Disposition if more than one disposition event is collected per subject. VISITNUM is not required in Adverse Events, Concomitant Medication, or Medical History. VISIT and VISITDY are permissible, but not required when VISITNUM is used. The following table shows an example of how the visit identifiers might be used: USUBJID 001 001 001 4.1.4.6 VISIT Week 1 Week 2 Week 2 Unscheduled VISITNUM 2 3 3.1 VISITDY 7 14 (Null) --DY 7 13 17

The calculation of study days within subdivisions of time in a clinical trial may be based on one or more sponsor-defined reference dates not adequately represented by RFSTDTC. The Sponsors Define data definition document should reflect the reference dates used to calculate such study days. CDISC will define an appropriate model for representing such sponsor-defined dates in a future version of this document. The CDISC V2/V3 domain models included the timing variables --ONGO and --PRIOR, but the meaning of these variables was found to differ dramatically among different sponsors, and even across different trials for the same sponsor. For example, AEONGO often meant an event was ongoing as of the end of a study, while MHONGO meant a condition was ongoing at the beginning of the study. The meaning of these variables was further confused because it was not always apparent as to which time point to use to define the beginning or end of a study. To clarify this, V3.1 of the SDTMIG defined two new timing variables, --STRF, Start Relative to Reference Period, and -ENRF, End Relative to Reference Period. --STRF identifies the start of an observation (Event, Intervention, or Finding) as being 'BEFORE', 'DURING', or 'AFTER' the sponsor-defined reference period. The sponsor-defined reference period is a continuous period of time defined by a discrete starting point (RFSTDTC) and a discrete ending point (RFENDTC). --STRF should be populated for observations with information such as check boxes on the CRF (e.g., 'PRIOR', 'ONGOING', or 'CONTINUING') that indicates when they occur in reference to the study timeline. --STRF should be set to U if its value is unknown. --ENRF identifies the end of the observation (Event, Intervention, or Finding) as being 'BEFORE', 'DURING', DURING/AFTER' or 'AFTER' the sponsor-defined reference period. The sponsor-defined reference period is a continuous period of time defined by a discrete starting point (RFSTDTC) and a discrete ending point (RFENDTC). --ENRF should be populated for observations with information such as check boxes on the CRF (e.g. 'PRIOR', 'ONGOING', or 'CONTINUING') that indicate when they occur in reference to the study timeline. --ENRF should be set to U if its value is unknown. Figure 4.1.4.7 below illustrates how to populate these variables in a CM domain. For both variables, sponsors should define the reference period in the study metadata.

4.1.4.7

CDISC, 2005. All rights reserved FINAL

Page 29 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Figure 4.1.4.7 Example of CMSTRF and CMENRF Variables

4.1.4.8

When the date and time of collection is reported in a domain based on the Findings general class, the date should go into the --DTC field -- e.g., EGDTC (Date/Time of ECG). When an interval (start and end) date/time is reported, the start date and time goes into --DTC and the end date and time goes into --ENDTC. For example, in the LB domain the LBDTC date variable is used for all single point collections or spot urine collections. For timed lab collections (e.g., 24 hour urine collections) the LBDTC variable is used for the start date/time of the collection and LBENDTC for the end date/time of the collection. This approach will allow the single point and interval collections to use the same date/time variables consistently across all datasets for the findings general class. The table below illustrates the proper use of these variables. Note --STDTC is not used for collection dates, so is blank in the following table. Date Collected Interval Collected --DTC X X --STDTC --ENDTC X

4.1.4.9

Dates are generally used only as timing variables to describe the timing of an event, intervention or collection activity, but there may be occasions when it may be preferable to model a date as a result in a Findings dataset. Note that using a date as a result to a findings question is unusual and atypical, and should be approached with caution, but this situation may occasionally occur when a) a group of questions (each of which has a date response) is asked and analyzed together; or b) the event(s) and intervention(s) in question are not medically significant (often the case when included in questionnaires). Consider the following cases: Calculated due date? Date of last day on the job? Date of birth of youngest child? End date of last reporting period?

Page 30 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) One approach to modeling these data would be to place the text of the question in --TEST and the response to the question, a date represented in ISO 8601 format, in --ORRES and --STRESC as long as these date results do not contain the dates of medically significant events or interventions. Again, use extreme caution when storing dates as the results of findings. Remember, in most cases, these dates should be timing variables associated with a record in an Intervention or Events dataset.

4.1.5
4.1.5.1

Other Assumptions
Original and Standardized Results of Findings The --ORRES variable contains the result of the measurement or finding as originally received or collected. --ORRES is an expected variable, and should always be populated with two exceptions: When --STAT = 'NOT DONE' When a record is derived (e.g., to represent an average or sum of collected values), in such a way as to make it impossible to meaningfully populate --ORRES.

Derived records are flagged with the --DRVFL variable. When the derived record comes from more than one visit, the sponsor must define the value for VISITNUM, addressing the correct temporal sequence. When --ORRES is populated, --STRESC must also be populated, regardless of whether the data values are character or numeric.--STRESC is derived either by the conversion of values in --ORRES to values with standard units, or by the assignment of the value of --ORRES (as in the PE Domain, where --STRESC could contain a dictionary-derived term). A further step is necessary when --STRESC contains numeric values. These are converted to numeric type and written to --STRESN. Because --STRESC may contain a mixture of numeric and character values, --STRESN may contain null values, as shown in the flowchart below. ORRES (all original values) STRESC (derive or copy all results) STRESN (copy numeric results only)

When the original measurement or finding is a selection from a defined code list, the --ORRES or --STRESC variables should contain results in decoded format, that is, the textual interpretation of whichever code was selected from the code list. Occasionally data that are intended to be numeric are collected with characters attached that cause the character-to-numeric conversion to fail. For example, numeric cell counts in the source data may be specified with a greater than (>) or less than (<) sign attached (e.g. >10,000 or <1). In order for these values to be reported as numeric in --STRESN, and assuming this is desirable, the sponsor must apply a rule to derive a numeric value for these strings and document this in the value-level metadata. For example the rule may be to remove the character symbol or to compensate for the symbol in some other way, such as adding to values using > or subtracting from values using <. Tests Not Done When an entire examination (Laboratory draw, ECG, Vital Signs or Physical Examination), or a group of tests (hematology or urinalysis), or an individual test (glucose, PR interval, blood pressure, or hearing) is not done, and this information is explicitly captured with a yes/no or done/not done question on the CRF, a record could be created in the dataset to record this information. See the examples below. If the data on the CRF is missing and yes/no or done/not done was not explicitly captured a record should not be created to indicate that the data was not collected. If a group of tests were not done --TESTCD should be --ALL --TEST should be <Name of the Module> --CAT should be <Name of Group of Tests> --ORRES should be NULL --STAT should be "NOT DONE"
Page 31 July 28, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

For example, if urinalysis is not done then LBTESTCD should be "LBALL" LBTEST should be "LAB" LBCAT should be "URINALYSIS" LBORRES should be NULL LBSTAT should be "NOT DONE.

Laboratory Examples: Numeric values that have been converted or copied (Rows 1 & 3). A character result that has been copied (Row 2). A result of 'TRACE' converted to a numeric value of 1 (Row 4). Value of 1+ converted to numeric 1 (Row 5). A result of 'BLQ' converted to a numeric value of zero (Row 6). Some original character results in LBORRES such as '<10,000', which are converted to numeric results per sponsor-defined algorithms (Row 7). A result is missing because the observation was 'NOT DONE', as reflected in the --STAT variable; neither LBORRES nor LBSTRESC are populated (Row 8). A result is derived from multiple records such as an average of baseline measurements for a baseline value (Row 9). None of the scheduled tests were completed as planned (Row 10) A group of tests were not completed as planned (Row 11)
Row 1 2 3 4 5 6 7 8 9 10 11 LBTESTCD GLUCOSE UBAC SGPT URBC UWBC KETONE WBC HCT MCHC LBALL LBALL LBSTAT LBCAT LBORRES 6.0 MODERATE 12.1 TRACE 1+ BLQ <10,000 LBORRESU mg/dL mg/L LBSTRESC LBSTRESN LBSTRESU LBDRVFL 60.0 60.0 mg/L MODERATE 12.1 12.1 mg/L 1 1 1 1 0 0 [sponsor[sponsordefined value] defined value] 33.8 NOT DONE NOT DONE HEMATOLOGY 33.8 g/dL Y

NOT DONE

ECG Examples: Numeric and character values that have converted or copied (Rows 1, 2, and 3). A result is missing because the test was 'NOT DONE', as reflected in the EGSTAT variable; neither EGORRES nor EGSTRESC are populated (Row 4). The overall interpretation is included as a new record (Row 5) The entire ECG was not done (Row 6)
Row 1 2 3 4 5 6 EGTESTCD QRS QT RHYMRATE PR INTP EGALL EGSTAT EGORRES 362 221 ATRIAL FLUTTER ABNORMAL, CLINICALLY SIGNIFICANT NOT DONE EGORRESU SEC MSEC EGSTRESC 362 .221 ATRIAL FLUTTER ABNORMAL, CLINICALLY SIGNIFICANT EGSTRESN 362 .221 EGSTRESU SEC SEC

NOT DONE

Page 32 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) Vital Signs Examples: Numeric and character values that have converted or copied (Rows 1 and 2). A result is missing because the Vital Signs test was 'NOT DONE', as reflected in the VSSTAT variable; neither VSORRES nor VSSTRESC are populated (Row 3). The result is derived by having multiple records for one measurement and the derived value is recorded in a new row with the derived record flagged. (Rows 4, 5, and 6). The entire examination was not done (Row 7) Row 1 2 3 4 5 6 7
VSTESTCD VSSTAT VSORRES VSORRESU VSSTRESC VSSTRESN VSSTRESU VSDRVFL

HT WT HR SYSBP SYSBP SYSBP VSALL

60 110 NOT DONE 96 100 NOT DONE

IN LB mmHg mmHg

152 50 96 100 98

152 50 96 100 98

cm kg mmHg mmHg mmHg

Questionnaire Examples: Character values that have converted to standard scores (Row1). A result is derived from multiple records (Row 2) A result is missing because the observation was 'NOT DONE', as reflected in the QSSTAT variable; neither QSORRES nor QSSTRESC are populated (Row 3). The entire questionnaire was not done (Row 4) Row 1 2 3 4
QSTEST QSSTAT QSORRES QSSTRESC QSSTRESN QSDRVFL

HEALTH HEALTH PERCEPTIONS (0-100) HEALTH QSALL

VERY GOOD NOT DONE NOT DONE

4.4

4.4 82 Y

4.1.5.2 Multiple observations may be linked (associated with each other) using the RELREC dataset described in Section 8 of this document. Comments should be stored in the Comments domain (CO). Additional Qualifier variables that were not specified in the model should be included in the Supplemental Qualifiers (SUPPQUAL) dataset. 4.1.5.3 CDISC realizes that some sponsors may have text strings longer than 200 characters for some variables. Because of the current requirement for Version 5 SAS transport file format, it will not be possible to store those long text strings using only one variable. Therefore, CDISC has defined a convention for storing a long text string by using a combination of the standard domain dataset and the Supplemental Qualifiers (SUPPQUAL) dataset. For a text string of more than 200 characters in length, the first 200 characters of text should be stored in the standard domain variable and each additional 200 characters of text should be stored as a record in the SUPPQUAL dataset. In this dataset, the value for QNAM should contain a sequential variable name, which is formed by appending a one-digit integer, beginning with 1, to the original standard domain variable name. When splitting a text string into several records, the text should be split between words to improve readability. As an example, if a sponsor had text for Medical History Reported Term (MHTERM) for one subject of 500 characters in length, the sponsor would put the first 200 characters of text in the standard domain variable and dataset (MHTERM in MH), the next 200 characters of text as a first supplemental record in the SUPPQUAL dataset, and the final 100 characters of text as a second record in the SUPPQUAL dataset. Variable QNAM would have the values MHTERM1 and MHTERM2 for the first and second records in SUPPQUAL, respectively, for this one particular text string. Sponsors should place the text itself into variable QVAL and the label of the original standard domain variable into variable QLABEL. In this case, IDVAR and IDVARVAL should be used in SUPPQUAL to relate the associated supplemental text records to the first 200 characters of text in the standard domain.
CDISC, 2005. All rights reserved FINAL Page 33 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1) In cases where the standard domain variable name is already 8 characters in length, sponsors should replace the last character with a digit when creating values for QNAM. As an example, for Other Action Taken in Adverse Events (AEACNOTH), values for QNAM for the SUPPQUAL records would have the values AEACNOT1, AEACNOT2, and so on. Examples of how SUPPQUAL would be populated are presented below:
STUDYID RDOMAIN USUBJID 12345 12345 MH MH 99-123 99-123 IDVAR IDVARVAL MHSEQ MHSEQ 6 6 QNAM MHTERM1 MHTERM2 QLABEL QVAL QORIG QEVAL CRF CRF Reported Term 2nd 200 for Condition or chars of text Procedure Reported Term last 200 for Condition or chars of text Procedure Other Action Taken last 200 chars of text

12345

AE

99-123

AESEQ

AEACNOT1

CRF

4.1.5.4 For the Interventions and Events observation classes, all data are assumed to be attributed to the Principal Investigator, or derived from such data by the sponsor. The observations recorded in the Findings class include the --EVAL qualifier because the observation may originate from more than one source (e.g., an Investigator or Central Reviewer). The QEVAL variable can be used to describe the evaluator for any data item in SUPPQUAL, but is not required when the data is objective. For all observations that have a primary and supplemental evaluation, sponsors should always put data from the primary evaluation into the standard domain dataset and data from the supplemental evaluation into the Supplemental Qualifiers dataset (SUPPQUAL see Section 8.4). In SUPPQUAL, each variable for the evaluation should be represented as a record. Within each record, the value for QNAM should be formed by appending a 1 to the corresponding standard domain variable name. In cases where the standard domain variable name is already 8 characters in length, sponsors should replace the last character with a 1 (incremented for each additional attribution). The result of, response to, or value associated with QNAM should be placed in variable QVAL. The label of the corresponding standard domain variable should be placed in variable QLABEL. As an example, if an adjudication committee evaluates an adverse event, data in SUPPQUAL for this evaluation would look as below. (Note that QNAM takes on the value AERELNS1, as the corresponding standard domain variable AERELNST is already 8 characters in length.) The adverse event data as determined by the primary investigator would reside in the standard AE dataset.
STUDYID RDOMAIN USUBJID IDVAR IDVARVAL 12345 12345 12345 AE AE AE 99-123 99-123 99-123 AESEQ AESEQ AESEQ 3 3 3 QNAM AESEV1 AEREL1 QLABEL Severity/ Intensity Causality QVAL MILD POSSIBLY RELATED Possibly related to aspirin use. QORIG CRF CRF CRF QEVAL ADJUDICATION COMMITTEE ADJUDICATION COMMITTEE ADJUDICATION COMMITTEE

AERELNS1 Relationship to Non-Study Treatment

Page 34 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

5 Models for Special Purpose Domains

5.1.1 Demographics Domain Model DM
dm.xpt, Demographics Version 3.1.1, August 26, 2005. One record per subject, Tabulation
Variable Name STUDYID DOMAIN USUBJID Variable Label Study Identifier Domain Abbreviation Unique Subject Identifier Type Controlled Terms or Format Origin CRF Derived Sponsor Defined Role CDISC Notes Core References SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4

Char Char **DM Char

SUBJID RFSTDTC

Subject Identifier for the Study Subject Reference Start Date/Time Subject Reference End Date/Time

Char Char ISO 8601

CRF Sponsor Defined Sponsor Defined

RFENDTC

Char ISO 8601

SITEID INVID INVNAM BRTHDTC

Study Site Identifier Investigator Identifier Investigator Name Date/Time of Birth

Char Char * Char Char ISO 8601

CRF or Derived CRF or Derived CRF or Derived CRF or Derived

Identifier Unique identifier for a study within the submission. Req Identifier Two-character abbreviation for the domain most relevant to the Req observation. Identifier Unique subject identifier within the submission. This can be a Req unique number or a compound identifier formed by concatenating STUDYID-SITEID-SUBJID. Since a subject may participate in more than one study (as in a follow-on), the USUBJID may differ from the SUBJID in some cases. Topic Subject identifier used within the study. Often the ID of the subject Req as recorded on a CRF. Timing Reference Start Date/time for the subject in ISO 8601 character Exp format. Usually equivalent to date/time of first intake of drug. Required for all randomized subjects; null for screen failures (if screen failures are submitted). Timing Reference End Date/time for the subject in ISO 8601 character Exp format. Usually equivalent to the date/time when subject was determined to have ended the trial, and often equivalent to date/time of last intake of drug. Required for all randomized subjects; null for screen failures (if screen failures are submitted). Record Unique identifier for a study site within a submission. Req Qualifier Record An identifier to describe the Investigator for the study. May be Perm Qualifier used in addition to the SITEID. Not needed if SITEID is equivalent to INVID. Synonym Name of the investigator for a site. Perm Qualifier Result Date/time of birth of the subject. Qualifier Perm

SDTMIG 4.1.4.1 SDTMIG 4.1.4.1

SDTMIG 4.1.4.1

CDISC, 2005. All rights reserved FINAL

Page 35 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name AGE AGEU

Variable Label

Type

Controlled Terms or Format

Origin CRF or Derived

Role

CDISC Notes

Core

References

Age in AGEU at RFSTDTC Num Age Units

Result Usually derived as (RFSTDTC -BRTHDTC), but BRTHDTC may Exp Qualifier not be available in all cases (due to subject privacy concerns). Variable Qualifier Result M, F, U for Male, Female, Unknown Qualifier Exp

SEX RACE

Sex Race

Char ** YEARS, CRF or MONTHS, Derived DAYS, HOURS, WEEKS Char **M, F, U CRF Char ** CRF

Req

ETHNIC ARMCD ARM COUNTRY DMDTC DMDY

Ethnicity Planned Arm Code

Char ** Char *

CRF CRF or Derived CRF or Derived CRF or Derived CRF or Derived Derived

Result Race of the subject. In cases of mixed race, primary race should be Exp Qualifier listed. If additional race values are collected, they should be represented as separate records in the Subject Characteristics domain with TESTCD of RACEOTH. The variable itself is required by FDA, but with ongoing discussions in Europe about the permissibility of collecting this data, the variable may be optional in future. Result Ethnicity of subject. Perm Qualifier Result Short version of ARM, used for sorting and programming Req Qualifier (Formerly TRTCD). Synonym Name given to the arm the subject was assigned to (Formerly Qualifier TRTGRP). Req

Description of Planned Arm Char * Country Date/Time of Collection Study Day of Collection Char **ISO 3166 3char. code Char ISO 8601 Num

Result Country of the investigational site in which the subject participated Req Qualifier in the trial. Timing Timing Use if collected. 1. Study day of collection of the demographic information, measured as integer days. 2. Algorithm for calculations must be relative to the sponsordefined RFSTDTC variable in Demographics. This formula should be consistent across the submission. Perm Perm SDTMIG 4.1.4.1 SDTMIG 4.1.4.4

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.

Page 36 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

5.1.1.1 Assumptions for Demographics (DM) domain model 1. Investigator (INVID) and/or site (SITEID) identification: Companies use different methods to distinguish sites and/or investigators. CDISC assumes that site will always be present, with investigator used as necessary to group data. This should be done consistently for the submission and the meaning of the variable made clear in the Define data definition document. Investigator name (INVNAM) may also be included as a support variable. 2. Subject identification: It is presumed that every subject in a study will have a subject identifier (SUBJID) and that in some cases a subject may be included in more than one study within a submission. To identify a subject uniquely across a submission, a unique identifier (USUBJID) should be assigned and included in all datasets within the submission. Recent ongoing discussions regarding subject privacy suggest that care must be taken with regard to the collection of variables like BRTHDTC. This variable is included in the Demographics model as an example of a variable that might be included in a submission; however, sponsors should follow regulatory guidelines and guidance as appropriate. Arm/Treatment identification: Note that in Version 3.x the variables ARM and ARMCD have replaced TRTGRP and TRTCD for consistency with the trial design, and to more clearly distinguish the planned arm from the actual treatment. When a sponsor is submitting trial design information, the values of ARM and ARMCD should be identical to the values defined for that subject in the Subject Elements (SE) dataset. The assignment of values should be consistent, if possible, within a submission. Justification for using SEX vs. GENDER: Page 71 of 'Providing Regulatory Submissions in Electronic Format - NDAs' (IT-3, January, 1999), available at http://www.fda.gov/cder/guidance/2353fnl.pdf specifically lists SEX as part of demographic data. Similarly, page 60 of 'Guidance for Industry, Providing Regulatory Submissions to the Center for Biologics Evaluation in Electronic Format - Biologics Marketing Applications' (November, 1999), available at http://www.fda.gov/cber/guidelines.htm specifically lists SEX as part of demographic data. SEX is used consistently in both documents except for one instance where GENDER is used (page 30 for Table 6 which may have been from another writer). 'ICH E3: Structure and Content of Clinical Study Reports' (November 30, 1999) only uses SEX (not GENDER). Attributions used to classify study populations for review and analysis, such as the COMPLT, SAFETY, ITT and PPROT variables discussed in the V3 DM domain, should be placed in the suppdm.xpt SUPPQUAL dataset as described in Section 8.4. When additional free text information is reported about subject's RACE using 'Other, Specify', sponsors should place the free text value in the Subject Characteristics (SC) dataset with TESTCD of RACEOTH as described in the CDISC note for RACE. Data for screen failure subjects, if submitted, should be included in the Demographics dataset, with ARMCD = SCRNFAIL" and ARM = Screen Failure to distinguish screen failures from randomized subjects. Sponsors may include a record in the Disposition dataset indicating when the screen dropout event occurred. See section "9.3.2 DS Examples" for examples of screen dropout records.

3.

4.

5.

6. 7. 8.

CDISC, 2005. All rights reserved FINAL

Page 37 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

5.1.2

Comments Domain Model CO

The COMMENTS special-purpose domain is a fixed domain that provides a solution for submitting free-text comments related to data in one or more SDS domains or collected on a separate CRF page dedicated to comments. Comments are generally not responses to specific questions; instead, comments usually consist of voluntary, free-text, unsolicited observations. COMMENTS has a structure that is similar to the Supplemental Qualifiers (SUPPQUAL) specialpurpose dataset but it allows for one comment to span multiple variables (COVAL-COVALn) in order to accommodate comments longer than 200 characters.1 Comments may be related to a Subject or to specific parent records in an SDTM domain. The approach for relating comments is described in Section 8.5. co.xpt -- Version 3.1.1. August 26, 2005. One record per comment per subject, Tabulation
Variable Name Variable Label Type Char Char Char Controlled Terms or Format **CO ** Origin CRF Derived Derived Sponsor Defined CRF or Derived ** Derived Derived * CRF Role CDISC Notes Core References SDTM 2.2.4 SDTM 2.2.4

STUDYID Study Identifier DOMAIN Domain Abbreviation RDOMAIN Related Domain Abbreviation USUBJID COSEQ IDVAR

Identifier Unique identifier for a study within the submission. Req Identifier Two-character abbreviation for the domain most relevant to the observation. Req Identifier Domain abbreviation of the parent record(s). Null for comments collected on a Exp general comments or additional information CRF page. Identifier Unique subject identifier within the submission. Req

Unique Subject Identifier Char Sequence Number Identifying Variable Num Char

SDTM 2.2.4 .

Identifier Sequence Number given to ensure uniqueness within a domain. Can be used to Req join related records Identifier Identifying variable in the dataset that identifies the related record(s). Perm Examples --SEQ, or --GRPID. Null for comments collected on separate CRFs. Identifier Value of identifying variable of the parent record(s). Used only when Perm individual comments are related to domain records. Null for comments collected on separate CRFs Record CRF page of parent record(s) to which the comment refers, or on which the Perm Qualifier comment was collected. COREF may be the page number, e.g. 650, or the page name e.g. DEMOG, or a combination of information that uniquely identifies the reference both e.g. 650-VITALS- VISIT 2. Timing Perm Result The text of the comment. COVAL cannot be null a value is required for the Req Qualifier record to be valid. The first 200 characters of the comment will be in COVAL, then next 200 in COVAL1, and continuing as needed to COVALn. Record Used to describe the originator of the comment. Controlled terminology will Perm Qualifier consist of values such as ADJUDICATION COMMITTEE, STATISTICIAN, DATABASE ADMINISTRATOR, CLINICAL COORDINATOR, and PRINCIPAL INVESTIGATOR.

IDVARVAL Identifying Variable Value Char COREF Comment Reference Char

CODTC COVAL

Date/Time of Comment Comment

Char Char

ISO 8601

CRF CRF

SDTMIG 4.1.4.1

COEVAL

Evaluator

Char

CRF or Derived

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.
1

Allowing for multiple variables per comment is an interim solution until the limitations posed by SAS Version 5 transport files are eliminated.
CDISC, 2005. All rights reserved FINAL

Page 38 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

5.1.2.1 1.

Assumptions for Comments (CO) domain model The COMMENTS dataset accommodates two sources of comments: 1) those collected alongside other data on topical CRF pages such as Adverse Events; and 2) those collected on a separate page specifically dedicated to comments. The value of the variable RDOMAIN for comments of the first type should be the domain code of the parent record(s). The value of the variable RDOMAIN of the second type should be null. Assumptions for the use of GRPID and SEQ as values for IDVAR are the same as for the Supplemental Qualifiers (SUPPQUAL) special purpose dataset described in Section 8, except that when there is no parent record (i.e., for comments collected on a dedicated comments CRF) IDVAR and IDVARVAL should be NULL. COMMENTS records without a parent record (i.e. those collected on a dedicated Comments CRF) should have CODTC filled if it is collected on the CRF. COMMENTS records with a parent record inherit timing variables from the parent, thus CODTC is NULL in COMMENTS. The variable COREF may be NULL unless it is used to identify the source of the comment. Example of use of COMMENTS special-purpose dataset --Subject with four comments: One comment collected on each of the AE, EX, and VS pages, and another collected on a separate Comments CRF 1) Columns 3 and 6 (RDOMAIN and IDVAR) represent information from the parent record, if applicable. 2) Row 4 contains general comments from the Comments CRF Page. There is no parent record for this comment, thus DOMAIN has the value CO and RDOMAIN, IDVAR and IDVARVAL are NULL. 3) Row 1 contains a comment text > 400 characters; Rows 2 and 3 contains text between 200 and 400 characters, Row 4 contains 150 characters of comment text.
ROW STUDYID DOMAIN RDOMAIN USUBJID COSEQ IDVAR IDVARVAL COREF COVAL COVAL1 COVAL2 COEVAL

2.

1 2

1234 1234

CO CO

AE EX

AB-99 AB-99

1 2

AESEQ EXGRPID

7 17

650 320-355

First 200 characters First 200 characters First 200 characters

Next 200 characters Next 200 characters More text

More Text

PRINCIPAL INVESTIGATOR PRINCIPAL INVESTIGATOR PRINCIPAL INVESTIGATOR PRINCIPAL INVESTIGATOR

3 4

1234 1234

CO CO

VS

AB-99 AB-99

3 4

VSGRPID

5 800

Total 150 characters

CDISC, 2005. All rights reserved FINAL

Page 39 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

6 Domain Models Based on the General Classes

6.1
6.1.1

INTERVENTIONS
Concomitant Medications CM

cm.xpt, Concomitant Medications Interventions, Version 3.1.1, August 26, 2005. One record per medication intervention episode per subject, Tabulation
Variable Name STUDYID DOMAIN USUBJID CMSEQ CMGRPID Variable Label Study Identifier Domain Abbreviation Unique Subject Identifier Sequence Number Group ID Type Char Char Char Num Char **CM Controlled Terms or Format Origin CRF Derived Sponsor Defined CRF or Derived Sponsor Defined Sponsor Defined Role Identifier Identifier Identifier Identifier Identifier CDISC Notes Unique identifier for a study within the submission. Core Req References SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTMIG 2.1; SDTM 2.2.4 SDTM 2.2.4

Two-character abbreviation for the domain most relevant to Req the observation. Unique subject identifier within the submission. Sequence number given to ensure uniqueness within a dataset for a subject. Can be used to join related records. Used to tie together a block of related records in a single domain to support relationships within the domain and between domains. Req Req Perm

CMSPID

Sponsor-Defined Identifier

Char

Identifier

Optional sponsor-defined reference number. Perhaps pre- Perm printed on the CRF as an explicit line identifier or defined in the sponsors operational database. Example: line number on a concomitant medication page.

Page 40 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name CMTRT CMMODIFY CMDECOD

Variable Label Reported Name of Drug, Med, or Therapy Modified Reported Name Standardized Medication Name

Type Char Char Char

Controlled Terms or Format

Origin CRF Sponsor Defined Derived

Role Topic Synonym Qualifier Synonym Qualifier

CDISC Notes Verbatim medication name that is either pre-printed or collected on a CRF.

Core Req

References SDTMIG 4.1.3.6 SDTMIG 4.1.3.5 SDTMIG 4.1.3.5

**

CMCAT CMSCAT

Category for Medication

Char

* *

Sponsor Defined Sponsor Defined CRF or Sponsor Defined

Grouping Qualifier Grouping Qualifier Record Qualifier

Subcategory for Medication Char

If CMTRT is modified, then CMMODIFY will contain the Perm modified text. Standardized or dictionary-derived text description of Perm CMTRT or CMMODIFY. Equivalent to the generic medication name in WHO Drug. The sponsor should specify the dictionary name and version in the Sponsor Comments column of the Define document. If an intervention term does not have a decode value in the dictionary then CMDECOD will be left blank. Used to define a category of medications/treatments. Perm Examples: ANTI-CANCER MEDICATION, or GENERAL CONMED. A further categorization of medications/ treatment. Perm Examples: CHEMOTHERAPY, RADIOTHERAPY, HORMONAL THERAPY, ALTERNATIVE THERAPY. Used when the use of specific medications is solicited To Perm indicate whether a medication was taken or not. Values are null for spontaneously reported medications. Note: Use of this variable is under evaluation by the SDS Team. The status indicates that the question was not asked. Perm

SDTMIG 2.1 SDTMIG 2.1

CMOCCUR

CM Occurrence

Char

**Y, N or Null

CMSTAT

Concomitant Medication Status Reason Medication Not Collected Indication Medication Class

Char

**NOT DONE

CRF or Sponsor Defined CRF CRF or Derived

Record Qualifier Record Qualifier Record Qualifier Variable Qualifier

SDTMIG 4.1.5.1.

CMREASND CMINDC CMCLAS

Char Char Char **

Describes the reason concomitant medication was not Perm collected. Used in conjunction with CMSTAT when value is NOT DONE. Denotes why a medication was taken or administered. Perm Examples: NAUSEA, HYPERTENSION. Use only when the dictionary used codes to a single class. If Perm using a dictionary that allows links to multiple classes, then omit CMCLAS from the dataset. For example, sponsors who use WHO Drug, which allows links from a medication to multiple ATC codes, would not include CMCLAS. SDTMIG 4.1.3.5

Derived

CDISC, 2005. All rights reserved FINAL

Page 41 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name CMCLASCD CMDOSE

Variable Label Medication Class Code Dose per Administration

Type Char Num

Controlled Terms or Format **

Origin Derived CRF or Sponsor Defined CRF or Derived

Role Variable Qualifier Record Qualifier Record Qualifier Variable Qualifier Variable Qualifier Variable Qualifier Record Qualifier Variable Qualifier Variable Qualifier Timing Timing Timing Timing

CDISC Notes Use only when the dictionary used codes to a single class. Amount of CMTRT taken.

Core Perm Perm

References SDTMIG 4.1.3.5

CMDOSTXT CMDOSU

Dose Description Dose Units

Char Char *

Dosing amounts or a range of dosing information collected Perm in text form. Example: 200-400. Units for CMDOSE and CMDOSTOT. Examples: ng, mg, Perm or mg/kg. Dose form for CMTRT. Examples: TABLET, LOTION. Perm

CRF or Sponsor Defined CRF or Sponsor Defined CRF or Sponsor Defined CRF or Derived CRF or Derived

CMDOSFRM

Dose Form

Char

CMDOSFRQ

Dosing Frequency Per Interval Total Daily Dose using CMDOSU Intended Dose Regimen

Char

Usually expressed as the number of dosings given per a specific interval. Examples: BID, QID. Total daily dose of CMTRT using the units in CMDOSU. Text description of the (intended) schedule or regimen for the Intervention. Examples: TWO WEEKS ON, TWO WEEKS OFF. Route of administration for CMTRT. Examples: ORAL, INTRAVENOUS.

Perm

CMDOSTOT CMDOSRGM

Num Char

Perm Perm

CMROUTE

Route of Administration

Char

CRF or Sponsor Defined CRF CRF Derived Derived

Perm

CMSTDTC CMENDTC CMSTDY CMENDY

Start Date/Time of Medication

Char

ISO 8601 ISO 8601

Perm Perm Study day of start of medication relative to the sponsordefined RFSTDTC. Study day of end of medication relative to the sponsordefined RFSTDTC. Perm Perm

SDTMIG 4.1.4.1 SDTMIG 4.1.4.1 SDTMIG 4.1.4.1 SDTMIG 4.1.4.1

End Date/Time of Medication Char Study Day of Start of Medication Study Day of End of Medication Num Num

Page 42 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name CMDUR

Variable Label Duration of Medication

Type Char

Controlled Terms or Format ISO 8601

Origin CRF

Role Timing

CDISC Notes

Core

References

CMSTRF

Start Relative to Reference Period

Char

** BEFORE, Derived DURING, AFTER

Timing

CMENRF

End Relative to Reference Period

Char

** BEFORE, Derived DURING, AFTER, DURING/ AFTER, U

Timing

Collected duration and unit of a treatment. Used only if Perm collected on the CRF and not derived from start and end date/times. Identifies the start of the medication with respect to the Perm sponsor-defined reference period. Sponsors should define the reference period in the study metadata. CMSTRF should be populated when a start date is not collected. If information such as "PRIOR", "ONGOING", or "CONTINUING" was collected, this information should be translated into CMSTRF. Identifies the end of the medication with respect to the Perm sponsor-defined reference period. Sponsors should define the reference period in the study metadata. Medications that are ongoing at the end of the reference period should have a value of AFTER for this variable. CMENRF should be populated when an end date is not collected. If information such as "PRIOR", "ONGOING", or "CONTINUING" was collected, this information should be translated into CMENRF.

SDTMIG 4.1.4.7

SDTMIG 4.1.4.7

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.
6.1.1.1 Assumptions for Concomitant Medications (CM) domain model 1. CM Definition a. CRF data that captures the Concomitant and Prior Medications/Therapies used by the subject. Examples are the Concomitant Medications/Therapies given on an as needed basis and the usual and background medications/therapies given for a condition. b. This domain should contain one record per each Concomitant Medication/Therapy intervention episode. 2. Concomitant Medications Description and Coding a. CMTRT captures the name of the Concomitant Medications/Therapy and it is the topic variable. It is a required variable and must have a value. b. CMMODIFY should be included if the sponsors procedure permits modification of a verbatim term for coding. The modified term is listed in CMMODIFY. c. CMDECOD is the standardized medication/therapy term derived by the sponsor from the coding dictionary. It is expected that the reported term (CMTRT) or the modified term (CMMODIFY) will be coded using a standard dictionary. The sponsor is expected to provide the dictionary name and version used to map the terms in the metadata (using the Comments column in the Define document). If a sponsor chooses to use more than one dictionary to code different types of data within the same dataset (e.g., drug and non-drug therapies), then the variables CMCAT or CMSCAT can be used to categorize the data by type. The metadata will specify which dictionary is applied to each category. d. CMCLAS is used only when the dictionary used codes to a single class. If using a dictionary that allows links to multiple classes, then omit CMCLAS from the submission dataset.
Page 43 July 28, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

3.

Additional Categorization and Grouping a. CMCAT and CMSCAT may be used as categorization variables to group data across subjects. Examples of CMCAT could include values such as ANTI-CANCER MEDICATION, or GENERAL CONMED. CMSCAT may be used to further categorize the Concomitant Medications/Therapy taken by the subject. Examples: CHEMOTHERAPY, RADIOTHERAPY, HORMONAL THERAPY or ALTERNATIVE THERAPY. Neither CMCAT nor CMSCAT are expected variables in the submission dataset. b. CMGRPID is used to link a block of related records at the subject level in a single domain to support relationships within the domain and between domains. Presence or Absence of Concomitant Medications a. Information regarding Concomitant Medications/Therapy is generally collected in two different ways, either by recording free text or using a prespecified list of terms. In the latter case, the solicitation of information on specific concomitant medications may affect the frequency at which they are reported; therefore, the fact that use of a specific medication was solicited may be of interest to reviewers. Consequently, the SDS Team is proposing the addition of a new variable --PRESP to the SDTM to indicate that the use of those specific medications was solicited. Until then, CMPRESP can be used as a supplemental qualifier. b. CMOCCUR may be used when a CRF includes a pre-specified list of Concomitant Medications/Therapies to indicate whether or not a medication/therapy was used. Proper use of this variable is under evaluation by the SDS Team. ConMed Structure a. The structure of the CM domain is one record per medication intervention episode per subject. It is the sponsor's responsibility to define an intervention episode. This definition may vary based on the sponsor's requirements for review and analysis. The submission dataset structure may differ from the structure at the time of collection. One common approach is to submit a new record when there is a change in the dosing regimen. Another approach is to collapse all records for a medication to a summary level with either a dose range or the highest dose level. Other approaches, including a visit-based record, may also be reasonable as long as they meet the sponsor's evaluation requirements.

4.

5.

Page 44 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

6.1.2

Exposure EX

ex.xpt, Exposure Interventions, Version 3.1.1 August 26, 2005. One record per constant dosing interval per subject, Tabulation
Variable Name STUDYID DOMAIN USUBJID EXSEQ Variable Label Study Identifier Domain Abbreviation Unique Subject Identifier Sequence Number Type Char Char Char Num **EX Controlled Terms or Format Origin CRF Derived Sponsor Defined CRF or Derived Sponsor Defined Sponsor Defined Role Identifier Identifier Identifier Identifier CDISC Notes Unique identifier for a study within the submission. Core Req References SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4

Two-character abbreviation for the domain most relevant Req to the observation. Unique subject identifier within the submission. Req

Sequence number given to ensure uniqueness within a Req dataset for a subject. Can be used to join related records. Used to tie together a block of related records in a single Perm domain to support relationships within the domain and between domains. Sponsor-defined reference number. Perhaps pre-printed Perm on the CRF as an explicit line identifier or defined in the sponsors operational database. Example: Line number on a CRF Page. Name of the intervention treatment usually the verbatim name of the investigational treatment given during the dosing period for the observation. Used to define a category of related records. Example: COMPARATOR CLASS. A further categorization of treatment. Req

EXGRPID

Group ID

Char

Identifier

SDTMIG 2.1; SDTM 2.2.4 SDTM 2.2.4

EXSPID

Sponsor-Defined Identifier

Char

Identifier

EXTRT

Name of Actual Treatment

Char

CRF or Topic Randomization File or Derived * * Sponsor Defined Sponsor Defined CRF or Sponsor Defined Grouping Qualifier Grouping Qualifier

SDTMIG 4.1.3.6

EXCAT EXSCAT EXDOSE

Category for Treatment Subcategory for Treatment Dose per Administration

Char Char Num

Perm Perm Exp

SDTMIG 2.1 SDTMIG 2.1

Record Qualifier Amount of EXTRT administered or given.

CDISC, 2005. All rights reserved FINAL

Page 45 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name EXDOSTXT EXDOSU

Variable Label Dose Description Dose Units

Type Char Char

Controlled Terms or Format

Origin CRF or Derived

Role

CDISC Notes

Core Perm

References

Record Qualifier Dosing amounts or a range of dosing information collected in text form. Example: 200-400. Variable Qualifier Variable Qualifier Variable Qualifier

CRF or Sponsor Defined CRF or Sponsor Defined CRF or Sponsor Defined CRF or Derived CRF or Derived

Units for EXDOSE and EXDOSTOT. Examples: ng, mg, Exp or mg/kg. Dose form for EXTRT. Examples: TABLET, LOTION,. Req

EXDOSFRM

Dose Form

Char

EXDOSFRQ

Dosing Frequency Per Interval Total Daily Dose Using EXDOSU Intended Dose Regimen Route of Administration

Char

Usually expressed as the number of dosings given per a specific interval. Examples: BID, QID.

Perm

EXDOSTOT EXDOSRGM EXROUTE

Num Char Char *

Record Qualifier Total daily dose of EXTRT using the units in EXDOSU. Perm Variable Qualifier Variable Qualifier Text description of the (intended) schedule or regimen for Perm the Intervention. Examples: two weeks on, two weeks off. Route of administration for EXTRT. Examples: ORAL, Perm INTRAVENOUS. Perm

CRF or Sponsor Defined CRF

EXLOT EXLOC

Lot Number Location of Dose Administration Treatment Vehicle

Char Char *

Record Qualifier Lot Number of the EXTRT product.

CRF or Sponsor Defined CRF or Sponsor Defined CRF or Sponsor Defined Sponsor Defined / Protocol

Record Qualifier Specifies anatomical location of administration Example: Perm LEFT ARM for a topical application. Record Qualifier Describes vehicle used for treatment. Example: SALINE. Perm

EXTRTV

Char

EXADJ

Reason for Dose Adjustment Char

TAETORD

Order of Element within Arm Num

Record Qualifier Describes reason or explanation of why a dose is adjusted Perm used only when an adjustment is represented in EX. May be used for variations from protocol-specified doses, or changes from expected doses. Timing Number that gives the order of the Element within the Perm arm.

Page 46 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name EXSTDTC EXENDTC EXSTDY EXENDY EXDUR

Variable Label

Type

Controlled Terms or Format ISO 8601 ISO 8601

Origin CRF or Derived CRF or Derived Derived Derived

Role Timing Timing Timing Timing Timing

CDISC Notes The time when the dose was administered. May not be relevant for injections. Study day of start of treatment relative to the sponsordefined RFSTDTC. Study day of end of treatment relative to the sponsordefined RFSTDTC.

Core Req

References SDTMIG 4.1.4.1

Start Date/Time of Treatment Char End Date/Time of Treatment Char Study Day of Start of Treatment Study Day of End of Treatment Duration of Treatment Num Num Char

Perm SDTMIG 4.1.4.1 Perm SDTMIG 4.1.4.1 Perm SDTMIG 4.1.4.1

ISO 8601

CRF

EXTPT

Planned Time Point Name

Char

CRF or Derived

Timing

EXTPTNUM EXELTM

Planned Time Point Number Num Planned Elapsed Time from Char Reference Pt Time Point Reference Char ISO 8601

Sponsor Defined Sponsor Defined Sponsor Defined

Timing Timing

Collected duration and unit of a treatment. Used only if Perm collected on the CRF and not derived from start and end date/times. 1. Text Description of time when a dose should be given. Perm 2. This may be represented as an elapsed time relative to a fixed reference point, such as time of last dose. See EXTPTNUM and EXTPTREF. Examples: Start or 5 min post. Numerical version of EXTPT to aid in sorting. Perm Elapsed time (in ISO 8601) format relative to the planned Perm fixed reference (EXTPTREF). This variable is useful where there are repetitive measures. Not a clock time. Name of the fixed reference point referred to by EXELTM, EXTPTNUM, and EXTPT. Examples: Previous Dose, Previous Meal. Perm

EXTPTREF

Timing

indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.
Assumptions for Exposure (EX) domain model

6.1.2.1

1. EX Definition a. The Exposure domain model records the details of a subjects exposure to protocol-specified study treatment. Study treatment may be any intervention that is prospectively defined as a test material within a study, and is typically but not always supplied to the subject. Examples include but are not limited to placebo, active comparator, and investigational product. Treatments that are not protocol-specified should be recorded in the Concomitant Medication (CM) domain.

CDISC, 2005. All rights reserved FINAL

Page 47 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

b. This domain should contain one record per constant dosing interval per subject. "Constant dosing interval" is sponsor-defined, and may include any period of time that can be described in terms of a known treatment given at a consistent dose. e.g., for a study with once-a-week administration of a standard dose for 6 weeks, exposure may be represented with a single record per subject, spanning the entire treatment phase. Or if the sponsor monitors each treatment administration and deviations in treatment or dose occur, there could be up to six records (one for each weekly administration). 2. Categorization and Grouping a. EXCAT and EXSCAT may be used when appropriate to categorize treatments into categories and subcategories. For example, if a study contains several active comparator medications, EXCAT may be set to 'ACTIVE COMPARATOR.' Since such categorization may not apply to many studies, these variables are permissible but not expected.

Page 48 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

6.1.3

Substance Use - SU

su.xpt, Substance Use Interventions, Version 3.1.1, August 26, 2005. One record per substance type per visit per subject, Tabulation
Variable Name Variable Label Type Controlled Terms or Format Origin Role CDISC Notes Core References

STUDYID DOMAIN USUBJID SUSEQ SUGRPID

Study Identifier Domain Abbreviation Unique Subject Identifier Sequence Number Group ID

Char Char Char Num Char **SU

CRF Derived Sponsor Defined CRF or Derived Sponsor Defined Sponsor Defined

Identifier Identifier Identifier Identifier Identifier

Unique identifier for a study within the submission. Two-character abbreviation for the domain most relevant to the observation. Unique subject identifier within the submission.

Req Req Req

SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTMIG 2.1; SDTM 2.2.4 SDTM 2.2.4

Sequence number given to ensure uniqueness within a dataset Req for a subject. Can be used to join related records. Used to tie together a block of related records in a single domain to support relationships within the domain and between domains. Perm

SUSPID

Sponsor-Defined Identifier

Char

Identifier

Sponsor-defined reference number. Perhaps pre-printed on the Perm CRF as an explicit line identifier or defined in the sponsors operational database. Example: Line number on a Tobacco & Alcohol use CRF page. Substance name. Examples: Cigarettes, Coffee. If SUTRT is modified, then the modified text is placed here. Req Perm

SUTRT SUMODIFY SUDECOD

Reported Name of Substance Modified Substance Name

Char Char **

CRF Sponsor Defined Derived

Topic Synonym Qualifier Synonym Qualifier

SDTMIG 4.1.3.6 SDTMIG 4.1.3.5 SDTMIG 4.1.3.5

Standardized Substance Name Char

Standardized or dictionary-derived text description of SUTRT Perm or SUMODIFY if the sponsor chooses to code the substance use. The sponsor should specify the dictionary name and version in the Sponsor Comments column of the Define document. If an intervention term does not have a decode value in the dictionary then SUDECOD will be null Used to define a category of related records. Examples: TOBACCO, ALCOHOL, or CAFFEINE. A further categorization of substance use. Examples: CIGARS, CIGARETTES, BEER, WINE Perm Perm

SUCAT SUSCAT

Category for Substance Use

Char

* *

Subcategory for Substance Use Char

Sponsor Defined Sponsor Defined

Grouping Qualifier Grouping Qualifier

SDTMIG 2.1 SDTMIG 2.1

CDISC, 2005. All rights reserved FINAL

Page 49 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name

Variable Label

Type

Controlled Terms or Format **Y, N or Null

Origin

Role

CDISC Notes

Core

References

SUOCCUR

SU Occurrence

Char

CRF or Sponsor Defined

Record Qualifier

SUSTAT SUREASND SUCLAS SUCLASCD SUDOSE

Substance Use Status Reason Substance Use Not Collected Substance Use Class Substance Use Class Code Substance Use Consumption

Char Char Char Char Num

**NOT DONE

CRF CRF

* *

Derived Derived CRF or Sponsor Defined CRF or Derived CRF or Sponsor Defined CRF or Sponsor Defined CRF or Sponsor Defined CRF or Derived CRF or Sponsor Defined CRF or Derived

Record Qualifier Record Qualifier Variable Qualifier Variable Qualifier Record Qualifier Record Qualifier Variable Qualifier Variable Qualifier Variable Qualifier Record Qualifier Variable Qualifier Timing

Used when the use of specific substances is solicited to Perm indicate whether a substance was taken or not. Values are null for spontaneously reported substances. Note: Use of this variable is under evaluation by the SDS Team. Used to indicate that the question was not asked. Perm Describes the reason substance use was not collected. Used in Perm conjunction with SUSTAT when value is NOT DONE. May be used when coding certain substance use cases such as Perm alcoholism or drug abuse. May apply when coding substance abuse use cases. Perm Amount of SUTRT consumed. Perm SDTMIG 4.1.5.3 SDTMIG 4.1.3.5 SDTMIG 4.1.3.1

SUDOSTXT SUDOSU

Substance Use Consumption Text Consumption Units

Char Char *

Substance use consumption amounts or a range of consumption information collected in text form. Units for SUDOSE. Examples: OUNCES, CIGARETTE EQUIVALENTS, or GRAMS.

Perm Perm

SUDOSFRM

Dose Form

Char

Dose form for SUTRT. Examples: INJECTABLE, LIQUID, Perm or POWDER. Usually expressed as the number of uses consumed per a specific interval. Examples: PER DAY, PER WEEK, OCCASIONAL. Total daily use of SUTRT using the units in SUDOSU. Route of administration for SUTRT. Examples: ORAL, INTRAVENOUS, INHALATION. 1. Protocol-defined description of clinical encounter. 2. May be used in addition to VISITNUM and/or VISITDY. Perm

SUDOSFRQ

Use Frequency Per Interval

Char

SUDOSTOT SUROUTE

Total Daily Consumption using Num SUDOSU Route of Administration Char

Perm Perm

VISIT

Visit Name

Char

Perm

SDTMIG 7.8; SDTMIG 4.1.4.5

Page 50 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name

Variable Label

Type

Controlled Terms or Format

Origin

Role

CDISC Notes

Core

References

VISITNUM

Visit Number

Num

CRF or Derived CRF or Derived ISO 8601 ISO 8601 CRF or Derived CRF or Derived Derived Derived ISO 8601 CRF

Timing

1. Clinical encounter number. 2. Numeric version of VISIT, used for sorting.

Req

SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 4.1.4.1 SDTMIG 4.1.4.1 SDTMIG 4.1.4.1 SDTMIG 4.1.4.1

VISITDY

Planned Study Day of Visit

Num

Timing

Perm

SUSTDTC SUENDTC SUSTDY SUENDY SUDUR

Start Date/Time of Substance Use End Date/Time of Substance Use

Char Char

Timing Timing Timing Timing Timing Study day of start of substance use relative to the sponsordefined RFSTDTC. Study day of end of substance use relative to the sponsordefined RFSTDTC. Collected duration and unit of substance use. Used only if collected on the CRF and not derived from start and end date/times.

Perm Perm Perm Perm Perm

Study Day of Start of Substance Num Use Study Day of End of Substance Num Use Duration of Substance Use Char

SUSTRF

Start Relative to Reference Period

Char

** BEFORE, Derived DURING, AFTER

Timing

Identifies the start of the substance use period with respect to Perm the sponsor-defined reference period. Sponsors should define the reference period in the study metadata. SUSTRF should be populated when a start date is not collected. If information such as "PRIOR", "ONGOING", or "CONTINUING" was collected, this information should be translated into SUSTRF. Identifies the end of the substance use period with respect to Perm the sponsor-defined reference period. Sponsors should define the reference period in the study metadata. Use of substances that is ongoing at the end of the reference period should have a value of AFTER for this variable. SUENRF should be populated when an end date is not collected. If information such as "PRIOR", "ONGOING", or "CONTINUING" was collected, this information should be translated into SUENRF.

SDTMIG 4.1.4.7

SUENRF

End Relative to Reference Period

Char

** BEFORE, Derived DURING, AFTER, DURING/ AFTER, U

Timing

SDTMIG 4.1.4.7

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.

CDISC, 2005. All rights reserved FINAL

Page 51 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

6.1.3.1 Assumptions for Substance Use (SU) domain model 1. The intent of the domain is to capture substance use information that may be used to assess the efficacy and/or safety of therapies that look to mitigate the effects of chronic substance use, or that could be used as covariates to support the efficacy and/or safety analyses of other topic variables. 2. SU Definition a. This information may be independent of planned study evaluations or may be a key outcome (e.g., planned evaluation) of a clinical trial. b. In many clinical trials detailed substance use information as represented by this domain may not be required (e.g., the only information collected may be a response to the question Have you ever smoked tobacco?); in this case many of the qualifying variables would not be submitted. Substance Use Description and Coding a. SUTRT captures the verbatim term collected for the substance. It is the topic variable for the SU dataset. SUTRT is a required variable and must have a value. b. SUMODIFY is a permissible variable and should be included if the sponsors procedure permits modification of a verbatim substance use term for coding. The modified term is listed in SUMODIFY. The variable may be populated as per the sponsors procedures; null values are permitted. c. SUDECOD is the preferred term derived by the sponsor from the coding dictionary. It is a permissible variable; Null values are permitted. Where deemed necessary by the sponsor the verbatim term (SUTRT) should be coded using a standard dictionary such as WHODRUG. The sponsor is expected to provide the dictionary name and version used to map the terms in the metadata (using the Comments column in the Define document). d. SUCLAS and SUCLASCD can be used to classify substances using a dictionary that codes to a single class. These variables would not be used if the sponsor were to use WHODRUG to classify substances.

3.

4. Additional Categorization and Grouping a. SUCAT and SUSCAT should not be redundant with the domain code or dictionary classification provided by SUDECOD. That is, they should provide a different means of defining or classifying SU records. For example, a sponsor may be interested in identifying all SU records that the investigator feels might represent opium use and thus will collect that categorization on the CRF. This categorization might differ from the categorization derived from the coding dictionary. It is not necessary to populate these variables. b. SUGRPID may be used to link (or associate) different records together to form a block of related records at the subject level within SU. It should not be used in place of SUCAT or SUSCAT, which are used to group data across subjects. It is not necessary to populate this variable. 5. Presence or Absence of Substance Use a. If specific substances are solicited (i.e., a pre-specified list of terms is collected), SUOCCUR may be used to indicate whether or not each substance was consumed. Use of this variable is under evaluation by the SDS Team. b. For cases when it is important to know whether the occurrence of specific substances is solicited (i.e., a pre-specified list of terms is collected), the SDS Team is proposing the addition of a new variable --PRESP to the SDTM. Until then, SUPRESP can be used as a supplemental qualifier. 6. Timing Variables a. SUSTDTC and SUENDTC may be populated as required. Null values are permissible. b. If substance use information is collected more than once within the CRF (indicating that the data are visit-based) or collected only once within the CRF then VISITNUM is required. VISITDY and VISIT are always permissible. 7. Other Qualifier Variables a. In those instances where detailed dosing information is collected, the variables SUDOSE, SUDOSTXT, SUDOSU, SUDOSFRM, SUDOSFRQ, SUDOSTOT and SUROUTE may be populated; null values are permitted. b. When the response for a substance use question on a CRF was not collected, then the value of SUSTAT = 'NOT DONE' and SUOCCUR will be null.
Page 52 August 26, 2005 CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

6.2
6.2.1

EVENTS
Adverse Events AE

ae.xpt, Adverse Events Events, Version 3.1.1, August 26, 2005. One record per adverse event per subject, Tabulation
Variable Name STUDYID DOMAIN USUBJID AESEQ Variable Label Study Identifier Domain Abbreviation Unique Subject Identifier Sequence Number Type Char Char Char Num Controlled Terms or Format **AE Origin CRF Derived Sponsor Defined CRF or Derived Sponsor Defined Sponsor Defined Sponsor Defined Role Identifier Identifier Identifier Identifier CDISC Notes Core References SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4

Unique identifier for a study within the submission. Req Two-character abbreviation for the domain most Req relevant to the observation. Unique subject identifier within the submission. Req Sequence number given to ensure uniqueness Req within a dataset for a subject. Can be used to join related records. Used to tie together a block of related records in a Perm single domain to support relationships within the domain and between domains. Optional internal or external identifier such as a Perm serial number on an SAE reporting form Optional Sponsor-defined reference number. Perm Perhaps pre-printed on the CRF as an explicit line identifier or defined in the sponsors operational database. Example: Line number on a Adverse Events page. Verbatim name of the event. Req If AETERM is modified, then AEMODIFY will Perm contain the modified text. Dictionary-derived text description of AETERM or Req AEMODIFY. Equivalent to the Preferred Term (PT in MedDRA). The sponsor should specify the dictionary name and version in the Sponsor Comments column of the Define document. Used to define a category of related records. Perm Example: BLEEDING, HYPOGLYCEMIA.

AEGRPID

Group ID

Char

Identifier

SDTM 2.2.4

AEREFID AESPID

Reference ID Sponsor-Defined Identifier

Char Char

Identifier Identifier

SDTM 2.2.4 SDTM 2.2.4

AETERM AEMODIFY AEDECOD

Reported Term for the Adverse Event Modified Reported Term Dictionary-Derived Term

Char Char Char **

CRF Sponsor Defined Derived

Topic Synonym Qualifier Synonym Qualifier

SDTMIG 4.1.3.6 SDTMIG 4.1.3.6 SDTMIG 4.1.3.5; SDTMIG 4.1.3.6

AECAT

Category for Adverse Event

Char

Sponsor Defined

Grouping Qualifier

SDTMIG 4.1.2.6

CDISC, 2005. All rights reserved FINAL

Page 53 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name AESCAT AEOCCUR

Variable Label Subcategory for Adverse Event Adverse Event Occurrence

Type Char Char

Controlled Terms or Origin Format * Sponsor Defined **Y, N or CRF or Null Sponsor Defined

Role Grouping Qualifier Record Qualifier

CDISC Notes

Core

References SDTMIG 4.1.2.6

AEBODSYS

Body System or Organ Class

Char

**

CRF or Derived CRF or Derived CRF CRF or Derived CRF

Record Qualifier Record Qualifier Record Qualifier Record Qualifier Record Qualifier

AELOC AESEV AESER AEACN

Location of the Reaction Severity/Intensity Serious Event Action Taken with Study Treatment

Char Char Char Char

* * **Y, N *

A further categorization of adverse event. Example: Perm NEUROLOGIC. Used when the occurrence of specific adverse Perm events is solicited to indicate whether an adverse event occurred or not. Values are null for spontaneously reported events. Note: Use of this variable is under evaluation by the SDS Team. Body system or organ class (Primary SOC) that is Exp involved in an event or measurement from the standard hierarchy (e.g., MedDRA). Describes anatomical location relevant for the Perm event. (e.g., LEFT ARM for skin rash). The severity or intensity of the event. Examples: Perm MILD, MODERATE, SEVERE. Is this is a serious event? Exp Describes changes to the study treatment as a result of the event. Examples include ICH E2B values: DRUG WITHDRAWN, DOSE REDUCED,DOSE INCREASED, DOSE NOT CHANGED, UNKNOWN or NOT APPLICABLE Describes other actions taken as a result of the event. Usually reported as free text. Example: Treatment unblinded. Primary care physician notified. Records the investigators opinion as to the causality of the event to the treatment. Examples: NOT RELATED,UNLIKELY RELATED, POSSIBLY RELATED, RELATED Records the investigator's opinion as to whether the event may have been to due to a treatment other than study drug. Reported as free text. Example: "More likely related to aspirin use. Used to indicate the pattern of the event over time. Examples: INTERMITTENT, CONTINUOUS, SINGLE EVENT. Exp

SDTMIG 4.1.3.5

AEACNOTH

Other Action Taken

Char

CRF

Record Qualifier

Perm

AEREL

Causality

Char

CRF

Record Qualifier

Exp

AERELNST

Relationship to Non-Study Treatment

Char

CRF

Record Qualifier

Perm

AEPATT

Pattern of Adverse Event

Char

CRF

Record Qualifier

Perm

Page 54 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name AEOUT

Variable Label Outcome of Adverse Event

Type Char

Controlled Terms or Origin Format * CRF

Role Record Qualifier

CDISC Notes

Core

References

AESCAN AESCONG AESDISAB AESDTH AESHOSP AESLIFE AESOD AESMIE AECONTRT AETOXGR

Involves Cancer Congenital Anomaly or Birth Defect Persist or Signif Disability/Incapacity Results in Death Requires or Prolongs Hospitalization Is Life Threatening Occurred with Overdose

Char Char Char Char Char Char Char

Other Medically Important Serious Char Event Concomitant or Additional Trtmnt GivenChar Standard Toxicity Grade Char

**Y, N or Null **Y, N or Null **Y, N or Null **Y, N or Null *Y, N or Null **Y, N or Null **Y, N or Null **Y, N or Null **Y, N **

CRF CRF CRF CRF or Derived CRF CRF CRF CRF CRF CRF or Derived

Record Qualifier Record Qualifier Record Qualifier Record Qualifier Record Qualifier Record Qualifier Record Qualifier Record Qualifier Record Qualifier Record Qualifier

Description of the outcome of an event. Examples Perm include ICH E2B values: RECOVERED/ RESOLVED, RECOVERING/ RESOLVING, NOT RECOVERED/ NOT RESOLVED, RECOVERED/ RESOLVED WITH SEQUELAE, FATAL or UNKNOWN. Was the event associated with the development of Perm cancer? Was the event associated with congenital anomaly Perm or birth defect? Did the event result in persistent or significant Perm disability/incapacity? Did the event result in death? Perm Did the event require or prolong hospitalization? Was the event life threatening? Did the event occur with an overdose? Do additional categories for seriousness apply? Perm Perm Perm Perm

AESTDTC AEENDTC

Start Date/Time of Adverse Event End Date/Time of Adverse Event

Char Char

ISO 8601 ISO 8601

CRF or Derived CRF or Derived

Timing Timing

Was another treatment given because of the Perm occurrence of the event? Toxicity grade according to a standard toxicity Perm scale such as Common Terminology Criteria for Adverse Events v3.0 (CTCAE). Sponsor should specify scale and version used in the Comments column of the Define data definition document. Value should contain valid numbers only; datatype will be changed to numeric in future version of SDTM. Exp Exp

SDTMIG 4.1.4.1; SDTMIG 4.1.4.2 SDTMIG 4.1.4.1; SDTMIG 4.1.4.2

CDISC, 2005. All rights reserved FINAL

Page 55 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name AESTDY AEENDY AEDUR

Variable Label Study Day of Start of Adverse Event Study Day of End of Adverse Event Duration of Adverse Event

Type Num Num Char

Controlled Terms or Format

Origin Derived Derived

Role Timing Timing Timing

CDISC Notes

Core

References SDTMIG 4.1.4.4 SDTMIG 4.1.4.4 SDTMIG 4.1.4.3

ISO 8601

CRF

Study day of start of adverse event relative to the Perm sponsor-defined RFSTDTC. Study day of end of event relative to the sponsor- Perm defined RFSTDTC. Collected duration and unit of an adverse event. Perm Used only if collected on the CRF and not derived from start and end date/times. Example: P1DT2H (for 1 day, 2 hours). Identifies the end of the event as being BEFORE, Perm DURING, DURING/AFTER or AFTER the sponsor-defined reference period. The sponsordefined reference period is a continuous period of time defined by a discrete starting point and a discrete ending point. Typically, this period is defined by the start (RFSTDTC) and end (RFENDTC) of the trial. Sponsors should define the reference period in the study metadata. Events that are ongoing at the end of the reference period should have a value of AFTER for this variable. If information such as "PRIOR", "ONGOING", or "CONTINUING" was collected, this information should be translated into AEENRF.

AEENRF

End Relative to Reference Period

Char

** Derived BEFORE, DURING, AFTER, DURING/ AFTER, U

Timing

SDTMIG 4.1.4.7

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.
6.2.1.1 Assumptions for Adverse Events (AE) domain model 1. AE Definition a. The Adverse Events dataset includes "any untoward medical occurrence in a patient or clinical investigation subject administered a pharmaceutical product and which does not necessarily have to have a causal relationship with this treatment" (ICH E2A). Adverse events may be captured either as free text or a pre-specified list of terms. AETERM captures the verbatim term collected for the event. It is the topic variable for the AE dataset. AETERM is a required variable and must have a value. AEMODIFY is a permissible variable and should be included if the sponsors procedure permits modification of a verbatim term for coding. The modified term is listed in AEMODIFY. The variable should be populated as per the sponsors procedures; null values are permitted.

2.

Adverse Event Description and Coding a. b.

Page 56 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

c. d. 3.

AEDECOD is the preferred term derived by the sponsor from the coding dictionary. It is a required variable and must have a value. It is expected that the reported term (AETERM) will be coded using a standard dictionary such as MedDRA. The sponsor is expected to provide the dictionary name and version used to map the terms in the metadata (using the Comments column in the Define document). AEBODSYS is the primary system organ class derived by the sponsor from the coding dictionary. It is expected that this variable will be populated. AECAT and AESCAT should not be redundant with the domain code or dictionary classification provided by AEDECOD and AEBODSYS. That is, they should provide a different means of defining or classifying AE records. For example, a sponsor may be interested in identifying all adverse events that the investigator feels might represent an infection and thus will collect that categorization on the CRF. This categorization might differ from the categorization derived from the coding dictionary. AEGRPID may be used to link (or associate) different records together to form a block of related records at the subject level within AE. It should not be used in place of AECAT or AESCAT, which are used to group data across subjects.. Adverse events are generally collected in two different ways, either by recording free text or using a pre-specified list of terms. In the latter case, the solicitation of information on specific adverse events may affect the frequency at which they are reported; therefore, the fact that a specific adverse event was solicited may be of interest to reviewers. Consequently, the SDS Team is proposing the addition of a new variable -PRESP to the SDTM to indicate that the use of those specific medications was solicited. Until then, AEPRESP can be used as a supplemental qualifier. If specific adverse events or a pre-specified list of terms are collected, AEOCCUR may be used to indicate whether or not each event occurred. Use of this variable is under evaluation by the SDS Team. If a study collects both pre-specified events as well as subject reported events, the value of AEOCCUR should be populated for all prespecified events and null for subject reported events. When adverse events are collected with the recording of free text, some sponsors may enter a record into the data management system to indicate 'no adverse events,' for a specific subject. For these subjects, do not include a record to indicate that there were no events unless AEOCCUR is being used to represent records for non-occurring events. Otherwise, records should be included in the submission dataset only for adverse events that have actually occurred. Either AEENDTC or AEENRF should be populated. If adverse events are collected more than once within the CRF (indicating that the data are visit-based), then VISITNUM is expected. If the adverse events are collected only once within the CRF (indicating that the data are event-based), then VISITNUM is permissible. VISITDY and VISIT are always permissible.

Additional Categorization and Grouping a.

b. 4.

Presence or Absence of Adverse Events a.

b. c. a.

5.

Timing Variables a. b.

CDISC, 2005. All rights reserved FINAL

Page 57 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

6.

Other Qualifier Variables a. If categories of serious events are collected secondarily to a leading question as in the example below, the values may be null. For example, if Serious is answered 'No,' the values for AESDTH, AESLIFE, AESHOSP, etc. may be null. However, if Serious is answered Yes, at least one of AESDTH, AESLIFE, AESHOSP, etc. will have a Y response. Serious? [ ] Yes [ ] No If yes, check all that apply: [ ] Fatal [ ] Life-threatening [ ] Inpatient hospitalization [ ] etc. On the other hand, if the CRF were structured so that a response is collected for each serious category, all category variables (e.g., AESDTH, AESHOSP) would be populated and AESER could be derived. When a description of Other Medically Important Serious Adverse Events category is collected on a CRF, sponsors should place the description in the SUPPQUAL dataset as described in Section 8.4. The serious categories Involves cancer (AESCAN) and Occurred with overdose (AESOD) are not part of the ICH definition of a serious adverse event, but these categories are available for use for studies conducted under guidelines that existed prior to the FDAs adoption of the ICH definition. In studies using toxicity grade according to a standard toxicity scale such as Common Terminology Criteria for Adverse Events v3.0 (CTCAE The Common Terminology is the current version for toxicity criteria published by the NCI (National Cancer Institute) at http://ctep.cancer.gov/reporting/ctc.html), AETOXGR should be used instead of AESEV. AESEV would not be included on the dataset. The sponsor is expected to specify which scale and version are used in the metadata (using the Comments column in the Define document).

b. c. d.

7.

AE Structure. a. The structure of the AE domain is one record per adverse event per subject. It is the sponsor's responsibility to define an event. This definition may vary based on the sponsor's requirements for characterizing and reporting product safety and is usually described in the protocol. The submission dataset structure may differ from the structure at the time of collection. One common approach is to collapse all records for an event to a summary level with the highest level of severity, causality, seriousness, etc. If there is a need to evaluate AEs at a more granular level, sponsors may choose to create a new record when severity, causality, or seriousness changes or worsens. Other approaches, including a visit-based record, may also be reasonable as long as they meet the sponsor's safety evaluation requirements.

Page 58 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

6.2.2

Disposition DS

ds.xpt, Disposition Events Version 3.1.1, August 26, 2005. One record per disposition status or protocol milestone per subject, Tabulation
Variable Name STUDYID DOMAIN USUBJID DSSEQ DSGRPID DSREFID DSSPID Variable Label Study Identifier Domain Abbreviation Unique Subject Identifier Sequence Number Group ID Reference ID Sponsor-Defined Identifier Type Char Char Char Num Char Char Char Controlled Terms or Format **DS Origin CRF Derived Sponsor Defined CRF or Derived Sponsor Defined Sponsor Defined Sponsor Defined Role Identifier Identifier Identifier Identifier Identifier Identifier Identifier CDISC Notes Unique identifier for a study within the submission. Two-character abbreviation for the domain most relevant to the observation. Unique subject identifier within the submission. Core Req Req Req References SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTMIG 2.1; SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4

Sequence number given to ensure uniqueness within a dataset for Req a subject. Can be used to join related records. Used to tie together a block of related records in a single domain Perm to support relationships within the domain and between domains. Optional internal or external identifier. Perm Optional Sponsor-defined reference number. Perhaps pre-printed Perm on the CRF as an explicit line identifier or defined in the sponsors operational database. Example: Line number on a Disposition page. Verbatim name of the event or protocol milestone. Some terms in Req DSTERM will match DSDECOD, but others, such as 'Subject moved,' will map to controlled terminology in DSDECOD, such as 'LOST TO FOLLOW-UP.'

DSTERM

Reported Term for the Disposition Event

Char

CRF

Topic

SDTMIG 4.1.3.6

CDISC, 2005. All rights reserved FINAL

Page 59 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name DSDECOD

Variable Label Standardized Disposition Term

Type Char

Controlled Terms or Origin Format * CRF or Derived

Role Synonym Qualifier

CDISC Notes Controlled terminology for the name of disposition event or protocol milestone. Examples of disposition events: COMPLETED, ADVERSE EVENT, DEATH, LACK OF EFFICACY, LOST TO FOLLOW-UP, NON-COMPLIANCE WITH STUDY DRUG, PHYSICIAN DECISION, PREGNANCY, PROGRESSIVE DISEASE, PROTOCOL VIOLATION, SCREEN FAILURE, STUDY TERMINATED BY SPONSOR, TECHNICAL PROBLEMS, WITHDRAWAL OF CONSENT, OTHER. Examples of protocol milestones: INFORMED CONSENT OBTAINED, RANDOMIZED Used to define a category of related records. Examples: DISPOSITION EVENT or PROTOCOL MILESTONE. A further categorization of disposition event. 1. Protocol-defined description of clinical encounter. 2. May be used in addition to VISITNUM and/or VISITDY. 1. Clinical encounter number. 2. Numeric version of VISIT, used for sorting.

Core Req

References SDTMIG 4.1.3.5

DSCAT DSSCAT VISIT VISITNUM VISITDY EPOCH DSDTC DSSTDTC DSSTDY

Category for Disposition Char Event Subcategory for Disposition Char Event Visit Name Char Visit Number Planned Study Day of Visit Trial Epoch Date/Time of Collection Start Date/Time of Disposition Event Study Day of Start of Disposition Event Num Num Char Char Char Num

* *

* ISO 8601 ISO 8601

Sponsor Defined Sponsor Defined CRF or Derived CRF or Derived CRF or Derived CRF or Derived CRF or Derived CRF or Derived Derived

Grouping Qualifier Grouping Qualifier Timing Timing Timing Timing Timing Timing Timing

Perm Perm Perm Perm Perm

SDTMIG 2.1 SDTMIG 2.1 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 7.4.4 SDTMIG 4.1.4.1 SDTMIG 4.1.4.1 SDTMIG 4.1.4.4

EPOCH may be used when DSCAT = 'DISPOSITION EVENT'. Perm Examples: TREATMENT PHASE, SCREENING, FOLLOW-UP Perm Exp Study day of start of the disposition event relative to the sponsor- Perm defined RFSTDTC.

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.
6.2.2.1 Assumptions for Disposition (DS) domain model 1. DS Definition a. The Disposition dataset provides an accounting for all subjects who entered the study and may include protocol milestones, such as randomization, as well as the subject's completion status or reason for discontinuation for each phase or segment of the study, including screening and post treatment follow-up. Sponsors may choose which disposition events and milestones to submit for a study.

Page 60 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

2.

DS Description and Coding a. DSTERM contains the disposition or other event information such as a protocol milestone. It is a required variable and must have a value. DSTERM contains either specific verbatim information of the disposition event and/or sponsor defined controlled terms with the same values as DSDECOD. Some terms in DSTERM will match those in DSDECOD, but others such as 'Subject moved' will map to controlled terminology in DSDECOD such as 'LOST TO FOLLOW-UP'. b. DSDECOD is the sponsor defined controlled term for the disposition event or other event information such as protocol milestones. It is a required variable and will always have a value. The value in DSDECOD may be equivalent to the value in DSTERM. For disposition events, DSDECOD = COMPLETE when the subject completed the phase or segment described in EPOCH; otherwise, DSDECOD indicates the reason the subject did not complete the EPOCH. Other Qualifier Variables a. DSCAT may be used to distinguish between disposition events and protocol milestones. It is a permissible variable, but its use is recommended for clarity when both types of events are submitted. b. Additional information about a death such as autopsy data could be modeled in a separate 'autopsy' domain based on the Findings class. Timing Variables a. DSSTDTC is expected as the date/time of the disposition event. Disposition events do not have start and end dates since disposition events do not span an interval (e.g. randomization date) but occur on a single date/time. b. EPOCH may bes used to indicate when the disposition event occurred. EPOCH is populated when DSCAT= 'DISPOSITION EVENT'. EPOCH should not be populated when DSCAT = 'PROTOCOL MILESTONE'.

3.

4.

CDISC, 2005. All rights reserved FINAL

Page 61 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

6.2.3

Medical History MH

mh.xpt, Medical History Events, Version 3.1.1, August 26, 2005. One record per medical history event per subject, Tabulation
Variable Name STUDYID DOMAIN USUBJID MHSEQ MHGRPID Variable Label Study Identifier Domain Abbreviation Unique Subject Identifier Sequence Number Group ID Type Char Char Char Num Char **MH Controlled Terms or Format Origin CRF Derived Sponsor Defined CRF or Derived Sponsor Defined Sponsor Defined Sponsor Defined Role Identifier Identifier Identifier Identifier Identifier CDISC Notes Unique identifier for a study within the submission. Two-character abbreviation for the domain most relevant to the observation. Unique subject identifier within the submission. Core Req Req Req References SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTMIG 2.1; SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4

Sequence number given to ensure uniqueness within a dataset Req for a subject. Can be used to join related records. Used to tie together a block of related records in a single domain to support relationships within the domain and between domains. Optional internal or external medical history identifier. Perm

MHREFID MHSPID

Reference ID Sponsor-Defined Identifier

Char Char

Identifier Identifier

Perm

Optional Sponsor-defined reference number. Perhaps prePerm printed on the CRF as an explicit line identifier or defined in the sponsors operational database. Example: Line number on a Medical History page. Verbatim or preprinted CRF term for the medical condition or Req event. If MHTERM is modified, then MHMODIFY will contain the Perm modified text. Dictionary-derived text description of MHTERM or Perm MHMODIFY. Equivalent to the Preferred Term (PT in MedDRA). The sponsor should specify the dictionary name and version in the Sponsor Comments column of the Define data definition document. Used to define a category of related records. Examples: CARDIAC or GENERAL Perm

MHTERM MHMODIFY MHDECOD

Reported Term for the Medical History Modified Reported Term Dictionary-Derived Term

Char Char Char **

CRF Sponsor Defined Derived

Topic Synonym Qualifier Synonym Qualifier

SDTMIG 4.1.3.6 SDTMIG 4.1.3.5 SDTMIG 4.1.3.5

MHCAT

Category for Medical History Char

Sponsor Defined

Grouping Qualifier

SDTMIG 2.1

Page 62 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name MHSCAT MHOCCUR

Variable Label Subcategory for Medical History

Type Char

Controlled Terms or Format * **Y, N or Null

Origin Sponsor Defined CRF or Sponsor Defined

Role Grouping Qualifier Record Qualifier

CDISC Notes A further categorization of the condition or event.

Core Perm

References SDTMIG 2.1

Medical History Occurrence Char

Used when the occurrence of specific medical history events Perm is solicited to indicate whether or not a medical condition had ever occurred. Values are null for spontaneously reported events. Note: Use of this variable is under evaluation by the SDS Team.

MHSTAT

Medical History Status

Char

**NOT DONE

CRF or Sponsor Defined CRF

Record Qualifier Record Qualifier Record Qualifier Timing Timing Timing Timing Timing Timing Timing

The status indicates that the question was not asked.

Perm

SDTMIG 4.1.5.1.

MHREASND MHBODSYS

Reason Medical History Not Char Collected Body System or Organ Class Char **

Describes the reason medical history was not collected. Used Perm in conjunction with MHSTAT when value is NOT DONE. Body system or organ class (Primary SOC) that is involved in Perm an event or measurement from the standard hierarchy (e.g., MedDRA). 1. Protocol-defined description of clinical encounter. Perm 2. May be used in addition to VISITNUM and/or VISITDY. 1. Clinical encounter number. 2. Numeric version of VISIT, used for sorting. Perm Perm Perm Perm Perm 1. Study day of medical history collection, measured as Perm integer days. 2. Algorithm for calculations must be relative to the sponsordefined RFSTDTC variable in Demographics. This formula should be consistent across the submission.

SDTMIG 4.1.5.3 SDTMIG 4.1.3.5

Derived

VISIT VISITNUM VISITDY MHDTC MHSTDTC MHENDTC MHDY

Visit Name Visit Number Planned Study Day of Visit Date/Time of History Collection Start Date/Time of Medical History Event End Date/Time of Medical History Event Study Day of History Collection

Char Num Num Char Char Char Num ISO 8601 ISO 8601 ISO 8601

CRF or Derived CRF or Derived CRF or Derived CRF or Derived CRF or Derived CRF or Derived Derived

SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 4.1.4.1

SDTMIG 4.1.4.1 SDTMIG 4.1.4.4

CDISC, 2005. All rights reserved FINAL

Page 63 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name MHENRF

Variable Label End Relative to Reference Period

Type Char

Controlled Terms or Format

Origin

Role Timing

CDISC Notes

Core

References SDTMIG 4.1.4.7

** BEFORE, Derived DURING, AFTER, DURING /AFTER, U

Identifies the end of the event as being BEFORE, DURING, Perm DURING/AFTER or AFTER the sponsor-defined reference period. Sponsors should define the reference period in the study metadata. Events that are ongoing at the end of the reference period should have a value of AFTER for this variable. If information such as "PRIOR", "ONGOING", or "CONTINUING" was collected, this information should be translated into MHENRF.

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.
6.2.3.1 1. Assumptions for Medical History (MH) domain model MH Definition a. The Medical History dataset generally includes the subject's prior history at the start of the trial. Note that treatments for prior and concomitant conditions should usually be submitted in an appropriate dataset from the Interventions class such as CM or SG. b. Examples of subject medical history information could include general medical history, gynecological history, and primary diagnosis. Medical History Description and Coding a. MHTERM captures the verbatim term collected for the condition or event. It is the topic variable for the MH dataset. MHTERM is a required variable and must have a value. b. MHMODIFY is a permissible variable and should be included if the sponsors procedure permits modification of a verbatim term for coding. The modified term is listed in MHMODIFY. The variable should be populated as per the sponsors procedures; null values are permitted. c. If the sponsor codes the reported term (MHTERM) using a standard dictionary, then MHDECOD will be populated with the preferred term derived from the dictionary. The sponsor is expected to provide the dictionary name and version used to map the terms in the metadata (using the Comments column in the Define document). d. If the sponsor codes the reported term (MHTERM) using a standard dictionary, then MHBODSYS will be populated with the primary system organ class derived from the dictionary. e. If a CRF collects medical history by pre-specified body systems and the sponsor also codes reported terms using a standard dictionary, then MHDECOD and MHBODSYS are populated using the information from the standard dictionary. MHCAT and MHSCAT should be used for the pre-specified body systems.

2.

Page 64 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

3.

Additional Categorization and Grouping a. MHCAT and MHSCAT may be populated with the sponsor's categorization of medical history events, which are often displayed on the CRF. MHCAT and MHSCAT should not be redundant with the domain code or dictionary classification provided by MHDECOD and/or MHBODSYS. That is, they should provide a different means of defining or classifying MH records. i. This categorization should not group all records (within the MH Domain) into one generic group such as 'Medical History' or 'General Medical History' because this is redundant information with the domain code. If no smaller categorization can be applied, then it is not necessary to include or populate this variable. ii. Examples of MHCAT could include 'General Medical History,' 'Allergy Medical History,' and 'Reproductive Medical History.' b. MHGRPID may be used to link (or associate) different records together to form a block of related records at the subject level within MH. It should not be used in place of MHCAT or MHSCAT, which are used to group data across subjects. For example, if a group of syndromes were related to a particular disease then this variable could be populated with the appropriate text. Presence or Absence of History a. Information on medical history is generally collected in two different ways, either by recording free text or using a pre-specified list of terms. In the latter case, the solicitation of information on specific medical history events may affect the frequency at which they are reported; therefore, the fact that a specific medical history event was solicited may be of interest to reviewers. Consequently, the SDS Team is proposing the addition of a new variable --PRESP to the SDTM to indicate that the occurrence of those specific medical history events was solicited. Until then, MHPRESP can be used as a supplemental qualifier. b. If history is collected by body system or category, MHOCCUR may be used to indicate the presence or absence of events for a given system or category. c. If specific medical history events are solicited (i.e., a pre-specified list of terms is collected), MHOCCUR may be used to indicate whether or not each event occurred. Use of this variable is under evaluation by the SDS Team. d. When the response for a medical history question or group of questions on a CRF was not collected, then the value of MHSTAT = 'NOT DONE' and MHOCCUR will be null. The reason that medical history was not done may be documented in MHREASND. e. If a study collects both pre-specified events as well as subject reported events, the value of MHOCCUR should be populated for all prespecified events and null for subject-reported events unless the response was not collected.

4.

CDISC, 2005. All rights reserved FINAL

Page 65 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

6.3
6.3.1

FINDINGS
ECG Test Results EG

eg.xpt, ECG Findings, Version 3.1.1, August 26, 2005. One record per ECG observation per time point per visit per subject, Tabulation
Variable Name STUDYID DOMAIN USUBJID EGSEQ EGGRPID EGREFID Variable Label Study Identifier Domain Abbreviation Unique Subject Identifier Sequence Number Group ID ECG Reference ID Type Char Char Char Num Char Char **EG Controlled Terms or Format Origin CRF Derived Sponsor Defined CRF or Derived Sponsor Defined Sponsor Defined or Derived Sponsor Defined Role Identifier Identifier Identifier Identifier Identifier Identifier CDISC Notes Unique identifier for a study within the submission. Two-character abbreviation for the domain most relevant to the observation. Unique subject identifier within the submission. Core Req Req Req References SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTMIG 2.1; SDTM 2.2.4

Sequence number given to ensure uniqueness within a dataset for Req a subject. Can be used to join related records. Used to tie together a block of related records in a single domain Perm to support relationships within the domain and between domains. Internal or external ECG identifier. Example: UUID for external Perm ECG Waveform File. Optional Sponsor-defined reference number. Perhaps pre-printed Perm on the CRF as an explicit line identifier or defined in the sponsor's operational database. Example: Line number from the ECG page. Short name of the measurement, test, or examination described in Req EGTEST. It can be used as a column name when converting a dataset from a vertical to a horizontal format. The value in EGTESTCD cannot be longer than 8 characters, nor can it start with a number (e.g.,'1TEST'). EGTESTCD cannot contain characters other than letters, numbers, or underscores. Examples: PR, QT, INTP.

EGSPID

Sponsor-Defined Identifier

Char

Identifier

SDTM 2.2.4

EGTESTCD

ECG Test or Examination Short Name

Char

**

CRF or Derived

Topic

Page 66 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name EGTEST EGCAT EGSCAT EGPOS EGORRES

Variable Label ECG Test or Examination Name Category for ECG Subcategory for ECG ECG Position of Subject

Type Char Char Char Char

Controlled Terms or Format ** * * *

Origin CRF Sponsor Defined Sponsor Defined CRF or Derived CRF or Derived

Role Synonym Qualifier Grouping Qualifier Grouping Qualifier Record Qualifier Result Qualifier Variable Qualifier Variable Qualifier

CDISC Notes

Core

References

Verbatim name of the test or examination used to obtain the Req measurement or finding. The value in EGTEST cannot be longer than 40 characters. Examples: PR Interval, QT Interval, etc. Used to categorize ECG observations. Examples: Perm MEASUREMENT or FINDING. A further categorization of the ECG. Position of the subject during a measurement or examination. Examples: SUPINE, STANDING: SITTING. Perm Perm

SDTMIG 2.1 SDTMIG 2.1

Result or Finding in Original Char Units Original Units Reference Range Indicator Char Char * *

EGORRESU EGNRIND

CRF or Derived CRF or Derived

Result of the ECG measurement or finding as originally received Exp or collected. Examples of expected values are 62 or 0.151 when EGTEST is a measured finding, or ATRIAL FIBRILLATION or QT PROLONGATION when EGTEST is 'interpretation'. Original units in which the data were collected. The unit for Perm EGORRES. Examples: SECONDS or MILLISECONDS. 1. Indicates where value falls with respect to reference range Perm which could be defined by EGORNRLO and EGORNRHI (if those variables are used) or by EGSTNRC. Examples: NORMAL, ABNORMAL, HIGH, LOW. 2. Should not be used to indicate clinical significance. Contains the result value for all findings, copied or derived from Exp EGORRES in a standard format or standard units. EGSTRESC should store all results or findings in character format; if results are numeric, they should also be stored in numeric format in EGSTRESN. For example, if a test has results 'NONE', 'NEG', and 'NEGATIVE' in EGORRES and these results effectively have the same meaning, they could be represented in standard format in EGSTRESC as 'NEGATIVE'. For other examples, see general assumptions. Additional examples of result data; SINUS BRADYCARDIA, ATRIAL FLUTTER, ATRIAL FIBRILLATION. Used for continuous or numeric results or findings in standard Perm format; copied in numeric format from EGSTRESC. EGSTRESN should store all numeric test results or findings. Standardized unit used for EGSTRESC or EGSTRESN. Perm

SDTMIG 4.1.5.1 SDTMIG 4.1.5.1

EGSTRESC

Character Result/Finding in Char Std Format

Derived

Result Qualifier

SDTMIG 4.1.5.1

EGSTRESN

Numeric Result/Finding in Standard Units

Num

Derived

Result Qualifier

SDTMIG 4.1.5.1

EGSTRESU

Standard Units

Char

CRF or Derived

Variable Qualifier

SDTMIG 4.1.5.1

CDISC, 2005. All rights reserved FINAL

Page 67 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name EGSTAT

Variable Label ECG Status

Type Char

Controlled Terms or Format **NOT DONE

Origin CRF or Derived CRF or Derived Sponsor Defined or Derived CRF or Derived

Role Record Qualifier Record Qualifier Record Qualifier Record Qualifier Synonym Qualifier

CDISC Notes

Core

References SDTMIG 4.1.5.1

Used to indicate an ECG was not done, or an ECG measurement Perm was not taken. Should be null if a result exists in EGORRES. Describes why a measurement or test was not performed. Examples: BROKEN EQUIPMENT or SUBJECT REFUSED. Used in conjunction with EGSTAT when value is NOT DONE. File name for the external ECG Waveform file Perm

EGREASND

Reason ECG Not Performed Char

EGXFN

ECG External file Name

Char

Perm

EGNAM EGLOINC

Vendor Name LOINC Code

Char Char **

Name or identifier of the laboratory or vendor who provides the Perm test results. 1. LOINC Code for EGTEST. LOINC codes are available for many but not all commonly used ECG test observations. 2. The LOINC version should be provided in the Sponsor Comments column in the data definition document. Perm SDTMIG 4.1.3.2

Derived

EGMETHOD EGBLFL EGDRVFL

Method of ECG Test Baseline Flag Derived Flag

Char Char Char

CRF

Record Qualifier Record Qualifier Record Qualifier

Method of the ECG test. Examples: MONITORING (1-LEAD), Perm 12-LEAD. Indicator used to identify a baseline value. Exp SDTMIG 4.1.3.7 SDTMIG 4.1.3.7

**Y or Null CRF or Derived **Y or Null Derived

Used to indicate a derived record. The value should be Y or null. Perm Records which represent the average of other records or questionnaire subscores, which do not come from the CRF are examples of records which would be derived for the submission datasets. If EGDRVFL=Y, then EGORRES should be NULL, with EGSTRESC, and (if numeric) EGSTRESN having the derived value. Role of the person who provided the evaluation. Used only for Exp results that are subjective (e.g., assigned by a person or a group). Should be null for records that contain collected or derived data. Examples: INVESTIGATOR, ADJUDICATION COMMITTEE, VENDOR. 1. Protocol-defined description of clinical encounter. 2. May be used in addition to VISITNUM and/or VISITDY. Perm

EGEVAL

Evaluator

Char

CRF

Record Qualifier

VISIT

Visit Name

Char

CRF or Derived

Timing

SDTMIG 7.8; SDTMIG 4.1.4.5

Page 68 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name VISITNUM VISITDY

Variable Label Visit Number

Type Num

Controlled Terms or Format

Origin CRF or Derived CRF or Derived

Role Timing Timing

CDISC Notes 1. Clinical encounter number. 2. Numeric version of VISIT, used for sorting.

Core Req Perm

References SDTMIG 7.8; 4.1.4.5 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 4.1.4.1 SDTMIG 4.1.4.4 SDTM 2.2.6

Planned Study Day of Visit Num

EGDTC EGDY

Date/Time of ECG Study Day of ECG

Char Num

ISO 8601

CRF or Derived Derived

Timing Timing 1. Study day of the ECG, measured as integer days. 2. Algorithm for calculations must be relative to the sponsordefined RFSTDTC variable in Demographics. This formula should be consistent across the submission. Numerical version of EGTPT to aid in sorting.

Exp Perm

EGTPTNUM EGTPT

Planned Time Point Number Num Planned Time Point Name Char

CRF or Derived CRF or Derived

Timing Timing

Perm

1. Text Description of time when measurement should be taken. Perm 2. This may be represented as an elapsed time relative to a fixed reference point, such as time of last dose. See EGTPTNUM and EGTPTREF. Examples: Start, 5 min post. Elapsed time (in ISO 8601) relative to a planned fixed reference Perm (EGTPTREF). This variable is useful where there are repetitive measures. Not a clock time or a date time variable. Examples: 'P15M' to represent the period of 15 minutes prior to the reference point indicated by EGTPTREF, or 'P8H' to represent the period of 8 hours after the reference point indicated by EGTPTREF. Name of the fixed reference point referred to by EGELTM, Perm EGTPTNUM, and EGTPT. Examples: Previous Dose, Previous Meal. SDTMIG 4.1.4.3

EGELTM

Elapsed Time from Reference Point

Char

ISO 8601

Derived

Timing

EGTPTREF

Time Point Reference

Char

Sponsor Defined

Timing

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.
6.3.1.1 Assumptions for ECG (EG) domain model 1. The method for QT interval correction is included in the test name by controlled terminology. The controlled terminology would be QTCF for Fridericia and QTCB for Bazett in EGTESTCD. EGTEST would contain either 'QT interval correction Fridericia' or 'QT interval correction Bazett'. The value of the calculated QT interval results goes into EGORRES. 2. Although EGLOINC is included to allow use of LOINC codes for ECG tests where appropriate, other coding schemes may be recommended by the CDISC terminology team in the future.
Page 69 July 28, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

6.3.2

Inclusion/Exclusion Exceptions IE

ie.xpt, Inclusion/Exclusion Exceptions Findings, Version 3.1.1, August 26, 2005. One record per inclusion/exclusion criteria exception per subject, Tabulation
Variable Name STUDYID DOMAIN USUBJID IESEQ IESPID Variable Label Study Identifier Domain Abbreviation Unique Subject Identifier Sequence Number Sponsor-Defined Identifier Type Char Char Char Num Char **IE Controlled Terms or Format Origin CRF Derived Sponsor Defined CRF or Derived Sponsor Defined Role Identifier Identifier Identifier Identifier Identifier CDISC Notes Unique identifier for a study within the submission. Core Req References SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4

Two-character abbreviation for the domain most relevant to Req the observation. Unique subject identifier within the submission. Sequence number given to ensure uniqueness within a dataset for a subject. Can be used to join related records. Req Req

Sponsor-defined reference number. Perhaps pre-printed on Perm the CRF as an explicit line identifier or defined in the sponsors operational database. Example: Question number on a questionnaire. Short name of the criterion described in IETEST. It can be Req used as a column name when converting a dataset from a vertical to a horizontal format. The value in IETESTCD cannot be longer than 8 characters, nor can it start with a number (e.g.'1TEST'). IETESTCD cannot contain characters other than letters, numbers, or underscores. Examples: IN01, EX01. Verbatim description of the inclusion or exclusion criterion Req that was the exception for the subject within the study. Used to define a category of related records. A further categorization of the exception criterion. Can be used to distinguish criteria for a sub-study. Req Perm

IETESTCD

Inclusion/Exclusion Criterion Char Short Name

CRF

Topic

IETEST IECAT IESCAT

Inclusion/Exclusion Criterion Char Inclusion//Exclusion Category Inclusion//Exclusion Subcategory Char Char

CRF **INCLUSION, Sponsor EXCLUSION Defined * Sponsor Defined

Synonym Qualifier Grouping Qualifier Grouping Qualifier

SDTMIG 2.1 SDTMIG 2.1

Page 70 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name IEORRES

Variable Label

Type

Controlled Terms or Format **Y, N

Origin CRF or Derived

Role

CDISC Notes

Core

References SDTMIG 4.1.5.1

I/E Criterion Original Result Char

Result Qualifier Original response to Inclusion/Exclusion Criterion question. Req Inclusion or Exclusion criterion met?

IESTRESC VISIT

I/E Criterion Result in Std Char Format

Visit Name Char

**Y, N

Derived CRF or Derived CRF or Derived CRF or Derived

Result Qualifier Response to Inclusion/Exclusion criterion result in standard Req format. (Same as IEORRES for the IE Domain). Timing 1. Protocol-defined description of clinical encounter. Perm 2. May be used in addition to VISITNUM and/or VISITDY. 1. Clinical encounter number. 2. Numeric version of VISIT, used for sorting. Perm

SDTMIG 4.1.5.1 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 4.1.4.1 SDTMIG 4.1.4.4 SDTM 2.2.5

VISITNUM

Visit Number

Num

Timing

VISITDY

Planned Study Day of Visit Num

Timing

Perm

IEDTC IEDY

Date/Time of Collection Study Day of Collection

Char Num

ISO 8601

CRF or Derived Derived

Timing Timing

Perm 1. Study day of collection of the inclusion/exclusion Perm exceptions, measured as integer days. 2. Algorithm for calculations must be relative to the sponsordefined RFSTDTC variable in Demographics. This formula should be consistent across the submission.

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.
6.3.2.1 Assumptions for Inclusion/Exclusion Exceptions (IE) domain model 1. The intent of the domain model is to only collect those criteria that cause the subject to be in violation of the Inclusion/Exclusion criteria not a response to each criterion. 2. 3. This domain should not be used to collect protocol deviations/violations incurred during the course of the study. A separate domain model (DV, available separately) will be developed for protocol deviations/violations in a future version of the Implementation Guide. Note that the SQL statement query "select distinct IETESTCD from IE; may not contain an exhaustive list of the inclusion / exclusion criteria because this domain only lists those not met by subjects. The complete list would be found in the TI trial inclusion/exclusion criteria dataset described in Section 7.9 .

CDISC, 2005. All rights reserved FINAL

Page 71 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

6.3.3

Laboratory Test Results LB

lb.xpt, Labs Findings, Version 3.1.1, August 26, 2005. One record per lab test per time point per visit per subject, Tabulation
Variable Name STUDYID DOMAIN USUBJID LBSEQ LBGRPID LBREFID Variable Label Study Identifier Domain Abbreviation Unique Subject Identifier Sequence Number Group ID Specimen ID Type Char Char Char Num Char Char **LB Controlled Terms or Format Origin CRF Derived Sponsor Defined CRF or Derived Sponsor Defined Role CDISC Notes Core Req Req Req Req Perm Perm References SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTMIG 2.1; SDTM 2.2.4

Identifier Unique identifier for a study within the submission. Identifier Two-character abbreviation for the domain most relevant to the observation. Identifier Unique subject identifier within the submission. Identifier Sequence number given to ensure uniqueness within a dataset for a subject. Can be used to join related records. Identifier Used to tie together a block of related records in a single domain to support relationships within the domain and between domains.

Sponsor Identifier Internal or external specimen identifier. Example: Specimen ID. Defined or Derived Sponsor Defined * CRF or Derived

LBSPID

Sponsor-Defined Identifier

Char

Identifier Optional Sponsor-defined reference number. Perhaps pre-printed on Perm the CRF as an explicit line identifier or defined in the sponsors operational database. Example: Line number on the Lab page. Topic Short name of the measurement, test, or examination described in Req LBTEST. It can be used as a column name when converting a dataset from a vertical to a horizontal format. The value in LBTESTCD cannot be longer than 8 characters, nor can it start with a number (e.g.'1TEST'). LBTESTCD cannot contain characters other than letters, numbers, or underscores. Examples: ALT, LDH. Req

SDTM 2.2.4

LBTESTCD

LAB Test or Examination Short Name

Char

LBTEST

LAB Test or Examination Name

Char

CRF

Synonym Verbatim name of the test or examination used to obtain the Qualifier measurement or finding. Note any test normally performed by a clinical laboratory is considered a lab test. The value in LBTEST cannot be longer than 40 characters. Examples: Alanine Aminotransferase, Lactate Dehydrogenase. Grouping Used to define a category of related records. Examples: such as Qualifier HEMATOLOGY, URINALYSIS, CHEMISTRY.

LBCAT LBSCAT LBORRES

Category for Lab Test Subcategory for Lab Test

Char Char

* *

Sponsor Defined Sponsor Defined CRF or Derived

Exp

SDTMIG 2.1 SDTMIG 2.1. SDTMIG 4.1.5.1

Grouping A further categorization of a test category such as DIFFERENTIAL, Perm Qualifier COAGULATON, LIVER FUNCTION or ELECTROLYTES. Result Result of the measurement or finding as originally received or Qualifier collected. Exp

Result or Finding in Original Char Units

Page 72 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name LBORRESU LBORNRLO LBORNRHI LBSTRESC

Variable Label Original Units Reference Range Lower Limit in Orig Unit Reference Range Upper Limit in Orig Unit

Type Char Char Char *

Controlled Terms or Format

Origin CRF or Derived CRF or Derived CRF or Derived Derived

Role

CDISC Notes

Core Exp

References SDTMIG 4.1.3.2

Variable Original units in which the data were collected. The unit for Qualifier LBORRES. Example: g/L.

Variable Lower end of reference range for continuous measurements in original Exp Qualifier units. Should be populated only for continuous results. Variable Upper end of reference range for continuous measurements in original Exp Qualifier units. Should be populated only for continuous results. Result Contains the result value for all findings, copied or derived from Exp Qualifier LBORRES in a standard format or standard units. LBSTRESC should store all results or findings in character format; if results are numeric, they should also be stored in numeric format in LBSTRESN. For example, if a test has results 'NONE', 'NEG', and 'NEGATIVE' in LBORRES and these results effectively have the same meaning, they could be represented in standard format in LBSTRESC as 'NEGATIVE'. For other examples, see general assumptions. Variable For normal range values that are character in ordinal scale (e.g., Perm Qualifier 'NEGATIVE TO TRACE', '-1 to +1', etc.). If categorical ranges were supplied. Result Used for continuous or numeric results or findings in standard format; Exp Qualifier copied in numeric format from LBSTRESC. LBSTRESN should store all numeric test results or findings. Variable Standardized unit used for LBSTRESC or LBSTRESN. Qualifier Variable Lower end of reference range for continuous measurements in Qualifier standardized units. Should be populated only for continuous results. Variable Upper end of reference range for continuous measurements in Qualifier standardized units. Should be populated only for continuous results. Exp Exp Exp SDTMIG 4.1.5.1 SDTMIG 4.1.5.1

Character Result/Finding in Char Std Format

LBSTNRC

Reference Range for Char Rslt-Std Units Numeric Result/Finding in Standard Units Standard Units Reference Range Lower Limit-Std Units Reference Range Upper Limit-Std Units Reference Range Indicator

Char

CRF or Derived Derived

LBSTRESN

Num

LBSTRESU LBSTNRLO LBSTNRHI LBNRIND

Char Num Num Char

CRF or Derived CRF or Derived CRF or Derived

CRF or Derived

Variable 1. Indicates where value falls with respect to reference range defined Exp Qualifier by LBORNRLO and LBORNRHI or by LBSTNRC. Examples: NORMAL, ABNORMAL, HIGH, LOW. 2. Should not be used to indicate clinical significance. Record Used to indicate exam not done. Should be null if a result exists in Qualifier LBORRES. Record Describes why a measurement or test was not performed such as Qualifier BROKEN EQUIPMENT, SUBJECT REFUSED, or SPECIMEN LOST. Used in conjunction with LBSTAT when value is NOT DONE. Perm Perm SDTMIG 4.1.5.1

LBSTAT LBREASND

Lab Status Reason Test Not Done

Char Char

**NOT DONE

CRF or Derived CRF or Derived

CDISC, 2005. All rights reserved FINAL

Page 73 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name LBNAM LBLOINC

Variable Label Vendor Name LOINC Code

Type Char Char

Controlled Terms or Format

Origin CRF or Derived

Role

CDISC Notes

Core Perm

References

Record Name or identifier of the laboratory or vendor who provides the test Qualifier results.

**

Derived

Synonym 1. LOINC Code for LBTEST. Perm Qualifier 2. The LOINC version should be provided in the comments column in the data definition file. Record Defines the type of specimen used for a measurement. Examples: Qualifier SERUM, PLASMA, URINE. Perm

SDTMIG 4.1.3.2

LBSPEC LBSPCCND LBMETHOD LBBLFL LBDRVFL

Specimen Type Specimen Condition Method of Test or Examination Baseline Flag Derived Flag

Char Char Char Char Char

CRF CRF

Record Free or standardized text describing the condition of the specimen e.g. Perm Qualifier HEMOLYZED, ICTERIC, LIPEMIC etc. Record Method of the test or examination. Examples EIA (Enzyme Qualifier Immunoassay), ELECTROPHORESIS, DIPSTICK Record Indicator used to identify a baseline value. Qualifier Perm Exp SDTMIG 4.1.3.7 SDTMIG 4.1.5.1 SDTMIG 4.1.3.7

* **Y or Null **Y or Null

CRF CRF or Derived CRF or Derived

Record Used to indicate a derived record. The value should be Y or null. Perm Qualifier Records which represent the average of other records or questionnaire subscores, which do not come from the CRF, are examples of records which would be derived for the submission datasets. If LBDRVFL=Y, then LBORRES should be NULL, with LBSTRESC, and (if numeric) LBSTRESN having the derived value. Record Indicator used to identify fasting status such as Y, N, U, or null if not Perm Qualifier relevant. Variable Description of toxicity quantified by LBTOXGR. Sponsor should Qualifier specify which scale and version is used in sponsor comments. Example: NCI CTCAE short name Perm

LBFAST LBTOX

Fasting Status Toxicity

Char Char

**Y, N, U or CRF Null * Derived

LBTOXGR

Standard Toxicity Grade

Char

**

CRF or Derived CRF or Derived CRF or Derived CRF or Derived

Variable Records toxicity grade value using a standard toxicity scale (such as Perm Qualifier the NCI CTCAE). Value should contain valid numbers only; datatype will be changed to numeric in future version of SDTM. Timing Timing Timing Timing 1. Protocol-defined description of clinical encounter 2.May be used in addition to VISITNUM and/or VISITDY 1. Clinical encounter number. 2. Numeric version of VISIT, used for sorting. Perm Req Perm Exp SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 4.1.4.1

VISIT VISITNUM VISITDY LBDTC

Visit Name Visit Number

Char Num

Planned Study Day of Visit Num Date/Time of Specimen Collection Char ISO 8601

CRF or Derived

Page 74 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name LBENDTC LBDY

Variable Label

Type

Controlled Terms or Format ISO 8601

Origin CRF or Derived Derived

Role Timing Timing

CDISC Notes

Core Perm

References SDTMIG 4.1.4.1 SDTMIG 4.1.4.8 SDTMIG 4.1.4.4

End Date/Time of Specimen Char Collection Study Day of Specimen Collection Num

1. Study day of specimen collection, measured as integer days. Perm 2. Algorithm for calculations must be relative to the sponsor-defined RFSTDTC variable in Demographics. This formula should be consistent across the submission. 1. Text Description of time when specimen should be taken. 2. This may be represented as an elapsed time relative to a fixed reference point, such as time of last dose. See LBTPTNUM and LBTPTREF. Examples: Start, 5 min post. Numerical version of LBTPT to aid in sorting. Elapsed time (in ISO 8601) relative to a planned fixed reference (LBTPTREF).. This variable is useful where there are repetitive measures. Not a clock time or a date time variable. Examples: 'Perm

LBTPT

Planned Time Point Name

Char

CRF or Derived

Timing

LBTPTNUM LBELTM

Planned Time Point Number Num Elapsed Time from Reference Point Char ISO 8601

CRF or Derived Derived

Timing Timing

Perm Perm SDTMIG 4.1.4.3

P15M' to represent the period of 15 minutes prior to the reference point indicated by LBTPTREF, or 'P8H' to represent the period of 8 hours after the reference point indicated by LBTPTREF.
LBTPTREF Time Point Reference Char Sponsor Defined Timing Name of the fixed reference point referred to by LBELTM, LBTPTNUM, and LBTPT. Examples: PREVIOUS DOSE, PREVIOUS MEAL. Perm

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.
6.3.3.1 Assumptions for Laboratory Test Results (LB) domain model 1. For lab tests that do not have continuous numeric results (e.g., urine protein as measured by dipstick, descriptive tests such as urine color), LBSTNRC could be populated either with normal range values that are character in an ordinal scale (e.g., 'NEGATIVE to TRACE') or a delimited set of values that are considered to be normal (e.g., 'YELLOW, AMBER'). LBORNRLO, LBORNRHI, LBSTNRLO, and LBSTNRHI should be null for these type of tests. 2. For lab tests where the specimen is collected over time, i.e., 24-hour urine collection, the start date/time of the collection goes into LBDTC and the end date/time of collection goes into LBENDTC.

CDISC, 2005. All rights reserved FINAL

Page 75 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

6.3.4

Physical Examinations PE

pe.xpt, Physical Examination Findings, Version 3.1.1, August 26, 2005. One record per body system or abnormality per visit per subject, Tabulation
Variable Name STUDYID DOMAIN USUBJID PESEQ PEGRPID Variable Label Study Identifier Domain Abbreviation Unique Subject Identifier Sequence Number Group ID Type Char Char Char Num Char Controlled Terms or Format **PE Origin CRF Derived Role Identifier Identifier CDISC Notes Core References SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTMIG 2.1; SDTM 2.2.4 SDTM 2.2.4

Unique identifier for a study within the submission. Req Two-character abbreviation for the domain most relevant to the Req observation. Unique subject identifier within the submission. Sequence number given to ensure uniqueness within a dataset for a subject. Can be used to join related records. Req Req

Sponsor Defined Identifier CRF or Derived Identifier

Sponsor Defined Identifier

Used to tie together a block of related records in a single Perm domain to support relationships within the domain and between domains. Sponsor-defined reference number. Perhaps pre-printed on the CRF as an explicit line identifier or defined in the sponsors operational database. Example: Question number on a questionnaire. Topic variable for PE. Short name for the value in PETEST, which can be used as a column name when converting the dataset from a vertical format to a horizontal format. The short names can be up to 8 characters. Examples: CV, GI. For subject-level exam, value should be 'ALL'. Verbatim term part of the body examined. The value in PETEST cannot be longer than 40 characters. Examples: CARDIOVASCULAR and GASTROINTESTINAL. For subject-level exam, value should be 'ALL'. If PEORRES is modified as part of a defined procedure, then PEMODIFY will contain the modified text. Used to define a category of examination. Examples: GENERAL, NEUROLOGICAL. Perm

PESPID

Sponsor-Defined Identifier Char

Sponsor Defined Identifier

PETESTCD

Body System Examined Short Name

Char

CRF or Derived

Topic

Req

PETEST

Body System Examined

Char

CRF

Synonym Qualifier

Req

PEMODIFY PECAT PESCAT PELOC

Modified Reported Term Category for Examination Subcategory for Examination

Char Char Char * * *

Sponsor Defined Synonym Qualifier Sponsor Defined Grouping Qualifier Sponsor Defined Grouping Qualifier CRF or Sponsor Record Defined Qualifier

Perm Perm

SDTMIG 4.1.3.6 SDTMIG 2.1 SDTMIG 2.1

A further categorization of the examination. Used if needed to Perm add further detail to PECAT. Can be used to specify where a physical exam finding occurred. Perm Example: LEFT ARM for skin rash.

Location of Physical Exam Char Finding

Page 76 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name PEBODSYS

Variable Label Body System or Organ Class

Type Char

Controlled Terms or Format **

Origin CRF or Derived

Role Result Qualifier

CDISC Notes

Core

References SDTMIG 4.1.3.5

1. Body system or organ class (Primary SOC) that is involved in Perm an event or measurement from the standard hierarchy (e.g., MedDRA). 2. May be predefined from CRF page or obtained from coding. Text description of any abnormal findings. If the examination Exp was completed and there were no abnormal findings, the value should be NORMAL. If the examination was not performed on a particular body system, or at the subject level, then the value should be null, and NOT DONE should appear in PESTAT. Original units in which the data were collected. The unit for PEORRES. If there are findings for a body system, then either the dictionary preferred term (if findings are coded using a dictionary) or PEORRES (if findings are not coded) should appear here. If PEORRES is null, PESTRESC should be null Perm Exp

PEORRES

Verbatim Examination Finding

Char

CRF or Derived

Result Qualifier

SDTMIG 4.1.5.1

PEORRESU PESTRESC

Original Units

Char

CRF or Derived Derived

Variable Qualifier Result Qualifier

Character Result/Finding in Char Std. Format

SDTMIG 4.1.5.1

PESTRESN

Numeric Result/Finding in Num Standard Units Standard Units Severity/Intensity Char Char * *

Derived

Result Qualifier Variable Qualifier Record Qualifier

Used for continuous or numeric results or findings in standard Exp format; copied in numeric format from PESTRESC. PESTRESN should store all numeric test results or findings. Standardized unit used for PESTRESC or PESTRESN. The severity or intensity of the exam finding. PESEV is only used to qualify the results of an observation -- it should not be used to record the answer to the question posed in PETEST.. Example: MODERATE could be used to describe a PEORRES of ACNE. Exp Perm

SDTMIG 4.1.5.1

PESTRESU PESEV

Derived CRF

PESTAT PEREASND

Examination Status Reason Not Examined

Char Char

**NOT DONE

CRF or Derived CRF or Derived

Record Qualifier Record Qualifier

Used to indicate exam not done. Should be null if a result exists Perm in ORRES. Describes why an examination was not performed or why a body system was not examined. Example: SUBJECT REFUSED. Used in conjunction with STAT when value is NOT DONE. Perm

SDTMIG 4.1.5.1

CDISC, 2005. All rights reserved FINAL

Page 77 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name PEBLFL PEEVAL

Variable Label Baseline Flag Evaluator

Type Char Char

Controlled Terms or Format

Origin

Role Record Qualifier Record Qualifier

CDISC Notes Indicator used to identify a baseline value.

Core Exp

References SDTMIG 4.1.3.7

**Y or Null CRF or Derived * CRF

Role of the person who provided the evaluation. Used only for Perm results that are subjective (e.g., assigned by a person or a group). Should be null for records that contain collected or derived data. Examples: INVESTIGATOR, ADJUDICATION COMMITTEE, VENDOR. 1. Protocol-defined description of clinical encounter. 2. May be used in addition to VISITNUM and/or VISITDY. 1. Clinical encounter number. 2. Numeric version of VISIT, used for sorting. Perm Exp Perm Exp 1. Study day of physical exam, measured as integer days. 2. Algorithm for calculations must be relative to the sponsordefined RFSTDTC variable in Demographics. This formula should be consistent across the submission. Perm SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 4.1.4.1 SDTMIG 4.1.4.4

VISIT VISITNUM VISITDY PEDTC PEDY

Visit Name Visit Number

Char Num

CRF or Derived CRF or Derived CRF or Derived ISO 8601 CRF or Derived Derived

Timing Timing Timing Timing Timing

Planned Study Day of Visit Num Date/Time of Examination Char Study Day of Examination Num

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.
6.3.4.1 Assumptions for Physical Examinations (PE) domain model 1. The PE domain provides an example where the result, PEORRES, is coded. This is in contrast to other domains (e.g., AE, CM, and MH), in which the topic variable (AETERM, CMTRT, and EXTRT, respectively) is the one coded.

Page 78 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

6.3.5

Questionnaires QS

qs.xpt, Questionnaires Findings, Version 3.1.1, August 26, 2005. One record per question per time point per visit per subject, Tabulation
Variable Name STUDYID DOMAIN USUBJID QSSEQ QSGRPID QSSPID Variable Label Study Identifier Domain Abbreviation Unique Subject Identifier Sequence Number Group ID Sponsor-Defined Identifier Type Char Char Char Num Char Char Controlled Terms or Format **QS Origin CRF Derived Sponsor Defined CRF or Derived Sponsor Defined Sponsor Defined * CRF or Derived Role CDISC Notes Core Req Req Req Req References SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTMIG 2.1; SDTM 2.2.4 SDTM 2.2.4

Identifier Unique identifier for a study within the submission. Identifier Two-character abbreviation for the domain most relevant to the observation. Identifier Unique subject identifier within the submission. Identifier Sequence number given to ensure uniqueness within a domain.

QSTESTCD

Question Short Name

Char

QSTEST

Question Name

Char

CRF or Derived * * Sponsor Defined Sponsor Defined CRF or Derived CRF or Derived

QSCAT QSSCAT QSORRES QSORRESU

Category of Question Subcategory for Question Finding in Original Units Original Units

Char Char Char Char

Identifier Used to tie together a block of related records in a single domain to Perm support relationships within the domain and between domains. Identifier Optional Sponsor-defined reference number. Perhaps pre-printed on Perm the CRF as an explicit line identifier or defined in the sponsors operational database. Example: Question number on a questionnaire. Topic Topic variable for QS. Short name for the value in QSTEST, which Req can be used as a column name when converting the dataset from a vertical format to a horizontal format. The value in QSTESTCD cannot be longer than 8 characters, nor can it start with a number (e.g.'1TEST'). QSTESTCD cannot contain characters other than letters, numbers, or underscores. Examples: COG01, GH1, PF1. Synonym Verbatim name of the question or group of questions used to obtain Req Qualifier the measurement or finding. The value in QSTEST cannot be longer than 40 characters. Examples: In General, Would You Say your Health Is. Grouping Used to define a category of related records that will be meaningful Req Qualifier to the Reviewer. Examples: HAMILTON DEPRESSION SCLAE, SF36, ADAS. Grouping A further categorization of the questions within the category. Perm Qualifier Examples: MENTAL HEALTH DOMAIN, DEPRESSION DOMAIN, WORD RECALL. Result Finding as originally received or collected (e.g. RARELY, Exp Qualifier SOMETIMES). Variable Original units in which the data were collected. The unit for Perm Qualifier QSORRES, such as minutes or seconds or the units associated with a visual analog scale.

SDTMIG 2.1 SDTMIG 2.1 SDTMIG 4.1.5.1

CDISC, 2005. All rights reserved FINAL

Page 79 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name QSSTRESC

Variable Label

Type

Controlled Terms or Format

Origin CRF or Derived

Role

CDISC Notes

Core

References SDTMIG 4.1.5.1

Character Result/Finding in Char Std Format

QSSTRESN QSSTRESU QSSTAT QSREASND QSBLFL QSDRVFL

Numeric Finding in Standard Num Units Standard Units Status of Question Reason Not Performed Baseline Flag Derived Flag Char Char Char Char Char **Y or Null **Y or Null * **NOT DONE

Derived CRF or Derived CRF or Derived CRF or Derived CRF or Derived CRF or Derived

VISIT VISITNUM VISITDY QSDTC QSDY

Visit Name Visit Number

Char Num

Planned Study Day of Visit Num Date/Time of Finding Study Day of Finding Char Num ISO 8601

CRF or Derived CRF or Derived CRF or Derived CRF or Derived Derived

Result Contains the finding for all questions or sub-scores, copied or Exp Qualifier derived from QSORRES in a standard format or standard units. QSSTRESC should store all findings in character format; if findings are numeric, they should also be stored in numeric format in QSSTRESN. If question scores are derived from the original finding, then the standard format is the score. Examples: 0, 1. Result Used for continuous or numeric findings in standard format; copied Perm Qualifier in numeric format from QSSTRESC. QSSTRESN should store all numeric results or findings. Variable Standardized unit used for QSSTRESC or QSSTRESN. Perm Qualifier Record Used to indicate a questionnaire or response to a questionnaire was Perm Qualifier not done. Should be null if a result exists in QSORRES. Record Describes why a question was not answered. Example: subject Perm Qualifier refused. Used in conjunction with QSSTAT when value is NOT DONE. Record Indicator used to identify a baseline value. Exp Qualifier Record The value should be Y or null. Records which represent the average Perm Qualifier of other records or questionnaire sub-scores which do not come from the CRF are examples of records which would be derived for the submission datasets. If QSDRVFL=Y, then QSORRES should be NULL, with QSSTRESC and (if numeric) QSSTRESN having the derived value. Timing 1. Protocol-defined description of clinical encounter. Perm 2. May be used in addition to VISITNUM and/or VISITDY. Timing 1. Clinical encounter number. Exp 2. Numeric version of VISIT, used for sorting. Timing Perm Timing Timing Exp 1. Study day of finding collection, measured as integer days. Perm 2. Algorithm for calculations must be relative to the sponsor-defined RFSTDTC variable in Demographics. This formula should be consistent across the submission. 1. Text Description of time when questionnaire is administered. Perm 2. This may be represented as an elapsed time relative to a fixed reference point, such as time of last dose. See QSTPTNUM and QSTPTREF.

SDTMIG 4.1.5.1

SDTMIG 4.1.5.1

SDTMIG 4.1.3.7 SDTMIG 4.1.5.1 SDTMIG 4.1.3.7

SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 7.8; SDTMIG 4.1.5 SDTMIG 4.1.4. SDTMIG 4.1.4.4

QSTPT

Planned Time Point Name

Char

CRF or Derived

Timing

Page 80 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name QSTPTNUM QSELTM

Variable Label

Type

Controlled Terms or Format

Origin CRF or Derived Derived

Role Timing Timing

CDISC Notes Numerical version of QSTPT to aid in sorting. Elapsed time (in ISO 8601) relative to a planned fixed reference (QSTPTREF). This variable is useful where there are repetitive measures. Not a clock time or a date time variable. Examples: '-

Core Perm Perm

References

Planned Time Point Number Num Elapsed Time from Reference Point Char ISO 8601

SDTMIG 4.1.4.3

P15M' to represent the period of 15 minutes prior to the reference point indicated by QSTPTREF, or 'P8H' to represent the period of 8 hours after the reference point indicated by QSTPTREF.
QSTPTREF Time Point Reference Char Sponsor Defined ISO 8601 CRF or Derived Timing Name of the fixed reference point referred to by QSELTM, Perm QSTPTNUM, and QSTPT. Examples: PREVIOUS DOSE, PREVIOUS MEAL. Evaluation Interval associated with a QSTEST question represented Perm in ISO 8601 character format. Example: "P2Y" to represent an interval of 2 years in the question "Have you experienced any episodes in the past 2 years?"

QSEVLINT

Evaluation Interval

Char

Timing

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.
6.3.5.1 Assumptions for Questionnaire (QS) domain model 1. Questionnaire data may include, but are not limited to the following: subject reported outcomes, validated or non-validated questionnaires, and validated or non-validated diaries (i.e., data with a common topicality). When objective numeric data with result qualifiers are collected in a questionnaire or diary format, the sponsor should consider whether this data actually belongs in a separate (new) domain. It is realized that questionnaires may include some objective questions and some of these may trigger additional data collection in other domains. 2. The names of the questionnaires should be described under the variable QSCAT in the questionnaire domain. For example, ADAS for Alzheimer's Disease Assessment Scale, SF36 for SF-36 Health Survey, PANSS for Positive and Negative Syndrome Scale, etc. 3. Names of categories/subcategories for groups of items/questions should be described under QSSCAT. Derived information such as, total scores, and sub scores, etc, should be stored in the QS domain as derived records with appropriate category/subcategory names (QSCAT/QSSCAT), item names (QSTESTCD/QSTEST), and results (QSSTRESN/QSSTRESC). Derived records should be flagged by QSDRVFL. .

CDISC, 2005. All rights reserved FINAL

Page 81 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

6.3.6

Subject Characteristics SC

sc.xpt, Subject Characteristics Findings, Version 3.1.1, August 26, 2005. One record per characteristic per subject, Tabulation
Variable Name STUDYID DOMAIN USUBJID SCSEQ SCGRPID SCSPID SCTESTCD Variable Label Study Identifier Domain Abbreviation Unique Subject Identifier Sequence Number Group ID Sponsor-Defined Identifier Type Char Char Char Num Char Char * Controlled Terms or Format **SC Origin CRF Derived Sponsor Defined CRF or Derived Sponsor Defined Sponsor Defined CRF or Derived Role Identifier Identifier Identifier Identifier Identifier Identifier Topic CDISC Notes Unique identifier for a study within the submission. Two-character abbreviation for the domain most relevant to the observation. Unique subject identifier within the submission. Sequence number given to ensure uniqueness within a dataset for a subject. Can be used to join related records. Used to tie together a block of related records in a single domain to support relationships within the domain and between domains. Core Req Req Req Req Perm References SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTMIG 2.1; SDTM 2.2.4 SDTM 2.2.4

Subject Characteristic Short Char Name

Sponsor-defined reference number. Perhaps pre-printed on the CRF as Perm an explicit line identifier or defined in the sponsors operational database. Example: Question number on a questionnaire. Short name of the measurement, test, or examination described in Req SCTEST. It can be used as a column name when converting a dataset from a vertical to a horizontal format. The value in SCTESTCD cannot be longer than 8 characters, nor can it start with a number (e.g.'1TEST'). SCTESTCD cannot contain characters other than letters, numbers, or underscores. Example: SUBJINIT, EYECD. Verbatim name of the test or examination used to obtain the Req measurement or finding. The value in SCTEST cannot be longer than 40 characters. Examples: Subject Initials, Eye Color. Used to define a category of related records. Perm A further categorization of the subject characteristic. Result of the subject characteristic as originally received or collected. Original Unit in which the data were collected. The unit for SCORRES. Perm Exp Perm

SCTEST

Subject Characteristic

Char

CRF

Synonym Qualifier Grouping Qualifier Grouping Qualifier Result Qualifier Variable Qualifier

SCCAT SCSCAT SCORRES

Category for Subject Characteristic Subcategory for Subject Characteristic

Char Char

* *

Sponsor Defined Sponsor Defined CRF or Derived CRF or Derived

SDTMIG 2.1 SDTMIG 2.1 SDTMIG 4.1.5.1

Result or Finding in Original Char Units SCORRESU Original Units Char

Page 82 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name SCSTRESC

Variable Label

Type

Controlled Terms or Format

Origin Derived

Role Result Qualifier

CDISC Notes

Core

References

Character Result/Finding in Char Std Format

SCSTRESN SCSTRESU SCSTAT

Numeric Result/Finding in Standard Units Standard Units Status of SC Measurement

Num Char Char Char Char Num ISO 8601 * **NOT DONE

Derived CRF or Derived CRF or Derived CRF or Derived CRF or Derived Derived

Result Qualifier Variable Qualifier Record Qualifier Record Qualifier Timing Timing

Contains the result value for all findings, copied or derived from Exp SDTMIG 4.1.5.1 SCORRES in a standard format or standard units. SCSTRESC should store all results or findings in character format; if results are numeric, they should also be stored in numeric format in SCSTRESN. For example, if a test has results 'NONE', 'NEG', and 'NEGATIVE' in SCORRES and these results effectively have the same meaning, they could be represented in standard format in SCSTRESC as 'NEGATIVE'. For other examples, see general assumptions. Used for continuous or numeric results or findings in standard format; Perm SDTMIG 4.1.5.1 copied in numeric format from SCSTRESC. SCSTRESN should store all numeric test results or findings. Standardized unit used for SCSTRESC or SCSTRESN. Perm Used to indicate that the measurement was not done. Should be null if a Perm SDTMIG 4.1.5.1 result exists in SCORRES. Describes why a question was not answered. Example: subject refused. Perm Used in conjunction with SCSTAT when value is NOT DONE. Exp 1. Study day of collection, measured as integer days. 2. Algorithm for calculations must be relative to the sponsor-defined RFSTDTC variable in Demographics. This formula should be consistent across the submission. SDTMIG 4.1.4.1

SCREASND Reason Not Performed SCDTC SCDY Date/Time of Collection Study Day of Examination

Perm SDTMIG 4.1.4.4

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.
6.3.6.1 Assumptions for Subject Characteristics (SC) domain model 1. Subject Characteristics is for data not collected in other domains that is subject-related. Examples: subject initials, other race details, eye color, childbearing status, etc. 2. The structure for demographic data is fixed and includes date of birth, age, sex , race, ethnicity and country. The structure of subject characteristics is based on the Findings general class and is an extension of the demographics data. Subject Characteristics consists of data that is collected once per subject (per test) and that is not expected to change during the trial. Subject characteristics contains data such as additional information about race, subject initials, economic information, and eye color.

CDISC, 2005. All rights reserved FINAL

Page 83 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

6.3.7

Vital Signs VS

vs.xpt, Vital Signs Findings, Version 3.1.1, August 26, 2005. One record per vital sign measurement per time point per visit per subject, Tabulation
Variable Name STUDYID DOMAIN USUBJID VSSEQ VSGRPID VSSPID Variable Label Study Identifier Domain Abbreviation Unique Subject Identifier Sequence Number Group ID Sponsor-Defined Identifier Type Char Char Char Num Char Char ** Controlled Terms or Format **VS Origin CRF Derived Sponsor Defined CRF or Derived Sponsor Defined Sponsor Defined CRF or Derived Role Identifier Identifier CDISC Notes Core Req Req Req Req Perm Perm Req Reference SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4 SDTMIG 2.1; SDTM 2.2.4

VSTESTCD Vital Signs Test Short Name Char

VSTEST

Vital Signs Test Name

Char

**

CRF

VSCAT VSSCAT

Category for Vital Signs

Char

* *

Sponsor Defined Sponsor Defined

Subcategory for Vital Signs Char

Unique identifier for a study within the submission. Two-character abbreviation for the domain most relevant to the observation. Identifier Unique subject identifier within the submission. Identifier Sequence number given to ensure uniqueness within a dataset for a subject. Can be used to join related records. Identifier Used to tie together a block of related records in a single domain to support relationships within the domain and between domains. Identifier Optional Sponsor-defined reference number. Perhaps pre-printed on the CRF as an explicit line identifier or defined in the sponsors operational database. Topic Short name of the measurement, test, or examination described in VSTEST. It can be used as a column name when converting a dataset from a vertical to a horizontal format. The value in VSTESTCD cannot be longer than 8 characters, nor can it start with a number (e.g.'1TEST'). VSTESTCD cannot contain characters other than letters, numbers, or underscores. Examples: SYSBP, DIABP, BMI. Synonym Verbatim name of the test or examination used to obtain the Qualifier measurement or finding. The value in VSTEST cannot be longer than 40 characters. Examples: Systolic Blood Pressure, Diastolic Blood Pressure, Body Mass Index. Grouping Used to define a category of related records. Qualifier Grouping A further categorization of a measurement or examination. Qualifier

Req

SDTMIG 2.1

Perm Perm

SDTMIG 2.1 SDTMIG 2.1

Page 84 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name VSPOS

Variable Label

Type

Controlled Terms or Format *

Origin CRF or Derived CRF or Derived

Role Record Qualifier Result Qualifier Variable Qualifier Result Qualifier

CDISC Notes

Core

Reference

Vital Signs Position of Char Subject VSORRES Result or Finding in Original Char Units VSORRESU Original Units Char VSSTRESC Character Result/Finding in Char Std Format

CRF or Derived Derived

VSSTRESN VSSTRESU VSNRIND

Numeric Result/Finding in Standard Units Standard Units Reference Range Indicator

Num Char Char * *

Derived CRF or Derived CRF or Derived

Result Qualifier Variable Qualifier Variable Qualifier

Position of the subject during a measurement or examination. Perm Examples: SUPINE, STANDING, SITTING. Result of the vital signs measurement as originally received or Exp collected. Original units in which the data were collected. The unit for Exp VSORRES. Examples: INCHES, FEET, POUNDS, BEATS PER MINUTE. Contains the result value for all findings, copied or derived from Exp VSORRES in a standard format or standard units. VSSTRESC should store all results or findings in character format; if results are numeric, they should also be stored in numeric format in VSSTRESN. For example, if a test has results 'NONE', 'NEG', and 'NEGATIVE' in VSORRES and these results effectively have the same meaning, they could be represented in standard format in VSSTRESC as 'NEGATIVE'. For other examples, see general assumptions. Used for continuous or numeric results or findings in standard Exp format; copied in numeric format from VSSTRESC. VSSTRESN should store all numeric test results or findings. Standardized unit used for VSSTRESC or VSSTRESN. Exp Perm

SDTMIG 4.1.5.1

SDTMIG 4.1.5.1

SDTMIG 4.1.5.1 SDTMIG 4.1.5.1

VSSTAT

Vitals Status

Char Char Char Char

**NOT DONE

CRF or Derived CRF or Derived

VSREASND Reason Not Performed VSLOINC VSLOC LOINC Code Location of Vital Signs Measurement

** *

Derived CRF or Derived

1. Indicates where value falls with respect to reference range (which could be defined by VSORNRLO and VSORNRHI or by VSSTNRC (if those variables are used). Examples: NORMAL, ABNORMAL, HIGH, LOW. 2. Should not be used to indicate clinical significance. Record Used to indicate that a vital sign measurement was not done. Qualifier Should be null if a result exists in VSORRES. Record Describes why a measurement or test was not performed. Qualifier Examples: BROKEN EQUIPMENT or SUBJECT REFUSED. Used in conjunction with VSSTAT when value is NOT DONE. Synonym 1. LOINC Code for VSTEST. Qualifier 2. The LOINC version should be provided in the Sponsor Comments column in the data definition document. Record Location relevant to the collection of Vital Signs measurement. Qualifier Example: LEFT ARM for blood pressure.

Perm Perm Perm Perm

SDTMIG 4.1.5.1 SDTMIG 4.1.5.1 SDTMIG 4.1.3.2 SDTMIG 4.1.3.7

CDISC, 2005. All rights reserved FINAL

Page 85 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Variable Name VSBLFL VSDRVFL

Variable Label Baseline Flag Derived Flag

Type Char Char

Controlled Terms or Format

Origin

Role Record Qualifier Record Qualifier

CDISC Notes Indicator used to identify a baseline value

Core Exp

Reference SDTMIG 4.1.3.7 SDTMIG 4.1.5.1 SDTMIG 4.1.3.7

**Y or Null CRF or Derived **Y or Null Derived

VISIT VISITNUM VISITDY VSDTC VSDY

Visit Name Visit Number

Char Num

CRF or Derived CRF or Derived CRF or Derived ISO 8601 CRF or Derived Derived

Timing Timing Timing Timing Timing

Planned Study Day of Visit Num Date/Time of Measurements Char Study Day of Vital Signs Num

Used to indicate a derived record. The value should be Y or null. Perm Records which represent the average of other records or questionnaire sub-scores which do not come from the CRF are examples of records which would be derived for the submission datasets. If VSDRVFL=Y, then VSORRES should be NULL, with VSSTRESC and (if numeric) VSSTRESN having the derived value. 1. Protocol-defined description of clinical encounter. Perm 2. May be used in addition to VISITNUM and/or VISITDY. 1. Clinical encounter number. Req 2. Numeric version of VISIT, used for sorting. Perm 1. Study day of vital signs measurements, measured as integer days. 2. Algorithm for calculations must be relative to the sponsordefined RFSTDTC variable in Demographics. This formula should be consistent across the submission. 1. Text Description of time when measurement should be taken. 2. This may be represented as an elapsed time relative to a fixed reference point, such as time of last dose. See VSTPTNUM and VSTPTREF. Examples: Start, 5 min post. Numerical version of VSTPT to aid in sorting. Elapsed time (in ISO 8601) relative to a planned fixed reference (VSTPTREF). This variable is useful where there are repetitive measures. Not a clock time or a date time variable. Examples: 'Exp Perm

SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 7.8; SDTMIG 4.1.4.5 SDTMIG 4.1.4.1 SDTMIG 4.1.4.4

VSTPT

Planned Time Point Name

Char

CRF or Derived

Timing

Perm

VSTPTNUM Planned Time Point Number Num VSELTM Elapsed Time from Char Reference Point

ISO 8601

CRF or Derived Derived

Timing Timing

Perm Perm

SDTMIG 4.1.4.3

P15M' to represent the period of 15 minutes prior to the reference point indicated by VSTPTREF, or 'P8H' to represent the period of 8 hours after the reference point indicated by VSTPTREF.
VSTPTREF Time Point Reference Char Sponsor Defined Timing Name of the fixed reference point referred to by VSELTM, VSTPTNUM, and VSTPT. Examples: PREVIOUS DOSE, PREVIOUS MEAL. Perm

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.

Page 86 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

7 Trial Design Datasets

7.1 INTRODUCTION
The Trial Design Model, which is included in the SDTM, allows description of key aspects of the planned conduct of a clinical trial in a standardized way. These standardized descriptions will allow reviewers to: clearly and quickly grasp the design of a clinical trial compare the designs of different trials search a data warehouse for clinical trials with certain features compare planned and actual treatments and visits for subjects in a clinical trial.

Modeling a clinical trial in this standardized way requires the explicit statement of certain decision rules that may not be addressed or may remain vague or ambiguous in the usual prose protocol document. Prospective modeling of the design of a clinical trial should lead to a clearer, better protocol. Retrospective modeling of the design of a clinical trial should ensure a clear description of how the trial was interpreted by the sponsor. The aspects of clinical trial design currently included in the Trial Design Model are as follows: the planned arms of the trial what happens to a subject in each arm (i.e., what series of treatment and non-treatment time periods [trial Elements] are planned for a subject assigned to that arm) the planned schedule of visits the inclusion and exclusion criteria for the trial.

Future versions of the Trial Design Model are expected to include additional aspects of clinical trials, such as planned exposures and planned assessments. The Trial Design Model is built upon the concepts of Elements, Arms, Epochs and Visits. An Element is the basic building block for time within a clinical trial, and has the following characteristics: A description of what happens to the subject during the Element. A definition of the start of the Element. A rule for ending the Element.

An Arm is a planned sequence of Elements, typically equivalent to a treatment group. Branches may take place between one Element and the next and some designs allow for some flexibility of Elements within an Arm. The term EPOCH (which is discussed in Section 7.4.4) is used to refer to the portion of a blinded trial that corresponds to an individual element (when the actual element may not be known). A Visit is defined as a clinical encounter that encompasses planned and unplanned trial interventions, procedures, and assessments that may be performed on a subject. A Visit has a start and an end, each described with a rule. A Visit need not be nested within an Element. In other words, it may start in one Element and end in another. In most blinded trials, the timing of Visits is the same for all subjects, regardless of the Arm to which they have been assigned. In these cases, the Arm is not needed to describe the timing of Visits, and is left blank in the Trial Visits domain. If the timing of Visits depends on Arm, then the complete set of Visits for each Arm should be represented in the domain. The Trial Design Model also includes the Trial Inclusion/Exclusion (TI) dataset to describe the inclusion/exclusion criteria used to screen subjects. The IE domain (subject specific inclusion/exclusion criteria not met) described in section 6.3.2 contains the actual exceptions to those criteria for enrolled subjects.
CDISC, 2005. All rights reserved FINAL Page 87 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

7.2

PLANNED ELEMENTS, ARMS, AND VISITS

The Trial Element table (Table 7.2.1) The Trial Arms table (Table 7.2.2) The Trial Visits table (Table 7.2.3).

Under the Trial Design Model, planned information is presented in a series of three tables:

7.2.1

Trial Elements
Type Controlled Origin Terms or Format Char Char **TE CRF Derived Role CDISC Notes Core

te.xpt, Trial Elements, Version 3.1.1, August 26, 2005. One record per Planned Element
Variable Variable Label Name STUDYID Study Identifier DOMAIN Domain Abbreviation

Identifier Unique identifier for a study within Req the submission. Identifier Two-character abbreviation for the Req domain most relevant to the observation. Topic Short name of ELEMENT used for Req programming and sorting. Req Req

ETCD

Element Code

Char

* *

Sponsor Defined

ELEMENT Description of Element Char TESTRL TEENRL Rule for Start of Element Char

Sponsor Defined / Protocol Synonym The name of the Element. Qualifier Sponsor Defined / Protocol Rule Sponsor Defined / Protocol Rule Expresses rule for beginning Element.

Rule for End of Element Char

Expresses rule for ending Element. Perm Either TEENRL or TEDUR must be present for each element. Planned Duration of Element in ISO 8601 format. Used when the rule for ending the Element is to applied after a fixed duration. Perm

TEDUR

Planned Duration of Element

Char

ISO 8601

Sponsor Defined / Protocol Timing

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.

Page 88 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

7.2.2
Variable Name STUDYID DOMAIN ARMCD ARM TAETORD ETCD ELEMENT

Trial Arms
Variable Label Type Controlled Origin Terms or Format Char CRF Derived Derived Derived Role CDISC Notes Core

ta.xpt, Trial Arms, Version 3.1.1, August 26, 2005. One record per Planned Element per Arm

Study Identifier

Identifier Unique identifier for a study within the submission. Identifier Two-character abbreviation for the domain most relevant to the observation. Topic Short name of ARM., used for programming and sorting.

Req Req Req Req Req Req Perm

Domain Abbreviation Char **TA Arm Code Description of Arm Order of Element within Arm Element Code Description of Element Char * Char * Num Char * Char * Char

Synonym Name given to an Arm or treatment group. Qualifier

Sponsor Defined Identifier Number that gives the order of the Element within the Arm. Sponsor Defined Record Short name of ELEMENT used for Qualifier programming and sorting. Sponsor Defined / Synonym The name of the Element. Protocol Qualifier Rule

TABRANCH Branch

Condition subject met, at a 'branch' in the trial Exp design at the end of this Element, to be included in this Arm; (e.g., randomization to Drug A). If the trial design allows a subject to transition Exp to an Element other than the next Element in sequence, then the conditions for transitioning to those other Elements, and the alternative Element sequences, are specified in this rule (e.g., responders go to washout). Name of the Trial Epoch with which this Element of the Arm is associated. Perm

TATRANS

Transition Rule

Char

Rule

EPOCH

Trial Epoch

Char *

Timing

Note: The same Element may occur more than once within an Arm. * indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.

CDISC, 2005. All rights reserved FINAL

Page 89 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

7.2.3

Trial Visits
Type Controlled Origin Terms or Format Char **TV CRF Derived Role CDISC Notes Core

tv.xpt, Trial Visits, Version 3.1.1, August 26, 2005. One Record per Planned Visit per Arm
Variable Variable Label Name STUDYID DOMAIN Study Identifier

Identifier Unique identifier for a study within the submission. Identifier Two-character abbreviation for the domain most relevant to the observation Topic 1. Clinical encounter number 2. Numeric version of VISIT, used for sorting.

Req Req

Domain Abbreviation Char

VISITNUM Visit Number

Num

CRF or Derived

Req

VISIT

Visit Name

Char

CRF or Derived

Synonym 1. Protocol-defined description of clinical encounter. Qualifier 2. May be used in addition to VISITNUM and/or VISITDY as a text description of the clinical encounter. Timing 1. Planned study day of VISIT. 2. Due to its sequential nature, used for sorting.

Perm

VISITDY

Planned Study Day of Num Visit

CRF or Derived

Perm

ARMCD

Arm Code

Char

Derived

Record 1 Short name of ARM, used for Qualifier programming and sorting. 2. If the timing of visits for a trial does not depend on which ARM a subject is in, then ARMCD should be null.

Exp

ARM

Description of Arm

Char

Derived

Synonym 1. Name given to an Arm or Perm Qualifier Treatment Group. 2. If the timing of Visits for a trial does not depend on which Arm a subject is in, then Arm should be left blank. Rule Rule describing when the Visit starts, Perm in relation to the sequence of Elements. Used only when Visits are dependent on occurrences within the study, not fixed by protocol. Rule describing when the Visit ends, Perm in relation to the sequence of Elements.

TVSTRL

Visit Start Rule

Char

Derived

TVENRL

Visit End Rule

Char

Sponsor Defined / Protocol

Rule

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.

Page 90 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

7.3

SUBJECT ELEMENTS AND VISITS

While the Trial Elements, Trial Arms and Trial Visits datasets describe the planned design of the study, it is also necessary to collect the corresponding actual data. Subject assignment to an Arm is captured in the ARM variable in Demographics. Actual Elements and Visits data for each subject are described in two additional datasets: The Subject Elements dataset (Table 7.3.1) The Subject Visits dataset (Table 7.3.2).

7.3.1

Subject Elements
Type Controlled Origin Terms or Format Char Char Char Num **SE CRF Derived Sponsor Defined Derived Role CDISC Notes Core

se.xpt ,Subject Elements, Version 3.1.1, August 26, 2005. One record per Actual Element per Subject
Variable Variable Label Name STUDYID Study Identifier DOMAIN Domain Abbreviation USUBJID Unique Subject Identifier SESEQ Sequence Number

Identifier Unique identifier for a study within the submission.

Req

Identifier Two-character abbreviation for the Req domain most relevant to the observation. Identifier Unique subject identifier within the submission. Identifier Sequence number given to ensure uniqueness within dataset. Can be used to join related records. Topic 1. Short name of ELEMENT, used for programming and sorting. 2. If an encountered Element differs from the planned Element to the point that it is considered a new Element, then use 'UNPLAN' as the value for ETCD to represent this Element. Req Req

ETCD

Subject Element Code

Char

Sponsor Defined

Req

ELEMENT Description of Subject Char Element

Sponsor Defined / Protocol

Synonym The name of the Element. If an Perm Qualifier encountered Element differs from the planned ELEMENT to the point that it is considered a new ELEMENT, then ELEMENT should be Null Timing Timing Start date/time for an Element for each subject. End date/time of an Element for each subject. Exp Exp Perm

SESTDTC Start Date/ Time of Element SEENDTC End Date/ Time of Element SEUPDES Description of Unplanned Element

Char Char Char

ISO 8601 ISO 8601

CRF or Derived CRF or Derived Sponsor Defined

Synonym Description of what happened to the Qualifier subject during this unplanned Element. Used only if ETCD has the value of 'UNPLAN'.

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.

CDISC, 2005. All rights reserved FINAL

Page 91 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

7.3.2

Subject Visits
Type Controlled Origin Terms or Format Char **SV CRF Derived Role CDISC Notes Core

sv.xpt, Subject Visits, Version 3.1.1, August 26, 2005. One record per Subject per Actual Visit
Variable Variable Label Name STUDYID DOMAIN Study Identifier

Identifier Unique identifier for a study within Req the submission. Identifier Two-character abbreviation for the domain most relevant to the observation. Req

Domain Abbreviation Char

USUBJID

Unique Subject Identifier

Char Num

Sponsor Defined CRF or Derived

Identifier Unique subject identifier within the Req submission. Topic 1. Clinical encounter number. Req (Decimal numbering may be useful for inserting unplanned visits.) 2. Numeric version of VISIT, used for sorting.

VISITNUM Visit Number

VISIT

Visit Name

Char

CRF or Derived

Synonym 1. Protocol-defined description of Qualifier clinical encounter. 2. May be used in addition to VISITNUM and/or VISITDY as a text description of the clinical encounter.

Perm

VISITDY SVSTDTC

Planned Study Day of Num Visit Start Date/Time of Visit Char Char Char ISO 8601 ISO 8601

CRF or Derived CRF or Derived CRF or Derived CRF or Derived

Timing Timing Timing

Planned study day of VISIT. Perm Start date/time for a Visit. End date/time of a Visit. Exp Exp

SVENDTC End Date/Time of Visit SVUPDES Description of Unplanned Visit

Synonym Description of what happened to the Perm Qualifier subject during an unplanned visit.

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.

Page 92 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

7.4

TRIAL ARMS

The core of the Trial Design models is the Trial Arms domain (TA). It contains one record for each occurrence of an Element in each Arm of the trial.

7.4.1

Identifying Trial Arms

Arms are a familiar concept in Clinical Trials. They usually correspond to treatment groups.

Example: Figure 1

Flowchart, Parallel Trial

Placebo Screen Run-in Drug A Drug B

Figure 1

In this example, the number of Arms is three. However, in some trials with multiple branchings, the number of Arms in the trial may not be obvious.

CDISC, 2005. All rights reserved FINAL

Page 93 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Example: Figure 2

Counting Arms in More Complex Trials

Treatment A Screen Treatment B Rescue
Randomization Response Evaluation

Open label Rescue Open label

Figure 2

In this example, subjects are randomized to one of two treatment groups. At the end of this first treatment period, they are assigned to one of two further treatment groups on the basis of their response to their initial period of treatment. The combination of these two major branchings results in four Arms for this trial. See Section 7.4.7 for additional considerations that arise in a multi-stage trial design like this one. See Section 7.4.3 for additional discussion of when a decision point in a trial design should be considered to give rise to a new Arm.

7.4.2

Developing the Trial Arms Table

In Figure 1 (shown previously), the trial has three Arms. This diagram is similar to the kind of study schema or flowchart usually included in most trial protocols. The 'building blocks' shown in these diagrams are referred to within the Trial Design Model as 'Elements.' Further details on how Elements are defined are in Section 7.5 Parallel Trial Example: Figure 3

Page 94 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Arms Representation, Parallel Trial

Screen Screen Screen Run-in Run-in Run-in Placebo Drug A Drug B

Randomization

Figure 3

The trial design can be represented as three parallel Arms, as in this diagram. The screen and run-in Elements, which occur before the branching at the point of randomization, are repeated in each Arm. Note that the randomization is represented by red lines at the ends of the run-in Elements in the three Arms. Randomizations, and other branchings in the trial design, are represented by such markers between Elements. One kind of branching, other than a randomization, is an assignment based on the subjects clinical response to treatment. This was mentioned in the example in Section 7.4.1. Branchings could also depend on a subjects disease severity, or any other clinical finding. Another kind of branching occurs in dose-escalation cohort studies. In this kind of trial design, there are rules that say whether the next subject should be in the current cohort (receive the current dose level) depending on how many subjects are in the current cohort and what their response to treatment has been. For example, the rule might say that five subjects should be entered in each cohort, and that additional subjects would be added to a cohort if a certain proportion of subjects experienced dose-limiting toxicity. The rule might say that entry of subjects (and the addition of new cohorts) would stop when a certain proportion of subjects in such an enlarged cohort experiences doselimiting toxicity.

CDISC, 2005. All rights reserved FINAL

Page 95 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

The diagram in Figure 3 can be further transformed into a table to aid in creating the trial design datasets. There is one row in the table for each Element in each Arm in Figure 3. Note how the column labeled 'Branch' captures the information represented by the red markers in the diagram. Note also that the value in the Branch column shows the randomization result that is needed for a subject to be in that Arm.
ARMCD Placebo Placebo Placebo A A A B B B TAETORD 1 2 3 1 2 3 1 2 3 ELEMENT Screen Run-In Placebo Screen Run-In Drug A Screen Run-In Drug B Randomized to Drug B Randomized to Drug A Randomized to Placebo TABRANCH

Crossover Trial Example: Figure 4

Flowchart, Crossover Trial

Placebo Rest Screen 5 mg 5 mg 5 mg Rest 10 mg Follow up Follow up

Rest Placebo Rest 10 mg Rest 10 mg

Rest Placebo Follow up

Figure 4

This trial is similar to the Parallel-Trial example, except that there are multiple treatment periods. Note that the Arms are distinguished by the order of treatments, rather than the presence of one treatment or another. Note also that rest Elements occur twice within each Arm.

Page 96 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

CrossOver Trial Example: Figure 5

Arms Representation, Crossover Trial

Screen Placebo Rest 5 mg Screen Screen 5 mg 5 mg Rest 10 mg Follow up

Rest Placebo Rest 10 mg Follow up Rest 10 mg Rest Placebo Follow up

Randomization

Figure 5

The trial design is represented with three parallel Arms. This transformation is similar to that shown in the Parallel Trial example (previously shown Figures1-3). As for the previous example, the diagram in Figure 5 can be further transformed into a table. Only the first 10 of the 21 rows of the table for this trial are shown in this example table.
ARM P-5-10 P-5-10 P-5-10 P-5-10 P-5-10 P-5-10 P-5-10 5-P-10 5-P-10 5-P-10 TAETORD 1 2 3 4 5 6 7 1 2 3 ELEMENT Screen Placebo Rest 5 mg Rest 10 mg Follow-up Screen 5 mg Rest Randomized to 5 mg - Placebo - 10 mg TABRANCH Randomized to Placebo - 5 mg - 10 mg

CDISC, 2005. All rights reserved FINAL

Page 97 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Oncology Trial Example: Figure 6

Flowchart, Oncology Trial

Repeat until disease progression

Trt A Rest Screen Trt B Rest

Follow up Follow up

Figure 6

Many oncology trials feature repeating cycles of treatment. As indicated, there is a condition for repeating the cycle.

Page 98 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Oncology Trial Example: Figure 7

Flowchart with Repeats, Oncology Trial

Trt A Rest Trt A Rest Trt A Rest Trt A Rest Follow up Screen Trt B Rest Trt B Rest Trt B Rest Trt B Rest Follow up
Note: If protocol does not specify maximum number of cycles, use actual maximum number of cycles that occurred in trial.

Figure 7

The first step in transforming this design into tabular format is to put in the repeating cycles, replacing the arrow indicating cycling through the treatment Element. If the protocol indicates a maximum number of permitted cycles, that number of cycles should be included. If the protocol does not specify a maximum, then use maximum number of cycles that actually occurred in the trial. At the time of submission, the trial has completed, and the maximum number of cycles is known.

CDISC, 2005. All rights reserved FINAL

Page 99 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Oncology Trial Example: Figure 8

Arm Representation, Oncology Trial

Transition rules: If disease progresses, skip forward

Screen Trt A Rest Trt A Rest Trt A Rest Trt A Rest Follow up

Screen Trt B Rest Trt B Rest Trt B Rest Trt B Rest Follow up
Randomization
Figure 8

This diagram shows parallel Arms, as before. The randomization is represented by red markers, as before. The green arrows represent the fact that subjects may not pass through all the cycles in the trial. In this case, a subject whose disease progresses will skip further treatment cycles and jump to the follow-up Element. In previous examples, the implicit rule was 'when the subject finishes one Element, move to the next.' In this trial, at the end of each Rest Element, the subject may move to one of two Elements. Thus there is a need for an explicit transition rule. As before, this diagram can be transformed into a table. However, this table requires another column, for the transition rule. If this column is blank, then the default transition rule, 'go to the next Element in sequence' is assumed. Only the rows for one Arm are shown in this example table. ARM A A A A A A A A A A TAETORD 1 2 3 4 5 6 7 8 9 10 ELEMENT Screen Trt A Rest Trt A Rest Trt A Rest Trt A Rest Follow-up If disease progression, go to 10th Element. If disease progression, go to 10th Element. If disease progression, go to 10th Element. TABRANCH Randomized to A TATRANS

Page 100 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

7.4.3

Distinguishing between Branches and Transitions

Both the Branch and Transition columns contain rules, but there are two columns because they represent two different types of rules. Branch rules represent forks in the trial flowchart, giving rise to separate Arms. The Branch rules thus appear in multiple Arms in the Trial Arms dataset. Within any one Arm, there is no choice in the value of the Branch condition. e.g., the subject must have been randomized to Arm A. Transition rules are used for choices within an Arm. In the Oncology example shown in Figure 8, subjects who receive 1, 2, 3, or 4 cycles are all considered to belong to Arm A. In modeling a trial, decisions may have to be made about whether a decision point in the flow chart represents the separation of two or more separate Arms, or simply to variations within the same Arm. This decision will depend on the comparisons of interest in the trial.

7.4.4

Trial Epoch Concept

In all the trials we have considered thus far, the several trial Arms are similar in that they have the same numbers of Elements and the same pattern of treatment and non-treatment Elements. Example: Figure 9

Epochs, Crossover Trial

1 Screen 2 1st Trt 3 1st Rest 4 2ndTrt 5 2nd Rest 6 3ndTrt 7 Follow-up

Screen Placebo Rest 5 mg Screen Screen 5 mg 5 mg

Rest 10 mg Follow up

Rest Placebo Rest 10 mg Follow up Rest 10 mg Rest Placebo Follow up

Figure 9

This diagram shows the similarity between Arms by means of dotted vertical lines that divide the trial as a whole, across all Arms, into 'Epochs.' When the trial elements form a grid in this way, the rows are Arms and the columns are Epochs. The numbers assigned to the columns correspond to TAETORD, the 'Order' column in the example tables. The names assigned to the columns are the names of the Epochs. The names suggested in this example represent familiar concepts. Only the terminology 'Epoch' is new. The great majority of clinical trials, particularly blinded clinical trials, have this grid structure of Epochs and Arms. During the conduct of a blinded study, when the Arm assignment of a subject is unknown, a subject's progress through the trial must be spoken of in terms of Epochs, since the subject's specific treatment Elements are unknown. The concept of Epochs is not universally applicable. The following examples show trials for which Epochs are applicable and are not applicable and some that fall in a gray area. In Figure 9, the Elements of the Arms fall into columns because the blocks representing the Elements in a column all have the same width. The widths of these blocks represent the Element durations, which are determined by their

CDISC, 2005. All rights reserved FINAL

Page 101 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

End Rules. (Element Start and End Rules are described in Section 7.5.) The following is a join of the TA and TE datasets for this trial (i.e., the TA dataset, to which the Start Rule (TESTRL) and Duration (TEDUR) from the TE dataset have been added).
ARM P-5-10 P-5-10 P-5-10 P-5-10 P-5-10 P-5-10 P-5-10 5-P-10 5-P-10 5-P-10 5-P-10 5-P-10 5-P-10 5-P-10 5-10-P 5-10-P 5-10-P 5-10-P 5-10-P 5-10-P 5-10-P TAETORD 1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 ELEMENT Screen Placebo Rest 5 mg Rest 10 mg Follow-up Screen 5 mg Rest Placebo Rest 10 mg Follow-up Screen 5 mg Rest 10 mg Rest Placebo Follow-up TESTRL Informed consent First dose of placebo 48 hrs after last dose drug First dose of 5 mg drug 48 hrs after last dose drug First dose of 10 mg drug 48 hrs after last dose drug Informed consent First dose of 5 mg drug 48 hrs after last dose drug First dose of placebo 48 hrs after last dose drug First dose of 10 mg drug 48 hrs after last dose drug Informed consent First dose of 5 mg drug 48 hrs after last dose drug First dose of 10 mg drug 48 hrs after last dose drug First dose of placebo 48 hrs after last dose drug TEENRL TEDUR P14D P14D 7D P14D 7D P14D P21D P14D P14D 7D P14D 7D P14D P21D P14D P14D 7D P14D 7D P14D P21D Randomized to 5 mg 10 mg - Placebo Randomized to 5 mg Placebo - 10 mg TABRANCH Randomized to Placebo - 5 mg - 10 mg

Page 102 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Sorting this dataset by TAETORD displays the similarity between Elements with the same value of TAETORD. For each value of TAETORD, the End Rules are the same (not shown with the example) and the Start Rules are either the same or differ only by drug dose. Each value of TAETORD corresponds to an Epoch, and the associated Start and End Rules of the Elements associated with that value of TAETORD represent, in effect, the Start and End Rules for the associated Epoch.
ARM 5-P-10 P-5-10 TAETORD 1 1 EPOCH Trial Screen Trial Screen ELEMENT Screen Screen TESTRL Informed consent Informed consent TEDUR P14D P14D TABRANCH Randomized to 5 mg/Placebo/10 mg Randomized to Placebo/5 mg/10 mg Randomized to 5 mg/10 mg/Placebo

5-10-P 5-P-10 P-5-10 5-10-P 5-P-10 P-5-10 5-10-P 5-P-10 P-5-10 5-10-P 5-P-10 P-5-10 5-10-P 5-P-10 P-5-10 5-10-P 5-P-10 P-5-10 5-10-P

1 2 2 2 3 3 3 4 4 4 5 5 5 6 6 6 7 7 7

Trial Screen First Treatment First Treatment First Treatment First Rest First Rest First Rest Second Treatment Second Treatment Second Treatment Second Rest Second Rest Second Rest Third Treatment Third Treatment Third Treatment Trial Follow-up Trial Follow-up Trial Follow-up

Screen 5 mg Placebo 5 mg Rest Rest Rest Placebo 5 mg 10 mg Rest Rest Rest 10 mg 10 mg Placebo Follow-up Follow-up Follow-up

Informed consent First dose of 5 mg drug First dose of placebo First dose of 5 mg drug 48 hrs after last dose drug 48 hrs after last dose drug 48 hrs after last dose drug First dose of placebo First dose of 5 mg drug First dose of 10 mg drug 48 hrs after last dose drug 48 hrs after last dose drug 48 hrs after last dose drug First dose of 10 mg drug First dose of 10 mg drug First dose of placebo 48 hrs after last dose drug 48 hrs after last dose drug 48 hrs after last dose drug

P14D P14D P14D P14D 7D 7D 7D P14D P14D P14D 7D 7D 7D P14D P14D 14D P21D P21D P21D

CDISC, 2005. All rights reserved FINAL

Page 103 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Inconsistent numbers and durations of Elements Example: Figure 10

Trial with Inconsistent Numbers and Durations of Elements

1 1 2 2 3 3 Screen Trt A Rest A 4 4 5 5 6 7 6 8 7 Trt B Rest B Follow-up
Figure 10

10 8

The trial in Figure 10 compares cancer therapies given as 3-week and 4-week cycles. The table below is a join of TA and TE datasets for this trial, sorted by TAETORD. Note that the number of Elements in Arm A is different from the number of Elements in Arm B and that the Start and End Rules for Elements with the same value of TAETORD are only loosely comparable for the 2nd through the 7th Elements. It is questionable whether it would be meaningful to compare data for the 2nd Elements through the 7th Elements, not meaningful to compare the 8th Elements, and there are no 9th and 10th Elements in Arm B for comparison It would be reasonable to compare data for the Screen and Follow-up Elements. The existing Element variable is available for selecting data for this comparison, so it is not necessary to define Epochs to facilitate this comparison.
ARM A B A B A B A B A B TAETORD 1 1 2 2 3 3 4 4 5 5 EPOCH ELEMENT Screen Screen Trt A Trt B Rest A Rest B Trt A Trt B Rest A Rest B TESTRL Informed consent Informed consent First dose A First dose B 48 hrs after last dose A 48 hrs after last dose B First dose A First dose B 48 hrs after last dose A 48 hrs after last dose B TEDUR 14D 14D 9D P7D 12D 21D 9D P7D 12D 21D If disease progresses, go to 10th Element If disease progresses, go to 8th Element If disease progresses, go to 10th Element If disease progresses, go to 8th Element TABRANCH Randomized to A Randomized to B TATRANS

Page 104 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

ARM A B A B A B A A

TAETORD 6 6 7 7 8 8 9 10

EPOCH

ELEMENT Trt A Trt B Rest A Rest B Trt A Follow-up Rest A Follow-up

TESTRL First dose A First dose B 48 hrs after last dose A 48 hrs after last dose B First dose A Decision not to treat further 48 hrs after last dose A Decision not to treat further

TEDUR 9D P7D 12D 21D 9D 90D 12D 90D

TABRANCH

TATRANS

If disease progresses, go to 10th Element

Inconsistent Element Durations Example: Figure 11

Trial with Inconsistent Element Durations

1 1 2 2 3 Screen Trt A Rest A 3 4 4 5 5 6 6 7 Trt B Rest B Follow-up
Figure 11

8 8

In the trial in Figure 11, the numbers of Elements are the same, and comparisons of the 2nd through 7th Elements of the two treatment Arms may be reasonable. The issue of whether Elements of different durations should be compared should be decided on the merits of the particular trial.

CDISC, 2005. All rights reserved FINAL

Page 105 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Dissimilar treatments: Figure 12

Trial with Dissimilar Treatments

1 Screen 1 123 Drug Treatment Behavior Modification Pre-Op Surgery Post-Op 4 Follow up A Follow up B Follow up C 2 2

Figure 12

The trial shown in Figure 12 compares dissimilar treatments with different durations. The usefulness of comparisons between data from these different treatment Elements would depend on the particular trial. If such comparisons were desired, an alternate model that uses one Surgery treatment Element (combining Pre-Op, Surgery and Post-Op in one element), instead of the three Elements shown here, might be considered. See Section 7.5.1 for additional discussion of 'lumping' versus 'splitting' in defining Elements.

7.4.5

Rules concept

The Branch and Transition columns shown in the example tables thus far are instances of a new concept being added to SDS in Version 3.1. A variable may have a Role of 'Rule.' The values of a Rule variable describe conditions for something happening. At the moment, values of Rule variables are text. At some point in the future, it is expected that these will be become executable code. Other Rule variables are present in the Trial Elements and Trial Visits domains. The value in the Branch rule gives conditions under which a subject was assigned to this Arm. A value in the Transition rule gives conditions under which a subject would move to an Element other than the next Element.

7.4.6
Variables

Recap of Trial Arms Variables

Purpose Identifiers Name the Arm Put Elements in order within the Arm Name an Element within the Arm Mark a branch in the trial design at the end of the Element. A branch may be a randomization or another method of assignment to Arms

STUDYID, DOMAIN ARM, ARMCD TAETORD ETCD, ELEMENT TABRANCH

Page 106 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Variables TATRANS EPOCH

Purpose A rule for where to go at the end of the Element. A blank represents the default rule, 'go to the next Element in sequence.' Permissible. Names the Trial Epoch with which the Element is associated.

7.4.7

Truncated Arms

In trials with multiple branch points, such as that shown in Figure 2, some subjects may drop out of the study before they reach the second branch point. This raises the question how a value of ARM should be assigned to such a subject. We recommended adopting names for the Arms of a multi-branching trial that reflect the various branches. For example, the four Arms in the trial shown in Figure 2 could be named A-Open, A-Rescue, B-Open, and B-Rescue. ARM values of A and B can be used for subjects who do not reach the second branch point. Note that it may be necessary to create an Arm name for subjects who fail to reach the branch point even for a trial with only one Branch. For example, it might be necessary to submit data on subjects who drop out during the run-in Element in the trial in Figure 1.

7.5

TRIAL ELEMENTS

Trial Elements (TE) contains one record for each type of Element in TA. It is like a key or legend for the TA domain. Note that although the same value of ELEMENT may appear in more than one record in the TA dataset, it will appear in just one record in the TE dataset. Examples: Figures 13, 14 and 15 show keys for Trial Elements added to the Trial Arms models shown previously in Figures 3, 5, and 8. Trial Elements Corresponding to Figure 3: Figure 13

Arms and Elements, Parallel Trial

Screen Screen Screen Screen Run-in Placebo Drug A Drug B
Figure 13

Run-in Run-in Run-in

Placebo Drug A Drug B 9 records in Trial Arms table

5 records in Trial Elements table

CDISC, 2005. All rights reserved FINAL

Page 107 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Trial Elements Corresponding to Figure 5: Figure 14

Arms and Elements, Crossover Trial

Screen Placebo Rest 5 mg Screen Screen Screen Placebo 5 mg 10 mg Rest Follow up
Trial Elements Corresponding to Figure 8: Figure 15
Figure 14

Rest 10 mg Follow up

5 mg 5 mg

Rest Placebo Rest 10 mg Follow up Rest 10 mg Rest Placebo Follow up 21 records in Trial Arms table

6 records in Trial Elements table

Arms and Elements, Oncology Trial

Screen Trt A Rest Trt A Rest Trt A Rest Trt A Rest Follow up Screen Trt B Rest Trt B Rest Trt B Rest Trt B Rest Follow up 20 records in Trial Arms table Screen Trt A Trt B Rest Follow up
Figure 15

5 records in Trial Elements table

Page 108 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

7.5.1

Identifying Trial Elements

An Element is a building block for creating Arms. Or, from another point of view, an Arm is composed of Elements (i.e., the trial design assigns subjects to Arms, which are comprised of steps called Elements). Trial Elements represent an interval of time in which certain activities are planned for the subjects participating in the trial. Both the activities and the time description are essential to the concept of the Element. The activities may be run-ins, wash-outs, drug treatments, or other interventions. The time description defines an interval of time, with a start and an end, during which the activities are planned to take place. Assumption: There are no gaps between Elements. The instant one Element ends, the next Element begins. A subject spends no time 'between Elements.' Since an Element is defined by both a period of time and what is happening to the subject during that time, 'Week 2 to Week 4' is not a valid Element. 'Two weeks of treatment with Drug A' is a valid Element. Elements may describe periods of non-treatment as well as periods of treatment. Common kinds of non-treatment Elements include washout, rest, and follow-up periods. Deciding how finely to divide trial time when identifying trial Elements is a matter of judgment. For example, a trial might include a dose titration, spelled out as one week each at a series of increasing doses until certain conditions are met. The trial design could be modeled in any of the following ways: using several one-week Elements at specific doses, followed by an Element of variable length at the chosen dose, as a titration Element of variable length followed by a constant dosing Element of variable length one Element with dosing determined by titration

The choice of model will depend on how the data will be analyzed and reported. If it is important to examine side effects or lab values at each individual dose, the first model is appropriate. If it is important only to identify the time to achieve titration, the second model might be appropriate. If the titration process is routine and of little interest, the third model might be adequate for the purposes of this trial. Oncology trials with cycles of treatment and rest present another example where the needs of the trial will determine whether the trial should be modeled with two Elements per cycle, as in Figure 8 (previously shown), or with one Element per cycle, as in Figure 16 below. Oncology Trial, One Element per Cycle: Figure 16

Transition rules: If disease progresses, skip forward

Screen

Trt A

Trt A

Trt A

Trt A

Follow up

Screen

Trt B

Trt B

Trt B

Trt B

Follow up

Randomization
Figure 16

CDISC, 2005. All rights reserved FINAL

Page 109 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

7.5.2

Developing the Trial Elements Table

An Elements name usually describes what happens to the subject during the Element. The time period associated with an Element is expressed in two Rules, the Element Start and Element End Rules. If the element is of fixed duration, then the Element Duration variable should also be populated. Element Start Rules identify the time point at which a subject is considered to have entered an Element. Many analyses depend on placing data in the proper Element, so the Element Start Rules should be as clear as possible, and should be chosen with consideration for planned analyses. Treatment Elements are usually considered to start with the beginning of the Element treatment (e.g., with the first dose of the relevant drug). The Start Rules for non-treatment Elements can be more difficult to define. They will often be defined relative to the end of the preceding treatment. For example, a wash-out period might be defined as starting 24 or 48 hours after the last dose of drug for the preceding treatment period. Defining a clear start point for the start of a non-treatment Element that follows another non-treatment Element can be difficult. The Element Start Rule for a treatment Element may be thought of as 'active' and the Element Start Rule for a nontreatment Element as 'passive.' The start of a treatment Element will not occur until a dose is given, no matter how long that dose is delayed. Once the last dose is given, the start of the subsequent non-treatment Element is inevitable, as long as another dose is not given. Element End Rules are rather different from Element Start Rules. The actual end of one Element is the beginning of the next Element. Thus the Element End Rule does not give the conditions under which an Element does end, but the conditions under which it should end. The most common type of Element End Rule is the rule for an Element of fixed duration. Such a rule might be expressed as '14 days after Element starts' or, perhaps more realistically, as '1216 days after Element starts.' Element End Rules may also depend on other conditions. For instance, a typical criterion for ending a rest Element between oncology chemotherapy treatment Elements would be, '15 days after start of Element and after WBC values have recovered'. There is not yet a standardized method for expressing rules in computer-executable form. However, for the most common type of Element End Rule, for elements of fixed duration, we can easily express element duration. Thus, a variable for Element Duration has been included in the model.

Page 110 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

Example: Trial Elements table for the parallel trial shown in Figure 13
ELEMENT Screen Run-In Placebo Drug A Drug B TESTRL Informed consent Eligibility confirmed First dose of placebo First dose of Drug A First dose of Drug B TEENRL 1 week after start of element 2 weeks after start of element 2 weeks after start of element 2 weeks after start of element 2 weeks after start of element TEDUR P7D P14D P14D P14D P14D

Example: Trial Elements table for the crossover trial shown in Figure 14 ELEMENT Screen Placebo 5 mg 10 mg Rest Follow-up TESTRL Informed consent First dose of placebo First dose of 5 mg drug First dose of 10 mg drug 48 hrs after last dose drug 48 hrs after last dose drug TEENRL 2 weeks after start of element 2 weeks after start of element 2 weeks after start of element 2 weeks after start of element 1 week after start of element 3 weeks after start of element TEDUR P14D P14D P14D P14D P7D P21D

The transition between one Element and the next can be complex, since it may involve meeting criteria for ending one Element, evaluating criteria for choosing which Element to go to next, and actually taking the action that initiates the new Element. For example, in the trial shown in Figure 15, the Trial Elements dataset might be as follows:
ELEMENT Screen Trt A Trt B Rest Follow-up TESTRL Consent First dose drug First dose drug Last dose + 24 hrs Decision not to treat further TEENRL 2 weeks after start of element 5 days after start of element 16 days after start of element and WBC recovers 16 days after start of element and WBC recovers 4 weeks P28D P14D P5D TEDUR

A subject in the 3rd Element (Trt B) must meet the Element End criteria (16 days pass and WBC recovers), then be evaluated to see whether disease has progressed. If it has not progressed, they must still wait until the first dose of the 4th Element is administered before they can be considered to be in the 4th Element. During the whole time from meeting the Element End criteria to meeting the new Element Start criteria, the subject stays in the 'old' Element. See Section 7.7.2 for a discussion of unplanned Elements.

CDISC, 2005. All rights reserved FINAL

Page 111 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

7.5.3
Variables

Recap of Trial Elements Variables

Purpose Identifiers Name the Element Rule describing the time point that defines the start of the Element Rule describing when the Element should end and its duration (if the element is of fixed duration)

STUDYID, DOMAIN ETCD, ELEMENT TESTRL TEENRL, TEDUR

7.5.4

Distinguishing Elements from Epochs

Figure 9, which was used to introduce the concept of trial Epoch is also useful for showing how the concepts of Elements and Epoch can be confused. Note that the natural name for the first Epoch is 'Screen,' which is also the natural name of the Element that appears in each Arm in this Epoch. The same is true for the 7th Epoch, where the Epoch and the Elements in the Epoch are both named 'Follow-up.' Clearly, it will be easy to confuse Element and Epoch in these cases. We recommend avoiding using the same names for Elements and Epochs. One possible solution is to prefix Epoch names with the word 'trial', on the grounds that an Epoch is a feature of the trial as a whole. In the example, 'Trial Screen' would be the Epoch, while 'Screen' would be the Element In the treatment Epochs of this trial, the Epoch describes a period of time in the trial, First, Second, or Third Treatment, while the Elements describe particular treatments which may occur during different Epochs of the trial. This trial has First and Second Rest Epochs, which are populated by the one Rest Element in this trial. It may be useful to think of a trial design as a table like the following, with a row for each Arm and a column for each Epoch. Trial Screen 1 P-5-10 5-P-10 5-10-P Screen Screen Screen First Treatment 2 Placebo 5 mg 5 mg First Rest 3 Rest Rest Rest Second Treatment 4 5 mg Placebo 10 mg Second Rest 5 Rest Rest Rest Third Treatment 6 10 mg 10 mg Placebo Trial Follow-up 7 Follow-up Follow-up Follow-up

The bold column headers are the names of the Epochs. The secondary column headers (1, 2, etc.) show how each value of TAETORD in this trial corresponds to an Epoch. The bold row labels are the names of the Arms. The cells are filled with the names of the Elements in the trial Arms. For some analysis purposes, one would want to group data by Element, but for other purposes, data would be grouped by Epoch or TAETORD. In this crossover trial, for instance, comparisons of incidences of side affects of the different treatments would group data by Element (Placebo vs. 5 mg vs. 10 mg). The analysis of PK results for this trial would include both treatment and period effects in the model. 'Treatment effect' derives from Element (Placebo vs. 5 mg vs. 10 mg), while 'Period effect' derives from the EPOCH (1st, 2nd, or 3rd Treatment), which is in turn associated with TAETORD (2nd vs. 4th. vs. 6th Element).

Page 112 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

7.6
7.6.1

TRIAL VISITS
Identifying Trial Visits

Trial Visits (TV) describes the planned visits in a trial.

The SDS definition of Visit is a clinical encounter. Assessments and procedures are planned at Visits. Assumption: Visits have both a start and an end. For many trials, the exact timing and duration of Visits are not important, but for other trials this detail is needed, so both start and end are included in the model. Assumption: A Visit may start in one Element and end in another. Example: Visits in clinical pharmacology trials are often planned to include a series of timed assessments (a profile) associated with a dose of drug. The dose is often the start of an Element. The planned Visit might start a hour before the dose (the end of the current Element) and end eight hours after the dose (the beginning of the new Element), to allow for pre-dose assessments and a series of post-dose assessments. The term 'Visit' reflects the fact that data in out-patient studies is usually collected when the subject Visits the clinic. Diary data and other data collected outside the clinic may not fit the usual concept of a Visit, but the planned times of collection of such data may be described as 'Visits' in the Trial Visits dataset. The description of planned 'Visits' may also require careful thought for studies of hospitalized subjects.

7.6.2

Developing the Trial Visits Table

For most trials, particularly blinded trials, the timing of Visits does not depend on Arm. However, planned Visits may be different for different Arms in the trial, so ARM and ARMCD are included in the Trial Visits dataset. If planned Visits do not depend on Arm, ARM and ARMCD can be left blank. If planned Visits do depend on Arm, records for planned Visits for each Arm should be submitted. The trial shown in Figure 12 would probably have different planned Visits for each Arm. Visit Start Rules are different from Element Start Rules because they usually describe when a Visit should occur, rather than the moment at which a Visit is considered to start. There are usually gaps between Visits, periods of time that do not belong to any Visit, so it usually not necessary to identify the moment when one Visit stops and another starts. However, some trials of hospitalized subjects may divide time into Visits in a manner more like that used for Elements. Visit Start Rules are usually expressed relative to the start or end of an Element, e.g., '1-2 hours before end of 3rd Element' or '8 weeks after end of 2nd Element.' Note that the Visit may or may not occur during the Element used as the reference for Visit Start Rule. For example, a trial with variable Elements might plan a Visit 6 months after the start of the first treatment period, regardless of the Element the subject is in at that time. Visit End Rules are similar to Element End Rules, describing when a Visit should end. They may be expressed relative to the start or end of an Element, or relative to the start of the Visit. Ranges may be used to describe the planned timing of Visits (e.g., 12-16 days after the start of 2nd Element), but this is different from the 'windows' that may be used in selecting data points to be included in an analysis associated with that Visit. For example, although Visit 2 was planned for 12-16 days after the start of treatment, data collected 1018 days after the start of treatment might be included in a 'Visit 1' analysis. The two ranges serve different purposes. Unplanned Visits are discussed further in Section 7.8.

CDISC, 2005. All rights reserved FINAL

Page 113 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Trial Visits Example: Figure 17

Trial Visits
1 Screen Screen Screen 2 Run-in Run-in Run-in 3 4 Placebo Drug A Drug B 5

Figure 17

The trial shown in Figure 17 might have the following Trial Visits dataset: VISITNUM 1 2 3 4 5 TVSTRL Start of 1 Element 30 minutes before end of 1st Element 30 minute before end of 2nd Element 1 week after start of 3rd Element 2 weeks after start of 3
rd st

TVENRL 1 hour after start of visit 30 minutes after start of 2nd Element 1 hour after start of 3rd Element 1 hour after start of visit 1 hour after start of visit

Element

7.6.3
Variables

Recap of Trial Visits Variables

Purpose Identifiers Name and order the visits Blank if visit schedule does not depend on Arm. Names of the ARM if visit schedule does depend on Arm. Rule describing when the visit should start. Usually expressed relative to starts or ends of Elements described using TAETORD.

STUDYID, DOMAIN VISIT, VISITNUM, VISITDY ARM, ARMCD TVSTRL

TVENRL

Rule describing when the visit should end. Usually expressed relative to the start of the Element, or to starts or ends of Elements described using TAETORD.

Page 114 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

7.7

SUBJECT ELEMENTS

Subject Elements (SE) describes the sequence of Elements a subject passed through, with their start and stop date/times. Note there is no need for a Subject Arms dataset, because the Arm for each subject is represented in the Demography (DM) dataset. The subject Element start date/times allow a comparison of the sequence of Elements a subject actually passed through to the planned sequence of Elements for the Arm to which they were assigned. The date/times in Subject Elements also allows any date/time to be placed within one Element. This assignment of data to actual Elements is often used in analyses. Subject Elements data involves judgment and interpretation.

7.7.1

Identifying Subject Elements

The Element Start Rules in the Trial Elements dataset should allow identification of the Elements the subject passed through and the derivation of start date/times for each subject Element. This derivation may be more or less difficult, depending on the trial design, on how well the Element Start Rules have been defined, and on how well the date/times associated with the transition from one Element to the next have been collected. Note that this process of matching the subjects actual experience with planned Elements involves judgment, and may be open to interpretation. The programming used to derive subject Elements will incorporate decisions about how close a match a time period and its treatments have to be to the plan to be considered a 'match.' For instance if the plan included Elements with treatment doses of 5 mg and 10 mg, is a period of treatment with a dose of 7 mg a '5mg' Element? If an Element is planned to last 2 weeks but actually lasts 6 weeks, is this a match? It is usually better to associate subject experience with one of the planned Elements, than to describe, note or summarize deviations from the plan, but this is a matter of judgment.

7.7.2

Unplanned Elements

It is usually better to parse a subjects time in the trial into segments labeled with names from the list of planned Elements, but occasionally a subject departs so much from plan that a decision is made to label a part of the subjects experience as an unplanned Element. Example: A crossover trial that involves three single doses of drug, one each week for three weeks. The period from one dose to the next is considered a treatment Element. A subject is enrolled in the trial and completes two of the doses. He departs for a month-long trip. On his return, he is given the third dose, then returns for follow-up a week later. It is decided that only the first week after the second dose should be considered as belonging to the second treatment Element. The remaining 3 weeks before the third dose is labeled an 'Unplanned' Element. Example: A trial involves treatment with an IV drug that is prepared in the hospital pharmacy. By error, the subject receives only one tenth of the planned dose. The error is discovered, and the subject is treated again, with the correct dose. It is decided to label the period starting with the initiation of the erroneous infusion and ending the initiation of the correct infusion an 'unplanned' Element. If an unplanned Element occurs, a record for the unplanned Element should be included in SE. For this record, ETCD cannot be populated with any of the values associated with planned Elements. It is recommended that the SE record for the unplanned Element use 'UNPLAN' as the value for ETCD, that ELEMENT be left blank, and that a description of the unplanned Element be given in SEUPDES.

7.7.3

Deriving SE End Date/Times

Because there are no gaps between Elements, once subject Element start date/times have been derived for all Elements, the derivation of all but one of the subject end date/times is trivial. For all subject Elements except the last, the end date/time of the Element is equal to the start date time of the next Element.

CDISC, 2005. All rights reserved FINAL

Page 115 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

It may be desirable to define RFENDTC (in the DM domain) as the end date/time for the end of the last subject element date/time.

7.7.4
Variables

Recap of Subject Elements Variables

Purpose Identifiers Name this Element in the subjects course Date/times of start and end of the Element for the subject Permissible variable. Blank except for unplanned Elements. Describes the unplanned Element.

STUDYID, DOMAIN, USUBJID ETCD, ELEMENT SESTDTC, SEENDTC SEUPDES

7.7.5

Using Subject Elements Data to Place Subject Data within an Element or Epoch

SDS V3 contained deprecated Period, Phase, and Cycle variables. The date/times in SE allow any date/time to be placed in one subject Element, or before the first subject Element (before the trial) or after the last subject Element (after the trial). Thus, if the subject Elements dataset is submitted, the deprecated period, phase, and cycle variables are not needed. An observation date/time may be compared to the date/times in the Subject Elements dataset in order to place that observation in the proper element. However, if the date/time of the observation is equal to the end date/time of one element and the start date/time of the following element, then it is not immediately clear whether the observation should be assigned to the element that is ending or to the element that is starting. Thus Sponsors should choose a convention for assigning observations to one element or the other, and document the convention. It will be particularly important to specify this convention if date/times are collected only to the nearest day, when "ties" of this kind will be fairly common. The sequence of subject Elements may be compared to the planned sequence of Elements for the Arm to which the subject was assigned. If the sequences match (or if the actual sequence is a truncated version of the planned sequence, as when a subject withdraws early) then it makes sense to associate a value of TAETORD with each subject Element. For trials with Epochs, values of EPOCH may also be associated with subject Elements. If a subjects planned and actual sequences of Elements do not match, then it may not make sense to associate values of TAETORD or EPOCH with the subject Elements, depending on the use that is being made of TAETORD or EPOCH.

7.8
7.8.1

SUBJECT VISITS
Identifying Subject Visits

Subject Visits (SV) describes the visits a subject passed through, with their start and end date/times.

Visit start and end date/times are usually collected directly. In many studies, only a single visit date is collected for each visit. This single date can be used to populate both start and end date/time. Unplanned Visits commonly occur in clinical trials, and records for unplanned visits should be included in SV. Mechanisms for labeling unplanned visits are in common use. SDS is not recommending particular conventions for populating VISIT, VISITNUM, and VISITDY for unplanned visits at this time.

Page 116 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

7.8.2
Variables

Recap of Subject Visits Variables

Purpose Identifiers Variables that name and order the visits Date/times of start and end of this visit for this subject Permissible variable. Blank except for unplanned Visits. Describes the unplanned Visit.

STUDYID, DOMAIN, USUBJID VISIT, VISITNUM, VISITDY SVSTDTC, SVENDTC SVUPDES

7.9

TRIAL INCLUSION/EXCLUSION CRITERIA

The Trial Inclusion Exclusion Domain (TI) is not subject oriented. It contains all the inclusion and exclusion criteria for the trial, and thus provides information that may not be present in the subject-level data on inclusion and exclusion criteria. The IE Domain (described in Section 6.3.2) contains records only for inclusion and exclusion criteria that a subject did not meet, and thus includes only criteria which one or more subjects failed to meet. Inclusion and Exclusion Criteria are rules, so the TI domain includes both the familiar text version of a criterion, IETEST, and a Rule version, TIRL. ti.xpt, Trial Inclusion/Exclusion Criteria, Version 3.1.1. August 26, 2005. One record per I/E criterion
Variable Name STUDYID DOMAIN IETESTCD Variable Label Type Controlled Origin Terms or Format Char Char Char **TI * CRF Derived Sponsor Defined Role CDISC Notes Core

Study Identifier Domain Abbreviation Inclusion/Exclusion Criterion Short Name

Identifier Unique identifier for a study within the submission.

Req

Identifier Two-character abbreviation for the domain. Req Topic Short name IETEST. It can be used as a Req column name when converting a dataset from a vertical to a horizontal format. The value in IETESTCD cannot be longer than 8 characters, nor can it start with a number (e.g.,'1TEST'). IETESTCD cannot contain characters other than letters, numbers, or underscores. The name IE prefix is used to ensure consistency with the IE domain.

IETEST

Inclusion/Exclusion Criterion Inclusion//Exclusion Category

Char

Sponsor Defined / Synonym Full text of the inclusion or exclusion Req Protocol Qualifier criterion. The prefix IE is used to ensure consistency with the IE domain. Grouping 'Used to categorization of the inclusion or Qualifier exclusion criteria. Req

IECAT

Char *INCLUSI Sponsor Defined ON, EXCLUSI ON Char Sponsor Defined

TIRL

Inclusion/Exclusion Criterion Rule

Rule

Optional rule that expresses the criterion in Perm computer-executable form.

CDISC, 2005. All rights reserved FINAL

Page 117 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

7.10

TRIAL SUMMARY INFORMATION

The Trial Summary Model allows the sponsors to submit a summary of the trial in a structured format. The Trial Summary dataset contains values for each characteristic described for a trial. Trial Summary is used to record basic information about the study such as trial phase, protocol title, and trial objectives. The Trial Summary domain contains information about the planned trial characteristics such as length of trial, planned number of subjects, planned maximum age of subject and others. The Trial Summary (TS) domain does not contain subject level data or data that can be derived from subject data. Thus, for example, it includes the number of subjects planned for the trial but not the number of subjects enrolled in the trial. ts.xpt, Trial Summary, Version 3.1.1, August 26, 2005. One record per trial summary parameter
Variable Name STUDYID DOMAIN Variable Label Study Identifier Type Char **TS Controlled Terms or Format Origin Derived Derived Role CDISC Notes Core Req Req Req Req

Identifier Unique identifier for a study within the submission.

Identifier Two-character abbreviation for the domain most relevant to the observation. TSSEQ Derived Identifier Sequence number given to ensure uniqueness within a dataset. Allows inclusion of multiple records for the same parm, and can be used to join related records. TSPARMCD Trial Char ** Derived Topic Short name for the Parameter described in TSPARM. The value in TSPARMCD cannot be longer than 8 Summary Parameter characters, nor can it start with a number. TSPARMCD cannot contain characters other than letters, numbers, or Short Name underscores. Examples: DESIGN, MASK, COMPTRT TSPARM Trial Char ** Derived Synonym Term for the Trial Summary Parameter. The value in Summary Qualifier TSPARM cannot be longer than 40 characters. Parameter Examples: Description Of Trial Design, Trial Blinding Schema, Comparative Name Of Treatment TSVAL Value of TSPARM. Example: 'ASTHMA' when Derived Result Parameter Char ** Qualifier TSPARM value is 'Trial Indications'. TSVAL cannot Value be null a value is required for the record to be valid. The first 200 characters of TSVAL will be in TSVAL, then next 200 in TSVAL1, and continuing as needed to TSVALn. * indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.

Domain Char Abbreviation Sequence Num Number

Req

Req

7.10.1.1 Assumptions for Trial Summary (TS) domain model 1. 2. 3. 4. 5. 6. The intent of the domain model is to provide a summary of the trial information. This is not subject level data. TSVAL may have controlled terminology depending on the value of TSPARMCD, see appendix 10.3.5 "Trial Summary" for more information. Trial Summary is similar to the COMMENTS special-purpose domain. It allows for one value to span multiple variables (TSVAL-TSVALn) in order to accommodate values longer than 200 characters.1 The TSPARMCD value 'PHASE' is not defined with any exactness since it can vary by drug class. For LENGTH OF TRIAL is defined as the planned length of time for a subject's participation. For "LENGTH OF TRIAL", units for duration are not needed. See Section 4.1.4.3. For some trials, there will be multiple records in the Trial Summary dataset for a single parameter such as. "TYPE OF TRIAL" since a trial could be for both the Safety and Efficacy parameters. In this case, when TSPARMCD=TYPE there will be two records for TSVAL, one with the value of 'SAFETY' and the other with the value of 'EFFICACY'.

Allowing for multiple variables per comment is an interim solution until the limitations posed by SAS Version 5 transport files are eliminated.
CDISC, 2005. All rights reserved FINAL

Page 118 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

7.11

HOW TO MODEL THE DESIGN OF A CLINICAL TRIAL

The following steps allow the modeler to move from familiar concepts, such as Arms, to less familiar concepts, such as Elements and Epochs. The actual process of modeling a trial may depart from these numbered steps. Some steps will overlap, there may be several iterations, and not all steps are relevant for all studies. 1. 2. 3. Start from the flow chart or schema diagram usually included in the trial protocol. This diagram will show how many arms the trial has, and the branch points, or decision points, where the arms diverge. Write down the decision rule for each branching point in the diagram. Does the assignment of a subject to an arm depend on a randomization? On whether the subject responded to treatment? On some other criterion? If the trial has multiple branching points, check whether all the branches that have been identified really lead to different Arms. The Arms will relate to the major comparisons the trial is designed to address. For some trials, there may be a group of somewhat different paths through the trial that are all considered to belong to a single Arm. For each arm, identify the major time periods of treatment and non-treatment a subject assigned to that arm will go through. These are the Elements, or building blocks, of which the Arm is composed. Define the starting point of each Element. Define the rule for how long the Element should last. Determine whether the element is of fixed duration. Re-examine the sequences of Elements that make up the various Arms and consider alternative Element definitions. Would it be better to 'split' some Elements into smaller pieces or 'lump' some Elements into larger pieces? Such decisions will depend on the aims of the trial and plans for analysis. Compare the various Arms. In most clinical trials, especially blinded trials, the pattern of Elements will be similar for all Arms, and it will make sense to define Trial Epochs. Assign names to these Epochs. During the conduct of a blinded trial, it will not be known which Arm a subject has been assigned to, or which treatment Elements they are experiencing, but the Epochs they are passing through will be known. Identify the Visits planned for the trial. Define the planned start timings for each visit, expressed relative to the ordered sequences of Elements that make up the Arms. Define the rules for when each visit should end. Identify the inclusion and exclusion criteria to be able to populate the TI domain.

4. 5. 6.

7.

8. 9.

CDISC, 2005. All rights reserved FINAL

Page 119 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

8 Representing Data Relationships

Although each domain in a regulatory submission is defined to stand alone as a separate dataset, there are many occasions when it is necessary or desirable to represent relationships among datasets or records, such as relationships between separate domains, or between domains and other special-purpose datasets. In some cases, this is necessary because the fixed structure of the SDTM observation classes restricts the ability of sponsors to represent data that do not entirely fit within those structures. The SDTM accommodates five distinct types of relationships, which are described in more detail in subsequent sections. Section 8.1 describes a relationship between a group of records for a given subject within the same domain. Section 8.2 describes a relationship between independent records (usually in separate domains) for a subject, such as a concomitant medication taken to treat an adverse event. Section 8.3 describes a relationship between two (or more) datasets where all the records of one (or more) dataset(s) have parent or counterpart record(s) in another dataset (or datasets). Section 8.4 describes a dependent relationship between data that cannot be represented by a standard variable in the general class and a parent record (or records) within a domain. Section 8.5 describes a dependent relationship between a comment and a parent record (or records) in other domains, such as a comment recorded with an adverse event. Section 8.6 discusses the concept of related datasets and whether to place additional data in a domain or supplemental qualifier table

See appendix 10.6 for a proposed approach for modeling findings data that may refer to interventions or events information. All relationships make use of the standard domain identifiers, STUDYID, DOMAIN, and USUBJID. In addition, the variables IDVAR and IDVARVAL are essential for identifying the record-level merge/join keys. The specific set of identifiers necessary to properly identify each type of relationship is described in detail in the following sections. The Sequence Number (--SEQ) variable uniquely identifies a record for a given USUBJID within a domain. --SEQ is required in all domains except DM. For example, if subject 1234-2003 has 25 adverse event records in the adverse event (AE) domain, AESEQ values for this subject should be the numbers 1 to 25, with numbering starting over with 1 for the next subject. Numbers may, however, be omitted from the sequence if the sponsor assigns the numbers early in the process and subsequently deletes some records. The Reference Identifier (--REFID) variable can be used to capture a sponsor-defined or external identifier, such as an identifier provided in an electronic data transfer. Some examples are lab-specimen identifiers, ECG identifiers, or image filename identifier (sometimes used to reference external files such as medical images or waveform ECGs). --REFID is permissible in all domains, but never required. Values for --REFID are sponsor defined and can be any alphanumeric strings the sponsor chooses, consistent with their internal practices. The Group Identifier (--GRPID) variable is explained below in section 8.1.

Page 120 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

8.1

RELATING GROUPS OF RECORDS WITHIN A DOMAIN

The optional grouping identifier variable --GRPID is permissible in all domains that are based on the general observation classes to identify relationships between records within a domain (such as Intervention records for a combination therapy). The relationship is defined by assigning the same unique (within USUBJID) character value to the --GRPID variable. All records for a subject in a domain are considered to be related when they have the same --GRPID value. The --GRPID values are only meaningful within the domain, and in the Relate Records (RELREC), Supplemental Qualifiers (SUPPQUAL), and Comments (COMMENTS) datasets, as described in subsequent sections. The values used for --GRPID can be any values the sponsor chooses; however, the philosophy for assigning values should be consistent across the submission to avoid confusion. For example, --GRPID could be any or all of the following: The therapy name for multiple exposure records associated with a combination therapy A sponsor-defined term used to group sets of questions together in the Questionnaire (QS) domain Any sponsor-defined constant value used to group multiple concomitant medications together to indicate they were all taken to treat the same condition. Using --GRPID in the SDS Domains can be helpful to reduce the number of records in the RELREC, SUPPQUAL, and COMMENTS datasets when it is necessary to use those datasets to capture/identify relationships/associations for records or values to a 'group' of parent/peer records.

8.1.1

--GRPID Example

The following table illustrates how to use --GRPID in the Concomitant Medications (CM) domain to identify a combination therapy. Note that the example only includes the most relevant variables from CM for this example; the complete CM domain might include other variables not shown below. In this example, subject 1234 has reported two combination therapies, each consisting of three separate medications. Each component of a combination is given the same value for GRPID.
STUDYID DOMAIN USUBJID CMSEQ 1234 1234 1234 1234 1234 1234 CM CM CM CM CM CM 1234 1234 1234 1234 1234 1234 1 2 3 4 5 6 CMGRPID COMBO THPY 1 COMBO THPY 1 COMBO THPY 1 COMBO THPY 2 COMBO THPY 2 COMBO THPY 2 CMTRT Verbatim Med A Verbatim Med B Verbatim Med C Verbatim Med D Verbatim Med E Verbatim Med F CMDECOD CMDOSE CMDOSU CMSTDTC CMENDTC Generic Med A Generic Med B Generic Med C Generic Med D Generic Med E Generic Med F 100 50 200 150 100 75 mg mg mg mg mg mg 2004-01-17 2004-01-17 2004-01-17 2004-01-21 2004-01-21 2004-01-21 2004-01-19 2004-01-19 2004-01-19 2004-01-22 2004-01-22 2004-01-22

CDISC, 2005. All rights reserved FINAL

Page 121 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

8.2

RELATING PEER RECORDS IN SEPARATE DATASETS

The Related Records (RELREC) dataset is used to describe relationships between records in two (or more) datasets, such as an Event record and an Intervention record, or a Finding record and an Event record. A relationship is defined by creating RELREC records for each of the records to be related and by assigning a unique character identifier value for the relationship. Each RELREC record contains keys that identify a record (or group of records) and the relationship identifier, which is stored in the RELID variable. The value of RELID can be any constant value chosen by the sponsor. Records expressing a relationship are specified using the keys STUDYID, RDOMAIN, and USUBJID, which are populated with the values for STUDYID, DOMAIN, and USUBJID from the SDS domain records, along with an IDVAR and an IDVARVAL. Single records can be related by using --SEQ in IDVAR. Groups of records can be specified by using --GRPID in IDVAR and the appropriate --GRPID value in IDVARVAL. Using the optional grouping identifier variable --GRPID (see section 8.1) to group a set of related records in the domains can be a more efficient method of representing relationships in RELREC, such as when relating an adverse event (or events) to a group of concomitant medications taken to treat the adverse event(s). For example, suppose adverse events are determined at the investigational site to be associated with the administration of a concomitant medication used in standard care. Since the adverse events and the concomitant medication are captured in separate domains, a separate record in the RELREC dataset must be used to describe the observed association among these records. The RELREC dataset should be used only when the sponsor either: 1. Collects information about explicit relationships, such as concomitant medications taken as a result of an adverse event 2. 3. Determines there is a scientific rationale or reporting need to identify relationships after data collection such as determining that an adverse event was due to a concomitant medication, or Collects information of a nature that necessitates using multiple datasets, as described in section 8.3.

This is not to imply that a sponsor should retrospectively identify all potential relationships in the RELREC dataset; instead, the purpose of the RELREC dataset is to give sponsors a standardized way to represent such relationships when a sponsor identifies a specific need to do so. The RELREC dataset should only be used to associate entire observations on a one-to-one or one-to-many level. Additional Qualifier variables that cannot be represented in existing variables in Findings, Interventions, or Events domains should be represented using the Supplemental Qualifiers (SUPPQUAL) dataset, which is described in Section 8.4.

Page 122 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

8.2.1

RELREC Dataset
relrec.xpt, Related Records, Version 3.1.1, August 26, 2005. One record per record or domain relationship
Variable STUDYID RDOMAIN Variable Label Study Identifier Related Domain Abbreviation Unique Subject Identifier Identifying Variable Identifying Variable Value Relationship Type Type Char Char* CDISC Notes Study Identifier of the SDS domain record(s). Domain Abbreviation of the SDS domain record(s).

USUBJID IDVAR IDVARVAL

Char Char** Char

Unique Subject Identifier of the SDS domain record(s). Value of the identifying variable in the general class dataset that identifies the related record(s). Examples include --SEQ and --GRPID. Value of identifying variable described in IDVAR. If --SEQ is the variable being used to describe this record, then the value of --SEQ would be entered here. Identifies the hierarchical level of the records in the relationship. Values should be either ONE or MANY. However, values are only necessary when identifying a relationship between datasets (as described in Section 8.3). Unique value within USUBJID that identifies the relationship. All records for the same USUBJID that have the same RELID are considered 'related/associated.' RELID can be any value the sponsor chooses, and is only meaningful within the RELREC dataset to identify the related/associated Domain records.

RELTYPE

Char**

RELID

Relationship Identifier

Char

* indicates variable may be subject to sponsor-controlled terminology ** indicates variable may be subject to external controlled terminology.

8.2.2
Example A:

RELREC Dataset Examples

Example A shows how to use the RELREC dataset when USUBJID 123456 had two lab tests (LBSEQ values of 47 and 48) and took two concomitant medications (CMSEQ values of 11 and 12) as the result of an adverse event (AESEQ value of 5). This example assumes that the lab values and the concomitant medications are not related to each other.
STUDYID EFC1234 EFC1234 EFC1234 EFC1234 EFC1234 EFC1234 RDOMAIN AE CM CM AE LB LB USUBJID 123456 123456 123456 123456 123456 123456 IDVAR AESEQ CMSEQ CMSEQ AESEQ LBSEQ LBSEQ IDVARVAL 5 11 12 5 47 48 RELTYPE RELID 1 1 1 2 2 2

NOTE: The two lab tests could have the same LBGRPID if they were on the same specimen. In such a case, the two RELREC records with RDOMAIN=LB could be replaced with one RELREC record with IDVAR=LBGRPID and IDVARVAL having the LBGRPID value assigned in the LB domain.
CDISC, 2005. All rights reserved FINAL Page 123 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Example B: Example B is the same scenario as Example A; however, the concomitant medications and lab tests are associated with each other in addition to being associated with the adverse event.
STUDYID EFC1234 EFC1234 EFC1234 EFC1234 EFC1234 RDOMAIN AE CM CM LB LB USUBJID 123456 123456 123456 123456 123456 IDVAR AESEQ CMSEQ CMSEQ LBSEQ LBSEQ IDVARVAL 5 11 12 47 48 RELTYPE RELID 1 1 1 1 1

Example C: Example C is the same scenario as Example B, however, both the concomitant medications are necessary for this treatment and have been assigned a CMGRPID of 'COMBO 1' by the sponsor, allowing the elimination of a record in the RELREC dataset.
STUDYID EFC1234 EFC1234 EFC1234 EFC1234 RDOMAIN AE CM LB LB USUBJID 123456 123456 123456 123456 IDVAR AESEQ CMGRPID LBSEQ LBSEQ IDVARVAL 5 COMBO1 47 48 RELTYPE RELID 1 1 1 1

8.3

RELATING DATASETS

The Related Records (RELREC) dataset can also be used to identify relationships between datasets (e.g., a one-tomany or parent-child relationship, as discussed in Section 8.3.1). The relationship is defined by including a single record for each related dataset that identifies the key(s) of the dataset that can be used to relate the respective records. Relationships between datasets should only be recorded in the RELREC dataset when the sponsor has found it necessary to split information between datasets that are related, and that may need to be examined together for analysis or proper interpretation. It is not necessary to use the RELREC dataset to identify associations from data in the SUPPQUAL dataset or the COMMENTS dataset, as both these datasets include the key variable identifiers of the parent record that are necessary to make the association.

8.3.1

RELREC Dataset Relationship Example

This example shows how to use the RELREC dataset to represent related information that is submitted as two datasets which have a one-to-many relationship. In the example below all the records in one domain are being related to all of the records in the other so both USUBJID and IDVAR are null.
STUDYID EFC1234 EFC1234 RDOMAIN MB MS USUBJID IDVAR MBSPID MSSPID IDVARVAL RELTYPE ONE MANY RELID A A

In the sponsor's operational database, these datasets may have existed as either separate datasets that were merged for analysis, or one dataset that may have included observations from more than one observation class (e.g., Events and Findings). The value in IDVAR should be the name of the key used to merge/join the two datasets. In the above example, the --SPID variable is used as the single key to identify the related observations. The values for the --SPID

Page 124 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

variable in the two datasets are sponsor defined. Although other variables may also serve as a single merge key when the corresponding values for IDVAR are equal, --SPID or --REFID are typically used for this purpose. Since IDVAR identifies the keys that can be used to merge/join all records between the datasets, the root values (i.e., --SPID in the above example) for IDVAR must be the same for both records with the same RELID. --SEQ cannot be used because --SEQ only has meaning within a dataset, not across datasets.

8.4

RELATING NON-STANDARD VARIABLE VALUES TO A PARENT DOMAIN

The Supplemental Qualifiers (SUPPQUAL) special purpose dataset model is used to capture non-standard variables and their association to parent records in domains, and allows capturing values for variables not presently included in the general-observation-class models. Because the SDTM does not allow the addition of new variables, it is necessary for sponsors to represent the metadata and data for each non-standard variable/value combination in a SUPPQUAL dataset. The SUPPQUAL dataset model is structured similarly to the RELREC dataset, in that it uses the same set of keys to identify related records. Each SUPPQUAL record also includes the name of the Qualifier variable being added (QNAM), the label for the variable (QLABEL), the actual value for each instance or record (QVAL), the origin (QORIG) of the value (whether it was collected via CRF, assigned or derived), and the Evaluator (QEVAL) to specify the role of the individual who assigned the value (such as ADJUDICATION COMMITTEE or SPONSOR). One common case for using a SUPPQUAL dataset is to capture attributions. An attribution is typically an interpretation or subjective classification of one or more observations by a specific evaluator, such as a population flag that classifies a subject or their data according to their evaluability for efficacy analysis. Since it is possible that different attributions may be necessary in some cases, SUPPQUAL provides a mechanism for incorporating as many attributions as are necessary. It is recognized that sponsors may also need to use a SUPPQUAL dataset to submit additional non-standard variables that cannot be represented in the general classes. Standard names for certain expected values for QNAM and QLABEL are included in Appendix 10.3.4. The combined set of values for the first six columns (STUDYIDQNAM) should be unique for every record. That is, there should not be multiple records in a SUPPQUAL dataset for the same QNAM value, as it relates to IDVAR/IDVARVAL for a USUBJID. For example, if two individuals provide a determination on whether an Adverse Event is Treatment Emergent (e.g., the investigator and an independent adjudicator) then separate QNAM values should be used for each set of information, perhaps AETRTEMI and AETRTEMA. This is necessary to ensure that reviewers can join/merge/transpose the information back with the records in the original domain without risk of losing information. When populating a SUPPQUAL dataset with population flags related to the Demographics domain (subject-level evaluability), there should be one record for each population flag for each subject. QVAL values for population flags should be Y or N, with no Null values. In the event that evaluability is based upon individual visits or CRF pages, additional population flags attached to other domains may be included in SUPPQUAL datasets. Just as use of the optional grouping identifier variable --GRPID can be a more efficient method of representing relationships in RELREC, it can also be used in a SUPPQUAL dataset to identify individual qualifier values (SUPPQUAL records) related to multiple domain records that could be grouped, such as relating an attribution to a group of ECG measurements.

CDISC, 2005. All rights reserved FINAL

Page 125 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

8.4.1
Variable

SUPPQUAL Dataset
Variable Label Type CDISC Notes References SDTM 2.2.4 SDTM 2.2.4 SDTM 2.2.4

suppqual.xpt, Supplemental Qualifiers, Version 3.1.1, August 26, 2005. One record per variable

STUDYID Study Identifier RDOMAIN Related Domain Abbreviation USUBJID Unique Subject Identifier IDVAR Identifying Variable IDVARVAL Identifying Variable Value QNAM Qualifier Variable Name

Char Study Identifier of the Parent record(s). Char* Domain Abbreviation of the Parent record(s). Null for an attribution (e.g., ITT) that may apply for all records in all domains for a subject. Char Unique Subject Identifier of the Parent record(s). Char* Identifying variable in the dataset that identifies the related record(s). Examples: --SEQ, --GRPID. Char Value of identifying variable of the parent record(s).

QLABEL

QVAL

QORIG

QEVAL

Char* The short name of the Qualifier variable, which is used as a column SDTMIG name in a domain view with data from the parent domain. The value in 4.1.5.2 QNAM cannot be longer than 8 characters, nor can it start with a number (e.g.,'1TEST'). QNAM cannot contain characters other than letters, numbers, or underscores. This will often be the column name in the sponsors operational dataset. Qualifier Variable Char This is the long name or label associated with QNAM. The value in Label QLABEL cannot be longer than 40 characters. This will often be the column label in the sponsors original dataset. Data Value Char Result of, response to, or value associated with QNAM. A value for this column is required; no records can be in SUPPQUAL with a null value for QVAL. Origin Char* Since QVAL can represent a mixture of collected (on a CRF), derived, or assigned items, QORIG is used to indicate the origin of this data. Examples include CRF, ASSIGNED, or DERIVED. Evaluator Char* Used only for results that are subjective (e.g., assigned by a person or a group). Should be null for records that contain objectively collected or derived data. Some examples include ADJUDICATION COMMITTEE, STATISTICIAN, DATABASE ADMINISTRATOR, CLINICAL COORDINATOR, PRIMARY INVESTIGATOR, etc.

* indicates variable may be subject to sponsor-controlled terminology; ** indicates variable may be subject to external controlled terminology.
Note all variables except QEVAL must be populated for each record. A record in a SUPPQUAL dataset relates back to its general observation class dataset parent record(s) via the key identified in the IDVAR variables. SUPPQUAL dataset records that are related to all subject records in all domains, such as the Intent To Treat (ITT) and Safety (SAFETY) subject-level population flags, will have both IDVAR and IDVARVAL set to null. This is because the only keys needed to identify the relationship/association to all subject records (for that subject) in all domains are STUDYID and USUBJID. A SUPPQUAL dataset can contain both objective data (where values are collected or derived algorithmically) and subjective data (attributions where values are assigned by a person or committee). For objective data, the value in QEVAL will be null. For subjective data (when QORIG=ASSIGNED), the value in QEVAL should reflect the role of the person or institution assigning the value (e.g., PRIMARY INVESTIGATOR or ADJUDICATION COMMITTEE). All records in the SUPPQUAL datasets must have a value for QVAL. Transposing source variables with missing/null values may inadvertently generate SUPPQUAL records with null values for QVAL, causing the SUPPQUAL dataset to be extremely large. When this happens, the sponsor should delete the records where QVAL is null prior to submission.

Page 126 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

8.4.2

Submitting Supplemental Qualifiers in Separate Datasets

Because sponsors often have a large number of supplemental qualifier variable values, the SDTM IG has defined a preferred approach to submit supplemental qualifiers by domain rather than placing all of the supplemental information within one dataset. Following this convention for the supplemental qualifiers produces a one-to-one correspondence between a domain dataset and its supplemental qualifier dataset. In such cases, the following rules should be met: The set of supplemental qualifiers for each domain should be included in a separate dataset with the name supp-- where -- denotes the source domain which the supplemental qualifiers relates back to. For example, population flags and other demographic qualifiers would be placed in suppdm.xpt. For users conforming to V3.1.1 and later versions of the SDTMIG, a SUPPQUAL dataset should NOT be submitted in addition to separate supp-- datasets for individual domains. In other words, separate supplemental qualifier domains cannot be used with some domains and SUPPQUAL for the other s).

8.4.3

SUPPQUAL Examples

The example below demonstrates how a set of SUPPQUAL datasets should be used to relate non-standard information to a parent domain. In the first two rows of suppdm.xpt, population flags are defined as supplemental information to a subjects demography data. Since IDVAR and IDVARVAL are not populated, we know that the qualifying information is applicable to the subject as a whole, not to any specific record. The next two rows of suppae.xpt add qualifying information to adverse event data. IDVAR defines the key variable that links this information to the AE data (AESEQ). IDVARVAL specifies the value of the key variable within the parent record that the SUPPQUAL record applies to. The remaining rows specify the supplemental variables names (AESOTHC and AETRTEM), labels, values, origin (CRF or DERIVED), and who made the evaluation. The final set of records in the suppqs.xpt example show how to record the language used for a questionnaire. The parent domain (RDOMAIN) is QS, and the merge key (IDVAR) is QSCAT. QNAM holds the name of the supplemental qualifier variable being defined (QSLANG). The language recorded in QVAL applies to all of a subjects records where IDVAR (QSCAT) equals IDVARVAL. In this case, IDVARVAL has values for two questionnaires (SF36 and ADAS) for two separate subjects. QVAL identifies the questionnaire language version (Flemish or Farsi) for each subject. suppdm.xpt: Supplemental Qualifiers for DM
STUDYID RDOMAIN USUBJID IDVAR IDVARVAL QNAM QLABEL 1996001 DM 99-401 ITT Intent to Treat 1996001 DM 99-401 PPROT Per Protocol Evaluable QVAL Y N QORIG DERIVED DERIVED QEVAL SPONSOR SPONSOR

suppae.xpt: Supplemental Qualifiers for AE

STUDYID RDOMAIN USUBJID IDVAR IDVARVAL 1996001 1996001 AE AE 99-401 99-401 AESEQ AESEQ 1 1 QNAM QLABEL QVAL QORIG CRF DERIVED SPONSOR QEVAL AESOTHC Desc of Other Med Impt Spontaneous Serious Event Abortion AETRTEM Treatment Emergent N

suppqs.xpt: Supplemental Qualifiers for QS

STUDYID RDOMAIN 1996001 1996001 1996001 1996001 QS QS QS QS USUBJID 99-401 99-401 99-802 99-802 IDVAR IDVARVAL QNAM QSCAT QSCAT QSCAT QSCAT SF36 ADAS SF36 ADAS QSLANG QSLANG QSLANG QSLANG QLABEL Questionnaire Language Questionnaire Language Questionnaire Language Questionnaire Language QVAL FLEMISH FLEMISH FARSI FARSI QORIG QEVAL CRF CRF CRF CRF

Examples of data that should not be in the SUPPQUAL dataset include: Subject-level objective data that fit in Subject Characteristics (SC) ECG Interpretation -- this should be added as an additional test called EGINTP in EGTESTCD, and given the same EGGRPID or EGREFID value for all records associated with that ECG.

CDISC, 2005. All rights reserved FINAL

Page 127 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

8.5

RELATING COMMENTS TO A PARENT DOMAIN

The Comments special-purpose domain, which is described in Section 5.1.2, is used to capture unstructured free text comments. It allows for the submission of comments related to a particular domain (e.g., adverse events) or those collected on separate general-comment pages not associated with a domain. Comments may be related to a Subject, a domain for a Subject, or to specific parent records in any domain. The Comments domain is structured similarly to the SUPPQUAL dataset, in that it uses the same set of keys (STUDYID, RDOMAIN, USUBJID, IDVAR and IDVARVAL) to identify related records. All comments are considered child records of data captured in domains. STUDYID, USUBJID, and DOMAIN (with the value CO) must always be populated. RDOMAIN, IDVAR, and IDVARVAL should be populated as follows: 1. 2. Comments related only to a subject in general would have RDOMAIN null, as the only key needed to identify the relationship/association to that subject is USUBJID. Comments related only to a specific domain (and not to any specific record) for a subject would populate RDOMAIN with the code for the domain with which they are associated. IDVAR and IDVARVAL would be null. Comments related to specific domain records for a subject would populate the RDOMAIN, IDVAR, and IDVARVAL variables with values from the parent record(s).

3.

Again, --GRPID can be used as the value in IDVAR to identify comments with relationships to multiple domain records, such as a comment that applies to a 'group' of concomitant medications.

8.5.1

COMMENTS Example

In the following example, Subject 1234-2004 has five separate comments: one comment on an adverse events (AE) record, one on an exposure (EX) record, and one on a vital signs (VS) record; one on the PE CRF page, and another on a separate comments CRF page. Another comments example is provided in Section 5.1.2.
STUDYID 1234 1234 1234 DO RDO USUBJID CO IDVAR IDVAR COREF CODTC MAIN MAIN SEQ VAL CO CO CO AE 1234-2004 EX 1234-2004 VS 1234-2004 1 AESEQ 2 EXSEQ 3 VSSEQ 7 17 5

COVAL
Same AE has occurred prior to treatment Some delays in taking medications because of vacation Value confirmed in range, but questionable based on other related measurements. Subject was very uncooperative during physical exam. Subject overall attitude has improved greatly during study.

COEVAL PRINCIPAL INVESTIGATOR PRINCIPAL INVESTIGATOR ADJUDICATION COMMITTEE PRINCIPAL INVESTIGATOR PRINCIPAL INVESTIGATOR

PAGE 650

1234 1234

CO CO

PE

1234-2004 1234-2004

4 5

PAGE 40 2003-11-08
2004-01-14

Page 128 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

8.6

HOW TO DETERMINE WHERE DATA BELONG IN THE SDTM

A domain is defined as a collection of observations on a particular topic. For example, Medical History is a collection of information related to events or conditions that happened before the clinical study. Typically the term 'domain' has been synonymous with a dataset. However, there may be cases where data related to the same topic might, at times, need to be captured in more than one domain or observation class. In such cases, information that may have been historically presented in a single dataset may now need to be represented in more than one SDTM dataset. It is often necessary to refer to the protocol or statistical analysis plan to decide how to model new data domains not modeled in the SDTMIG and understand whether a collection of information logically belongs together in a single domain or dataset. When determining how to model a new type of data, it is best to identify the domain topicality first and classify its primary observation class as an Event, Intervention, or Finding (see section 2.4). To do this, begin by evaluating the overall logical content of the information rather than analyzing the physical structure. Applying this principle to Medical History, for example, note that the collected information is related to previous conditions or events, regardless of whether dates were collected. Just because dates were not collected or available in specific cases does not change the fact that the primary topicality fits Events. Conversely, if the topic of the information is more related to specific measurements rather than just recording occurrences, it is typically modeled in the Findings class. Once the domain topicality and its general observation class are identified, then the physical structure of the information should be evaluated to verify whether all collected information can be represented in a single observation class. Just because the information is collected /recorded on the same individual CRF pages does not necessarily mean that all the data collected/recorded on those individual CRF pages belongs in the same domain. Information grouped together on CRFs for convenience but obviously having different topicalities (such as Concomitant Medications collected on SAE forms) should be mapped to multiple domains. If the sponsor feels that its important for the reviewer to easily identify the relationship between the AE and CM records, then RELREC can be used (see Section 8.2.1). Information grouped together because it is dependent on other information, such as findings collected as a result of an event, requires careful evaluation to determine whether all the information should be in a single domain, split into multiple datasets, or submitted in a supplemental qualifiers dataset. The following guidelines should be used in making this determination: Submit data in a supplemental qualifiers special-purpose dataset (see Section 8.4) if the information that does not fit consists only of additional variables that would not be classified as Identifiers, Topic, or Timing variables in a separate domain. In other words, the information consists only of what would be classified as additional Qualifiers to further describe an existing record, and would not be considered sufficient to exist as an entire 'record' of information by itself in any observation class. A general way of thinking about this is if the data would simply have been another column in the sponsors original dataset. Submit data as a separate dataset if the information that does not fit could comprise a complete record in that dataset (it contains variables that would be classified as Identifiers, Topic, or Timing variables). Section 8.2 describes how information should be stored in a second dataset of another general observation class for the same Domain when the information has, at minimum a Topic Variable, its own Timing variables and (for Findings domains) at least one Result Qualifier.

General rules for creating a separate related dataset of common topicality with another are as follows: 1. Typically, additional datasets should be created for either a. b. c. Interventions with associated Findings, or Events with associated Findings Two sets of Findings that do not fit logically in a single dataset.

CDISC, 2005. All rights reserved FINAL

Page 129 July 28, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

2.

The topicality of the information and the context determine the Parent (or master) dataset general observation class (Interventions, Events or Findings). In some cases, a Parent (master) dataset may be required, even if the records in that dataset have limited information Both datasets should include a common identifier variable (e.g., --REFID) to capture the identification (name) of the record to be used as a merge/join key: however, it is recognized that in some cases it would be more appropriate to use --GRPID. Values for the --REFID or --GRPID variables would need to be the same for related records in the two (or more) datasets. Both datasets should use a different two-character code for a value in the Domain variable, and as the variable prefix (MB and MS for example). The dataset names for the two datasets should be the twocharacter Domain codes chosen for the datasets (MB and MS respectively). Since it is not self-evident that a Domain is captured in multiple datasets using this combination of dataset names (a reviewer might not automatically know that MB and MS are related to each other), sponsors must populate the RELREC dataset (see Section 8.2.1) with rows that identify the dataset relationship and document the approach used in the Define file (metadata) and/or Annotated CRF. This approach is fully compatible with the current V3.x standard as long as the dataset relationship is clearly documented.

3.

4.

Page 130 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

9 Implementation Examples
9.1 DEMOGRAPHICS EXAMPLE
This is an example of demographics that illustrates the use of the RACE and ETHNIC variables. Both RACE and ETHNIC are subjected to the rules of controlled terminology. Race (RACE) and ethnicity (ETHNIC) are independent of each other, for example, the first subject (USUBJID=1) is 'WHITE' for RACE and 'HISPANIC OR LATINO' for ETHNIC. Ethnicity can only have two values, either 'HISPANIC OR LATINO' or 'NOT HISPANIC OR LATINO'. In cases of mixed race, the RACE variable should be used for the primary race, and the SC domain should be used for additional non-primary race values. For example, in Row 4, the 4th subject (USUBJID=4), who is recorded as 'ASIAN' for RACE could have the value 'INDIAN' in SC where SCTESTCD equals RACEOTH.
Row Row 1 Row 2 Row 3 Row 4 Row 5 Row 6 STUDYID ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 DOMAIN DM DM DM DM DM DM USUBJID 1 2 3 4 5 6 RACE WHITE WHITE BLACK OR AFRICAN AMERICAN ASIAN AMERICAN INDIAN OR ALASKA NATIVE NATIVE HAWAIIAN OR OTHER PACIFIC ISLANDERS ETHNIC HISPANIC OR LATINO NOT HISPANIC OR LATINO NOT HISPANIC OR LATINO NOT HISPANIC OR LATINO NOT HISPANIC OR LATINO NOT HISPANIC OR LATINO

9.2
9.2.1

INTERVENTIONS EXAMPLES
CM Example: Intermittent Use of Concomitant Medications

Sponsors collect the timing of concomitant medication use with varying specificity, depending on the pattern of use; the type, purpose, and importance of the medication; and the needs of the study. Often the same information can be conveyed with a range of dates and the daily frequency rather than presenting every unique instance of medication use. If appropriate, medications taken intermittently or sporadically over a time period may be reported with a date range and a frequency of 'PRN'. The example below shows three subjects who took the same medication at the same times. For the first subject (USUBJID=1), each instance is recorded separately, and frequency (CMDOSFRQ) is ONCE. For the second subject (USUBJID=2), the second record (CMSEQ=2) shows that aspirin was taken twice on January 7th, so the frequency is BID. The frequency is also included for the other daily records to avoid confusion. The third subjects (USUBJID=3) records are
CDISC, 2005. All rights reserved FINAL Page 131 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1) collapsed (this is shown as an example only, not as a recommendation) into a single entry that spans the relevant time period, with a frequency of PRN. This approach assumes that knowing exactly when aspirin was used is not important for evaluating this studys safety and efficacy.
Row Row 1 Row 2 Row 3 Row 4 Row 5 Row 6 Row 7 Row 8 Row 9 Row 10 STUDYID ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 DOMAIN CM CM CM CM CM CM CM CM CM CM USUBJID 1 1 1 1 1 1 2 2 2 3 CMSEQ 1 2 3 4 5 6 1 2 3 1 CMTRT Aspirin Aspirin Aspirin Aspirin Aspirin Aspirin Aspirin Aspirin Aspirin Aspirin CMDOSFRQ ONCE ONCE ONCE ONCE ONCE ONCE QD BID QD PRN CMSTDTC 2004-01-01 2004-01-02 2004-01-03 2004-01-07 2004-01-07 2004-01-09 2004-01-01 2004-01-07 2004-01-09 2004-01-01 CMENDTC 2004-01-01 2004-01-02 2004-01-03 2004-01-07 2004-01-07 2004-01-09 2004-01-03 2004-01-07 2004-01-09 2004-01-09

Page 132 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

9.2.2

EX Examples:

Exposure Example 1: This is an example of an Exposure dataset for a parallel design study. In this example, subjects were randomized to one of three treatment groups: Drug A 40 mg q.d., Drug A 20 mg q.d., or Drug B 150 mg bid. Drug A was in capsule form and Drug B was in tablet form. To maintain the blind, subjects took both capsules and tablets with only one form being active. The study included 8 weeks of treatment, with subjects remaining on the same treatment throughout the study. With respect to timing of doses, the sponsor only collected the start and stop dates of uninterrupted periods of treatment. Note below that Subject 12345003 missed taking study medications on Study Days 23 and 24.
STUDYID DOMAIN Row 1 Row 2 Row 3 Row 4 Row 5 Row 6 Row 7 Row 8 12345 12345 12345 12345 12345 12345 12345 12345 EX EX EX EX EX EX EX EX USUBJID 12345001 12345001 12345002 12345002 12345003 12345003 12345003 12345003 EXDOSTO T 40 0 20 0 0 300 0 300 EXSEQ 1 2 1 2 1 2 3 4 EXTRT DRUG A PLACEBO DRUG A PLACEBO PLACEBO DRUG B PLACEBO DRUG B EXDOSE EXDOSU 40 0 20 0 0 150 0 150 MG MG MG MG EXDOSFRM CAPSULE TABLET CAPSULE TABLET CAPSULE TABLET CAPSULE TABLET

EXDOSFRQ Row 1 (cont) Row 2 (cont) Row 3 (cont) Row 4 (cont) Row 5 (cont) Row 6 (cont) Row 7 (cont) Row 8 (cont) QD BID QD BID QD BID QD BID

EXROUTE ORAL ORAL ORAL ORAL ORAL ORAL ORAL ORAL

EXSTDTC 2002-01-10 2002-01-10 2002-01-10 2002-01-10 2002-01-11 2002-01-11 2002-02-04 2002-02-04

EXENDTC 2002-03-08 2002-03-08 2002-03-07 2002-03-07 2002-02-01 2002-02-01 2002-03-06 2002-03-06

EXSTDY 1 1 1 1 1 1 25 25

EXENDY 58 58 57 57 22 22 55 55

CDISC, 2005. All rights reserved FINAL

Page 133 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1) Exposure Example 2: This is an example of an Exposure dataset for a two-way crossover study comparing once daily oral administration of Drug A 20 mg with Drug B 30 mg. Study drug was taken for 3 consecutive mornings 30 minutes prior to a standardized breakfast. There was a 5-day washout period between treatments.
STUDYID Row 1 Row 2 Row 3 Row 4 Row 5 Row 6 Row 7 Row 8 Row 9 Row 10 Row 11 Row 12 56789 56789 56789 56789 56789 56789 56789 56789 56789 56789 56789 56789 DOMAIN EX EX EX EX EX EX EX EX EX EX EX EX USUBJID 56789001 56789001 56789001 56789001 56789001 56789001 56789003 56789003 56789003 56789003 56789003 56789003 EXSEQ 1 2 3 4 5 6 1 2 3 4 5 6 EXSTDTC 2002-07-01T07:30 2002-07-02T07:30 2002-07-03T07:32 2002-07-09T07:30 2002-07-10T07:30 2002-07-11T07:34 2002-07-03T07:30 2002-07-04T07:24 2002-07-05T07:24 2002-07-11T07:30 2002-07-12T07:43 2002-07-13T07:38 EXGRPID 1 1 1 2 2 2 1 1 1 2 2 2 EXTRT DRUG A DRUG A DRUG A DRUG B DRUG B DRUG B DRUG B DRUG B DRUG B DRUG A DRUG A DRUG A EXDOSE EXDOSU 20 20 20 30 30 30 30 30 30 20 20 20 EXSTDY 1 2 3 9 10 11 1 2 3 9 10 11 MG MG MG MG MG MG MG MG MG MG MG MG EXENDY 1 2 3 9 10 11 1 2 3 9 10 11 EXDOSFRM CAPSULE CAPSULE CAPSULE TABLET, COATED TABLET, COATED TABLET, COATED TABLET, COATED TABLET, COATED TABLET, COATED CAPSULE CAPSULE CAPSULE EXTPT 30 MINUTES PRIOR TO STD BREAKFAST 30 MINUTES PRIOR TO STD BREAKFAST 30 MINUTES PRIOR TO STD BREAKFAST 30 MINUTES PRIOR TO STD BREAKFAST 30 MINUTES PRIOR TO STD BREAKFAST 30 MINUTES PRIOR TO STD BREAKFAST 30 MINUTES PRIOR TO STD BREAKFAST 30 MINUTES PRIOR TO STD BREAKFAST 30 MINUTES PRIOR TO STD BREAKFAST 30 MINUTES PRIOR TO STD BREAKFAST 30 MINUTES PRIOR TO STD BREAKFAST 30 MINUTES PRIOR TO STD BREAKFAST
CDISC. 2005. All rights reserved FINAL

EXDOSFRQ QD QD QD QD QD QD QD QD QD QD QD QD

EXDOSTOT EXROUTE Row 1 (cont) Row 2 (cont) Row 3 (cont) Row 4 (cont) Row 5 (cont) Row 6 (cont) Row 7 (cont) Row 8 (cont) Row 9 (cont) Row 10 (cont) Row 11 (cont) Row 12 (cont)
Page 134 August 26, 2005

EXENDTC 2002-07-01T07:30 2002-07-02T07:30 2002-07-03T07:32 2002-07-09T07:30 2002-07-10T07:30 2002-07-11T07:34 2002-07-03T07:30 2002-07-04T07:24 2002-07-05T07:24 2002-07-11T07:30 2002-07-12T07:43 2002-07-13T07:38

20 20 20 30 30 30 30 30 30 20 20 20

ORAL ORAL ORAL ORAL ORAL ORAL ORAL ORAL ORAL ORAL ORAL ORAL

CDISC SDTM Implementation Guide (Version 3.1.1)

9.2.3

SU Example:

Example and Usage Notes for V3.1.1 Substance Use (SU) Domain Model The use of the Interventions general observation class to model substance use data meets the needs of the complex, detailed dosing, frequency, and duration information requirements for analysis where collection of this information is key to the outcome of a clinical trial (e.g., smoking cessation). This example is for clinical trials where detailed substance use dosing, frequency, and duration information is required (e.g., smoking cessation trials). In such cases, detailed information (e.g., smoking substance, total amount per day, start and stop dates) about the substance used is collected. This data could then be used in primary efficacy analyses to support therapy effectiveness claims. Figure 9.2.3.1 presents a CRF that captures data of this nature. The information in the gray bar/ellipses is the sponsor's original annotation.

CDISC, 2005. All rights reserved FINAL

Page 135 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1) Figure 9.2.3.1 - Sample Case Report Form for Smoking Status Visit x SMOKING Smoking Status

Please identify one of the following smoking categories for the subject:

SMOKCAT
1

Not assessed

NOTDONE

Non-smoker (never smoked)

Start date

SMOKSDT

Stop date

SMOKEDT

Ex smoker
d d m m y y y y d d m m y y y y

Current smoker
d d m m y y y y

If Ex or Current smoker, complete detail section below, checking all that apply

SMOKTYPE
Cigars per day. Round to next highest whole number. Specify average daily amount smoked during the time the subject smoked. bowls per day. Round to next highest whole number. Specify average daily amount smoked during the time the subject smoked. per day. Round to next highest whole number. Specify average daily amount smoked during the time the subject smoked. TOBACONO

Pipe

Cigarettes

Page 136 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) Table 1 presents data that could be captured using the detailed smoking status CRF presented in Figure 9.2.3.1 using the Sponsors Operational database structure. Row 1: Subject is a non-smoker. Row 2: Subject is an ex-smoker. Detailed substance use information is provided. Rows 3, 4: Subject is an ex-smoker of more than one type of tobacco. Detailed substance use information is provided. Row 5: Subject is a current smoker. Detailed substance use information is provided. Rows 6, 7: Subject came in for 2 visits, for the first visit subject was an ex-smoker. At the 2nd visit, subject began to smoke again. Detailed substance use information is provided. Row 8: Subject did not have an assessment done. Table 1: Data Captured Using the Detailed Smoking Status CRF in the Sponsors Operational Database
STUDYID Row 1 Row 2 Row 3 Row 4 Row 5 Row 6 Row 7 Row 8 1234 1234 1234 1234 1234 1234 1234 1234 USUBJID 1234005 1234006 1234007 1234007 1234008 1234009 1234009 1234010 VISIT 1 1 1 1 1 1 2 1 SEQ 1 1 1 2 1 1 2 1 SMOKCAT NON EX EX EX CURRENT EX CURRENT CIGARS PIPE CIGARETTES CIGARETTES CIGARS CIGARETTES 1 1 10 20 1 20 1977 01-01-2000 01-01-1990 1945 1965 01-01-2003 1 1970 2003 04-02-2003 04-02-1993 SMOKTYPE TOBACONO SMOKSDT SMOKEDT NOTDONE

Table 2 presents the data structured according to the SDTM. If this was a multi-page CRF, collecting multiple substance use information (for example, tobacco use and alcohol consumption) then SUCAT would have values of TOBACCO and ALCOHOL. As the primary physiological impact of tobacco use is due to nicotine content, the value of SUDECOD was set to NICOTINE. SUSTRF and SUENRF hold the values of SMOKCAT as they are interpreted based on the subject's use status at sponsor-defined start and end reference timeframes. Row 1: Because this subject did not consume tobacco products, the value of SRTRT was set to 'Tobacco'. In addition, the values of SUSTRF and SUENRF are set to NULL, and the value of SUOCCUR could be set to "N" because this subject did not use any tobacco products. Row 8: Because the assessment for this subject was not done, the value of SUSTAT was set to NOT DONE.
CDISC, 2005. All rights reserved FINAL Page 137 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Table 2: Data from Example 3 Modeled in SDS V3.x SU Domain

STUDYID Row 1 Row 2 Row 3 Row 4 Row 5 Row 6 Row 7 Row 8 1234 1234 1234 1234 1234 1234 1234 1234 USUBJID 1234005 1234006 1234007 1234007 1234008 1234009 1234009 1234010 SUSEQ 1 1 1 2 1 1 2 1 SUTRT TOBACCO CIGARS PIPE SUDECOD NICOTINE NICOTINE NICOTINE SUCAT TOBACCO TOBACCO TOBACCO TOBACCO TOBACCO TOBACCO TOBACCO TOBACCO NOT DONE BEFORE BEFORE BEFORE BEFORE BEFORE DURING BEFORE BEFORE BEFORE ONGOING BEFORE ONGOING AVG/DAY AVG/DAY AVG/DAY AVG/DAY AVG/DAY AVG/DAY 1 1 10 20 1 20 SUSTAT SUSTRF SUENRF SUDOSFRQ SUDOSTOT

CIGARETTES NICOTINE CIGARETTES NICOTINE CIGARS NICOTINE

CIGARETTES NICOTINE

VISITNUM Row 1 (cont) Row 2 (cont) Row 3 (cont) Row 4 (cont) Row 5 (cont) Row 6 (cont) Row 7 (cont) Row 8 (cont) 1 1 1 1 1 1 2 1

SUSTDTC

SUENDTC

1977 2000-01-01 1990-01-01 1945 1965 2003-01-01

2003 2003-02-04 1993-02-04

1970

Not all variables specified in the CDISC SU domain are present on the example CRF. Since they are not used, the variables are not present on the SU dataset.

Page 138 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

9.3
9.3.1

EVENTS EXAMPLES AE Example

In this example, an AE CRF collects AE terms as free text. There is a section at the bottom of the form for free-text comments. AEs are collected one time per study. In this example, first study drug was administered to the subject on 13Oct2003 at 12:00. 3 AEs were reported. AEs are coded using MedDRA and the sponsors procedures include the possibility of modifying the reported term to aid in coding. Notes: 1. 2. 3. 4. 5. Not all variables specified in the CDISC AE domain are present on the CRF (e.g., AEACNOTH, AERELNST, AEDUR). Since they are not used, the variables are not present in the AE dataset. However, all required and expected variables are present and populated. Comments would be placed in the Comments domain (CO). The CRF is structured so that serious category variables (e.g., AESDTH, AESHOSP) are checked only when an overall seriousness question is answered 'Y.' If a subject had no adverse events reported, no records are included for that subject in the AE dataset. If a subject has a value for AEENDTC then AEENRF is null.

Sample data that illustrates the above:

STUDYID DOMAIN USUBJID AESEQ Row 1 ABC123 Row 2 ABC123 Row 3 ABC123 AE AE AE 123101 123101 123101 1 2 3 AETERM POUNDING HEADACHE PULMONARY EMBOLISM AESTDTC 2003-10-12 2003-10-21 AESEV SEVERE MODERATE AESER N N Y AEACN DOSE REDUCED DOSE REDUCED AESMIE AEENDTC 2003-10-12 AEMODIFY HEADACHE AEDECOD HEADACHE BACK PAIN PULMONARY EMBOLISM AEREL PROBABLY RELATED PROBABLY NOT RELATED

BACK PAIN FOR 6 HOURS 2003-10-13T13:05 2003-10-13T19:00 BACK PAIN

AEBODSYS Row 1 (cont) Row 2 (cont) Row 3 (cont) AEOUT Row 1 (cont) Row 2 (cont) Row 3 (cont) RECOVERED/RESOLVED RECOVERED/RESOLVED RECOVERING/RESOLVING NERVOUS SYSTEM DISORDERS VASCULAR DISORDERS

NOT APPLICABLE DEFINITELY NOT RELATED

MUSCULOSKELETAL AND CONNECTIVE TISSUE DISORDERS MODERATE

AESCONG AESDISAB AESDTH AESHOSP AESLIFE

AESTDY AEENDY AEENRF -1 1 -1 1 . AFTER

CDISC, 2005. All rights reserved FINAL

Page 139 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

9.3.2

DS Examples

Disposition Example 1: In this example, a DS CRF collects multiple disposition events at different time points in the study indicated by EPOCH. There are also several protocol milestones which are indicated by DSCAT = 'PROTOCOL MILESTONE'. DSTERM is populated with controlled terminology with the same value as DSDECOD except in the case when there is free text for DSTERM such as 'Subject moved'. In this case, the controlled terminology is only in DSDECOD (LOST TO FOLLOW-UP). Notes 1. 2. 3. Not all variables specified in the CDISC DS domain are present on the CRF (e.g., DSSCAT, DSSTDY). Since they are not used; the variables are not present on the DS dataset. However, all required and expected variables are present and populated. In the Events observation class, the date/time of the Disposition Event is the DSSTDTC variable. The date of collection of the disposition information is the DSDTC variable EPOCH can be derived from VISITNUM. In the example below, when VISITNUM = 0, then EPOCH = SCREENING. EPOCH is populated when DSCAT is a disposition event and null when DSCAT is a protocol milestone.

There are multiple disposition events for each subject: 1. Subject 123101 has 3 records to indicate the completion of 3 stages of the study, which are screening, treatment phase and follow-up. Protocol milestone records for informed consent and randomization are also included. 2. Subject 123102 is a screen drop. Screen drops are identified by a DSDECOD that is not equal to 'COMPLETED' for the SCREENING stage. This is an example of the submission of the verbatim reason for discontinuation in DSTERM. 3. Subject 123103 completed the screening stage but did not complete the treatment stage. 4. Subject 123104 died on October 29, 2003 (see DSSTDTC) after the completion of treatment, but prior to the completion of follow-up. Note that the date of collection of the event information was on October 31, 2003 (DSDTC). 5. Subject 123105 discontinued study treatment due to an AE, but went on to complete the follow-up phase of the trial.

Page 140 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) Sample data that illustrates the above:
STUDYID DOMAIN USUBJID DSSEQ Row 1 Row 2 Row 3 Row 4 Row 5 Row 6 Row 7 Row 8 Row 9 Row 10 Row 11 Row 12 Row 13 Row 14 Row 15 Row 16 Row 17 Row 18 Row 19 Row 20 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 ABC123 DS DS DS DS DS DS DS DS DS DS DS DS DS DS DS DS DS DS DS DS DS 123101 123101 123101 123101 123101 123102 123102 123103 123103 123103 123103 123104 123104 123104 123104 123104 123105 123105 123105 123105 123105 1 2 3 4 5 1 2 1 2 3 4 1 2 3 4 5 1 2 3 4 5 DSTERM DSDECOD DSCAT VISITNUM 0 0 1 2 3 0 0 0 0 1 1 0 0 1 2 1 0 0 1 2 3 SCREENING SCREENING EPOCH DSDTC DSSTDTC INFORMED CONSENT INFORMED CONSENT PROTOCOL MILESTONE OBTAINED OBTAINED COMPLETED COMPLETED DISPOSITION EVENT RANDOMIZED RANDOMIZED PROTOCOL MILESTONE COMPLETED COMPLETED DISPOSITION EVENT COMPLETED COMPLETED DISPOSITION EVENT INFORMED CONSENT INFORMED CONSENT PROTOCOL MILESTONE OBTAINED OBTAINED DISPOSITION EVENT SUBJECT DENIED MRI PROTOCOL PROCEDURE VIOLATION INFORMED CONSENT INFORMED CONSENT PROTOCOL MILESTONE OBTAINED OBTAINED COMPLETED COMPLETED DISPOSITION EVENT RANDOMIZED RANDOMIZED PROTOCOL MILESTONE SUBJECT MOVED LOST TO FOLLOW-UP DISPOSITION EVENT 2003-09-21 2003-09-21 2003-09-29 2003-09-29 2003-09-30 2003-09-30 TREATMENT 2003-10-31 2003-10-31 PHASE FOLLOW-UP 2003-11-15 2003-11-15 2003-11-21 2003-11-21 SCREENING 2003-11-22 2003-11-20 2003-09-15 2003-09-15 2003-09-22 2003-09-22 2003-09-30 2003-09-30

INFORMED CONSENT INFORMED CONSENT PROTOCOL MILESTONE OBTAINED OBTAINED COMPLETED COMPLETED DISPOSITION EVENT RANDOMIZED RANDOMIZED PROTOCOL MILESTONE COMPLETED COMPLETED DISPOSITION EVENT SUBJECT DIED DEATH DISPOSITION EVENT INFORMED CONSENT INFORMED CONSENT PROTOCOL MILESTONE OBTAINED OBTAINED COMPLETED COMPLETED DISPOSITION EVENT RANDOMIZED RANDOMIZED PROTOCOL MILESTONE ANEMIA ADVERSE EVENT DISPOSITION EVENT COMPLETED COMPLETED DISPOSITION EVENT

TREATMENT 2003-10-31 2003-10-31 PHASE 2003-09-15 2003-09-15 2003-09-22 2003-09-22 2003-09-30 2003-09-30 TREATMENT 2003-10-15 2003-10-15 PHASE FOLLOW-UP 2003-10-31 2003-10-29 2003-09-28 2003-09-28 2003-10-02 2003-10-02 2003-10-02 2003-10-02 TREATMENT 2003-10-17 2003-10-17 PHASE FOLLOW-UP 2003-11-02 2003-11-02 SCREENING SCREENING

Row 21 ABC123

CDISC, 2005. All rights reserved FINAL

Page 141 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1) Disposition Example 2: In this example, the sponsor has chosen to simply submit whether or not the subject completed the study. Examples are shown of 1 subject who completed the study and 2 who discontinued prior to completion. In this very straightforward situation, only required and expected variables have been included.
STUDYID Row 1 Row 2 Row 3 ABC456 ABC456 ABC456 DOMAIN DS DS DS USUBJID 456101 456102 456103 DSSEQ 1 1 1 DSTERM COMPLETED SUBJECT TAKING STUDY MED ERRATICALLY LOST TO FOLLOW-UP DSDECOD COMPLETED PROTOCOL VIOLATION LOST TO FOLLOW-UP DSSTDTC 2003-09-21 2003-09-29 2003-10-15

Disposition Example 3: In this example, the sponsor collects whether or not the subject completes the treatment and follow-up phases as well as whether the subjects treatment assignment was unblinded. The date of the unblinding is represented in DSSTDTC. Note that maintaining the blind as per protocol is not considered to be an event since there is no change in the subjects state.
STUDYID Row 1 Row 2 Row 3 Row 4 Row 5 ABC789 ABC789 ABC789 ABC789 ABC789 DOMAIN DS DS DS DS DS USUBJID 789101 789101 789102 789102 789102 DSSEQ 1 2 1 2 3 DSCAT DISPOSITION EVENT DISPOSITION EVENT DISPOSITION EVENT EPOCH TREATMENT PHASE FOLLOW-UP TREATMENT PHASE DSTERM COMPLETED COMPLETED SKIN RASH DSDECOD COMPLETED COMPLETED ADVERSE EVENT DSSTDTC 2004-09-12 2004-12-20 2004-09-30 2003-10-01 2004-12-28

OTHER EVENTS TREATMENT SUBJECT HAD TREATMENT PHASE SEVERE RASH UNBLINDED DISPOSITION EVENT FOLLOW-UP COMPLETED COMPLETED

Page 142 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

9.3.3 MH Example:
In this example, a study collects medical history information using 3 separate CRFs. A general medical history CRF collects descriptions of events by body system (e.g., Endocrine, Metabolic) and asks whether the conditions were ongoing or not at study start (Rows 1-3). The reported events are coded using MedDRA. Another CRF asks whether the subject had any of a list of specific risk factors. The third CRF collects the primary diagnosis and onset date. Risk factors and primary diagnosis are not coded using MedDRA. Notes a. b. c. d. e. Not all variables specified in the CDISC MH domain are present on the CRF (e.g., MHENDTC). Since they are not used, the variables are not present in the MH dataset shown. However, all required variables are present and populated. Comments, if provided, would be placed in the Comments domain (CO). MHCAT is used to specify the category of data collected (GENERAL MEDICAL HISTORY, RISK FACTORS, PRIMARY DIAGNOSIS). MHSCAT displays the body systems specified on the general medical history CRF. The Ongoing at Study Start question on the general medical history has been translated to MHENRF. If Yes is specified, MHENRF=DURING/AFTER; if No is specified, MHENRF=BEFORE.
MHTERM ASTHMA FREQUENT HEADACHES BROKEN LEG HYPERTENSION TIA STROKE MHSTDTC MHDECOD ASTHMA HEADACHE BONE FRACTURE MHCAT GENERAL MEDICAL HISTORY GENERAL MEDICAL HISTORY GENERAL MEDICAL HISTORY RISK FACTORS RISK FACTORS PRIMARY DIAGNOSIS MHENRF DURING/AFTER DURING/AFTER BEFORE MHSCAT RESPIRATORY CNS OTHER

Sample data that illustrates the above:

STUDYID DOMAIN USUBJID MHSEQ Row 1 ABC123 Row 2 ABC123 Row 3 ABC123 Row 4 ABC123 Row 5 ABC123 Row 6 ABC123 MH MH MH MH MH MH 123101 123101 123101 123101 123101 123101 1 2 3 5 6 4 MHBODSYS Row 1 (cont) Row 2 (cont) Row 3 (cont) Row 4 (cont) Row 5 (cont) Row 6 (cont) 2004-09-17T07:30 RESPIRATORY SYSTEM DISORDERS CENTRAL AND PERIPHERAL NERVOUS SYSTEM DISORDERS MUSCULOSKELTAL SYSTEM DISORDERS

CDISC, 2005. All rights reserved FINAL

Page 143 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

9.4
9.4.1

FINDINGS EXAMPLES
EG Examples

ECG Example 1:
This is an example of an ECG dataset for one subject at one visit where cycle measurements, findings, and an interpretation were delivered to the sponsor by an ECG provider. EGREFID is an internal or external ECG identifier that links to the external ECG Waveform Files. EGXFN is the file name for the external ECG Waveform file. Rows 1-6 - Show how ECG cycle measurements are represented. Rows 2-6 - Show the data in original units of measure in EGORRES and in the converted value in EGSTRESC and EGSTRESN Rows 5-6 - Show QTCB and QTCF. The actual data should have a 'Y' in the EGDRVFL column since these results are derived by the sponsor in this example. See assumption 4.1.5.1. Rows 7-10 - Show how ECG findings are represented. Row 11 Row 12 - Shows a way of representing comments that are important to the overall understanding of the ECG, but which are not truly findings or interpretations. - Shows how interpretations are represented.

Page 144 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

STUDYID DOMAIN USUBJID XYZ-US-701-002 XYZ-US-701-002 XYZ-US-701-002 XYZ-US-701-002 XYZ-US-701-002 XYZ-US-701-002 XYZ-US-701-002 XYZ-US-701-002 XYZ-US-701-002 XYZ-US-701-002 XYZ-US-701-002 XYZ-US-701-002 EGORRES EGSEQ EGGRPID EGREFID EGTESTCD 1 2 3 4 5 6 7 8 9 10 11 12 EGORRESU bpm sec sec sec sec sec 1 1 1 1 1 1 1 1 1 1 1 1 334PT89 334PT89 334PT89 334PT89 334PT89 334PT89 334PT89 334PT89 334PT89 334PT89 334PT89 334PT89 EGSTRESC 62 150 103 406 469 446 ATRIAL FIBRILLATION ATRIAL FLUTTER QT PROLONGATION LEFT VENTRICULAR HYPERTROPHY INCORRECT ELECTRODE PLACEMENT ABNORMAL, CLINICALLY INSIGNIFICANT VRATE PR QRS QT QTCB QTCF RHYRATE RHYRATE QTABN VCABN TECHPROB INTP EGTEST Ventricular Rate PR Interval QRS Interval QT Interval QTc - Bazett QTc - Fridericia Rhythm and Rate Rhythm and Rate QT Abnormalities Ventricular Conduction Abnormalities Technical Problems Interpretation EGEVAL EGXFN PQW436789-07 PQW436789-07 PQW436789-07 PQW436789-07 PQW436789-07 PQW436789-07 PQW436789-07 PQW436789-07 PQW436789-07 EGNAM Test Lab Test Lab Test Lab Test Lab Test Lab Test Lab Test Lab Test Lab Test Lab Test Lab Test Lab EGPOS SUPINE SUPINE SUPINE SUPINE SUPINE SUPINE SUPINE SUPINE SUPINE SUPINE SUPINE SUPINE EGDY -36 -36 -36 -36 -36 -36 -36 -36 -36 -36 -36 -36

Row 1
Row 2 Row 3 Row 4 Row 5 Row 6 Row 7 Row 8 Row 9 Row 10 Row 11 Row 12

XYZ XYZ XYZ XYZ XYZ XYZ XYZ XYZ XYZ XYZ XYZ XYZ

EG EG EG EG EG EG EG EG EG EG EG EG VISIT

EGSTRESN EGSTRESU 62 150 103 406 469 446 bpm msec msec msec msec msec

VISITNUM 1 1 1 1 1 1 1 1 1 1 1

EGDTC 2003-04-15T11:58 2003-04-15T11:58 2003-04-15T11:58 2003-04-15T11:58 2003-04-15T11:58 2003-04-15T11:58 2003-04-15T11:58 2003-04-15T11:58 2003-04-15T11:58 2003-04-15T11:58 2003-04-15T11:58 2003-04-15T11:58

62 0.15 0.103 0.406 0.469 0.446 ATRIAL FIBRILLATION Row 8 (cont) Screening 1 ATRIAL FLUTTER Row 9 (cont) Screening 1 QT PROLONGATION

Row 1 (cont) Row 2 (cont) Row 3 (cont) Row 4 (cont) Row 5 (cont) Row 6 (cont) Row 7 (cont)

Screening 1 Screening 1 Screening 1 Screening 1 Screening 1 Screening 1 Screening 1

Row 10 (cont) Screening 1 LEFT VENTRICULAR HYPERTROPHY Row 11 (cont) Screening 1 Row 12 (cont) Screening 1 INCORRECT ELECTRODE PLACEMENT ABNORMAL, CLINICALLY INSIGNIFICANT

PRINCIPAL INVESTIGATOR

CDISC, 2005. All rights reserved FINAL

Page 145 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

ECG Example 2:
Example 2 shows results for one subject across multiple visits where only the overall assessment was collected. In addition the ECG done at Visit 4 was read by the principal investigator and a cardiologist. In this example the EGGRPID is the same number and the EGSEQ increments by one. Row 2 - The only one selected as Baseline. Row 3 - Shows a date/time in ISO 8601 representation where both the date and time were collected.
STUDYID DOMAIN Row 1 Row 2 Row 3 Row 4 Row 5 ABC ABC ABC ABC ABC EG EG EG EG EG USUBJID ABC-99-CA-456 ABC-99-CA-456 ABC-99-CA-456 ABC-99-CA-456 ABC-99-CA-456 EGSEQ EGGRPID EGTESTCD EGTEST EGPOS EGORRES EGSTRESC 1 2 3 4 5 1 2 3 4 4 ECGINT ECGINT ECGINT ECGINT ECGINT ECG Result SUPINE NORMAL NORMAL

ECG Result SUPINE ABNORMAL ABNORMAL ECG Result SUPINE ABNORMAL ABNORMAL ECG Result SUPINE ABNORMAL ABNORMAL ECG Result SUPINE ABNORMAL ABNORMAL

EGSTRESN EGBLFL EGEVAL Row 1 (cont) Row 2 (cont) Row 3 (cont) Row 4 (cont) Row 5 (cont) Y PI PI PI PI Cardiologist

VISIT SCREEN I SCREEN II DAY 10 DAY 18 DAY 18

VISITNUM 1 2 3 4 4

VISITDY -2 -1 10 15 15

EGDTC 2003-11-26 2003-11-27 2003-12-07T09:02 2003-12-12 2003-12-12

EGDY -2 -1 10 15 15

Page 146 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

9.4.2

IE Example

This example shows records for four subjects, each of whom had a single inclusion/exclusion exception. Subject XYZ-0007 met exclusion criteria number 17, but was included in the trial. The other three subjects each failed to meet inclusion criteria number 3, but were included in the trial.
STUDYID DOMAIN Row 1 Row 2 Row 3 Row 4 XYZ XYZ XYZ XYZ IE IE IE IE USUBJID XYZ-0007 XYZ-0023 XYZ-0047 XYZ-0096 IESEQ 1 1 1 1 IETESTCD EXCL17 INCL03 INCL03 INCL03 IETEST Ventricular Rate IECAT EXCLUSION IESPID 17 3 3 3 IEORRES Y N N N IESTRESC Y N N N

Acceptable mammogram from local INCLUSION radiologist? Acceptable mammogram from local INCLUSION radiologist? Acceptable mammogram from local INCLUSION radiologist?

VISIT Row 1 (cont) Row 2 (cont) Row 3 (cont) Row 4 (cont) WEEK-8 WEEK-8 WEEK-8 WEEK-8

VISITNUM 1 1 1 1

VISITDY -63 -63 -63 -63

IEDTC 1999-01-10 1999-01-06 1999-01-12 1999-01-13

IEDY -58 -62 -56 -55

CDISC, 2005. All rights reserved FINAL

Page 147 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

9.4.3

LB Examples

Example 1:
Row 1 - Shows a value collected in one unit, but converted to selected standard unit.

Rows 2, 3, 4 - Rows 2 and 3 show two records for Alkaline Phosphatase done at the same visit, one day apart. Row 4 shows how to create a derived record (average of the records 2 and 3) and flagged derived and as the one to use as baseline. Rows 6, 7 Row 9 Row 10 - Show a suggested use of the LBSCAT variable. It could be used to further classify types of tests within a laboratory panel (i.e., DIFFERENTIAL). - Shows the proper use of the LBSTAT variable to indicate "NOT DONE", where a reason was collected when a test was not done. - Shows one subject's data collected at two different visits. For Visit 1 the subject had cholesterol measured. The normal range for this test is <200 mg/dL. The value <200 may not be used in the normal range variables LBORNRHI or LBSTNRHI; however, a sponsor may decide, for example, to enter '0' into LBORNRLO and '199' in LBORNRHI. The sponsor must define the appropriate value for all of the normal range variables.

STUDYID DOMAIN Row 1 Row 2 Row 3 Row 4 Row 5 Row 6 Row 7 Row 8 Row 9 Row 10 Row 11 ABC ABC ABC ABC ABC ABC ABC ABC ABC ABC ABC LB LB LB LB LB LB LB LB LB LB LB

USUBJID ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001

LBSEQ LBTESTCD 1 2 3 4 5 6 7 8 9 10 11 ALB ALKP ALKP ALKP WBC LYMPH NEUT PH ALB CHOL WBC

LBTEST Albumin Alkaline Phosphatase Alkaline Phosphatase Alkaline Phosphatase WBC Count Lymphocytes Neutrophils pH Albumin Cholesterol WBC Count

LBCAT

LBSCAT

LBORRES LBORRESU LBORNRLO LBORNRHI 34 398 350 5.9 6.7 5.1 7.5 229 5.9 g/L IU/L IU/L IU/L 10^9/L % 10^9/L 35 40 40 4 25 2 5.0 0 4 VISIT BASELINE BASELINE BASELINE 50 160 160 11 40 8 9.0 199 11

Chemistry Chemistry Chemistry Chemistry Hematology Hematology DIFFERENTIAL Hematology DIFFERENTIAL Urinalysis Chemistry Chemistry Hematology

mg/dL 10^9/L

LBSTRESC LBSTRESN LBSTRESU LBSTNRLO LBSTNRHI LB STAT LBREASND LBBLFL LBFAST LBDRVFL Row 1 (cont) Row 2 (cont) Row 3 (cont)
Page 148 August 26, 2005

VISITNUM LBDTC 1 1 1 1999-06-19 1999-06-19 1999-06-20

3.4 398 350

3.4 398 350

g/dL units/L units/L

3.5 40 40

5 160 160

Y Y Y

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) LBSTRESC LBSTRESN LBSTRESU LBSTNRLO LBSTNRHI LB STAT LBREASND LBBLFL LBFAST LBDRVFL Row 4 (cont) Row 5 (cont) Row 6 (cont) Row 7 (cont) Row 8 (cont) Row 9 (cont) Row10 (cont) Row11 (cont) 374 5.9 0.4 5.1 7.5 374 5.9 0.4 5.1 units/L 10^3/uL 10^9/L 10^9/L 40 4 1.2 2 5.00 160 11 3 8 9.00 NOT INSUFFICIENT DONE SAMPLE 229 5.9 229 5.9 mg/dL 10^3/uL 0 4 5.85 11 Y Y Y Y Y Y Y Y Y Y Y Y VISIT BASELINE BASELINE BASELINE BASELINE BASELINE VISIT 1 VISIT 1 VISIT 1 VISITNUM LBDTC 1 1 1 1 1 2 2 2 1999-06-19 1999-06-19 1999-06-19 1999-06-19 1999-06-19 1999-07-21 1999-07-21 1999-07-21

CDISC, 2005. All rights reserved FINAL

Page 149 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Example 2:
This is an example of three sequential timed urine collections measuring glucose. Rows 1 - Shows an example of a pre-dose urine collection interval (from 4 hours prior to dosing until 15 minutes prior to dosing) with a negative value for LBELTM that reflects the end of the interval in reference to the fixed reference LBTPTREF, the date of which is recorded in LBRFTDTC. --RFTDTC is a new timing variable added to version 1.1 of the SDTM, and is incorporated in SDTM 3.1.1 to meet this purpose.

Rows 2, 3 - Show an example of post-dose urine collection intervals with values for LBELTM that reflect the end of the intervals in reference to the fixed reference LBTPTREF, the date of which is recorded in LBRFTDTC.

STUDYID DOMAIN Row 1 Row 2 Row 3 ABC ABC ABC LB LB LB

USUBJID ABC-001-001 ABC-001-001 ABC-001-001

LBSEQ LBTESTCD 1 2 3

LBTEST

LBCAT Urine Urine Urine

LBORRES LBORRESU LBORNRLO LBORNRHI 7 11 9 VISIT BASELINE BASELINE BASELINE LBELTM -P15M P8H P16H LBTPTREF Dosing Dosing Dosing mg/dL mg/dL mg/dL 1 1 1 VISITNUM 1 1 1 LBRFTDTC 1999-06-19T08:00 1999-06-19T08:00 1999-06-19T08:00 15 15 15

GLUCOSE Glucose GLUCOSE Glucose GLUCOSE Glucose LBSTNRLO 0.1 0.1 0.1 LBTPT Pre-dose 0-8 hours after dosing 8-16 hours after dosing

LBSTRESC Row 1 (cont) Row 2 (cont) Row 3 (cont) 0.38 0.61 0.5 LBDTC Row 1 (cont) Row 2 (cont) Row 3 (cont)

LBSTRESN LBSTRESU 0.38 0.61 0.5 mmol/L mmol/L mmol/L LBENDTC 1999-06-19T07:45 1999-06-19T16:00 1999-06-20T00:00

LBSTNRHI 0.8 0.8 0.8

LBTPTNUM 1 2 3

1999-06-19T04:00 1999-06-19T08:00 1999-06-19T16:00

Page 150 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

9.4.4

PE Example

The example shows data for one subject collected at two different visits. In all of the records except 6 and 11 the data comes from the general physical examination. In this case PECAT is "General". Additional data collected on about an ophthalmologic examination is also added to this domain, refer to rows eight and thirteen. Row 1 Row 2 Row 4-6 Row 4-7 - Shows how PESTRESC is populated if result is 'NORMAL'. - Shows the proper use of the STAT variable to indicate "NOT DONE", and when PEREASND is used to indicate why a body system (PETEST) was not examined. - Show how PESPID is used to show the sponsor-defined identifier, which in this case is a sequence number used for identifying abnormalities within a body system. - Show how PESTRESC is populated if abnormality is dictionary coded.

Rows 8, 13 - Show how the PECAT variable can be used to indicate a different type of physical examination. In this case, the ophthalmologic examination may have been collected in a separate dataset in the operational database.

CDISC, 2005. All rights reserved FINAL

Page 151 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1) STUDYID DOMAIN USUBJID PESEQ PETESTCD Row 1 Row 2 Row 3 Row 4 Row 5 Row 6 Row 7 Row 8 Row 9 Row 10 Row 11 Row 12 Row 13 ABC ABC ABC ABC ABC ABC ABC ABC ABC ABC ABC ABC ABC PE PE PE PE PE PE PE PE PE PE PE PE PE PESTRESC Row 1 (cont) Row 2 (cont) Row 3 (cont) Row 4 (cont) Row 5 (cont) Row 6 (cont) Row 7 (cont) Row 8 (cont) Row 9 (cont) Row 10 (cont) Row 11 (cont) Row 12 (cont) Row 13 (cont) NORMAL NOT DONE NORMAL ACNE NOS MILD DERMATITIS MODERATE RASH SEVERE CARDIAC NORMAL NORMAL NORMAL NORMAL NORMAL NORMAL INVESTIGATOR ERROR ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 PESEV 1 2 3 4 5 6 7 8 9 10 11 12 13 PETEST PECAT PELOC PEBODSYS PEORRES

NORMAL HEAD Head GENERAL RESP Respiratory GENERAL NORMAL ENT Ear/nose/throat GENERAL SKIN Skin GENERAL FACE SKIN AND ACNE SKIN Skin GENERAL HANDS SKIN AND ALLERGIC REACTION SKIN Skin GENERAL LEFT ARM SKIN AND SKINRASH CARDIO Cardiovascular GENERAL INVESTIGATIONS SYSTOLIC NORMAL FUNDOSCP Fundoscopic OPHTALMOLOGIC NORMAL RESP Respiratory GENERAL NORMAL ENT Ear/nose/throat GENERAL NORMAL NECK Neck GENERAL NORMAL CARDIO Cardiovascular GENERAL NORMAL FUNDOSCP Fundoscopic OPHTALMOLOGIC PEREASND PESPID 1 1 1 1 2 3 1 1 1 1 1 1 1 VISIT BASELINE BASELINE BASELINE BASELINE BASELINE BASELINE BASELINE BASELINE VISIT 1 VISIT 1 VISIT 1 VISIT 1 VISIT 1 VISITNUM VISITDY 1 1 1 1 1 1 1 1 2 2 2 2 2 1 1 1 1 1 1 1 1 45 45 45 45 45 PEDTC 1999-06-06 1999-06-06 1999-06-06 1999-06-06 1999-06-07 1999-06-08 1999-06-06 1999-06-06 1999-07-21 1999-07-21 1999-07-21 1999-07-21 1999-07-21 PEDY -3 -3 -3 -3 -2 -1 -3 -3 45 45 45 45 45

PESTAT

Page 152 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

9.4.5

QS Examples

Example 1:
This is an example of data from a questionnaire from one subject at one visit with standard text answers that have an associated score. In this example the subject answered all of the questions in Rows 1-4 and Rows 7-9. The standard text (e.g., very good) translates to a score of 4.4. The value of 4.4 is populated in both QSSTRESN and QSSTRESC. Since this is the baseline data there is a flag in all records in QSBLFL. The values in Rows 5, 6, 10, and 11 are derived from previous records and are flagged with a Y in QSDRVFL. The example shows how the textual answer is handled in the QSORRES Variable, while the QSSTRESC and QSSTRESN contain the standardized score value. Rows 5, 6, 10, 11 - Show derived records. Notice how QSORRES is blank for derived records because there is no corresponding text value for the numeric value shown (see section 4.1.5.1). STUDYID Row 1 Row 2 Row 3 Row 4 Row 5 Row 6 Row 7 Row 8 Row 9 Row 10 Row 11 STUDYX STUDYX STUDYX STUDYX STUDYX STUDYX STUDYX STUDYX STUDYX STUDYX STUDYX DOMAIN QS QS QS QS QS QS QS QS QS QS QS QSORRES Row 1 (cont) Row 2 (cont) Row 3 (cont) Row 4 (cont) Row 5 (cont) Row 6 (cont) Row 7 (cont) Row 8 (cont) Row 9 (cont) Row 10 (cont) Row 11 (cont) VERY GOOD MOSTLY FALSE MOSTLY TRUE DEFINITELY FALSE NO NO NO USUBJID P0001 P0001 P0001 P0001 P0001 P0001 P0001 P0001 P0001 P0001 P0001 QSSEQ 1 2 3 4 5 6 7 8 9 10 11 QSTESTCD GH1 GH11A GH11B GH11C GH GHINDEX RP4A RP4B RP4C RP RPINDEX QSSTRESC 4.4 4 4 5 21.4 82 2 2 2 8 100 QSTEST Heath Sick a litter easier Healthy as anybody Expect health to get worse SF-36 General health perceptions SF-36 General health perceptions (0-100) Phys. Health-cut down time spent Phys. Health-accomplished less Phys. Health-limit kind of work SF-36 Role-physical SF-36 Role-physical (0-100) QSDRVFL VISITNUM 2 2 2 2 2 2 2 2 2 2 2 QSCAT SF36 SF36 SF36 SF36 SF36 SF36 SF36 SF36 SF36 SF36 SF36 QSSCAT GENERAL HEALTH GENERAL HEALTH GENERAL HEALTH GENERAL HEALTH GENERAL HEALTH GENERAL HEALTH ROLE-PHYSICAL ROLE-PHYSICAL ROLE-PHYSICAL ROLE-PHYSICAL ROLE-PHYSICAL QSDTC 2001-03-28 2001-03-28 2001-03-28 2001-03-28 2001-03-28 2001-03-28 2001-03-28 2001-03-28 2001-03-28 2001-03-28 2001-03-28 QSDY -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2

QSSTRESN 4.4 4 4 5 21.4 82 2 2 2 8 100

QSBLFL Y Y Y Y Y Y Y Y Y Y Y

VISIT BASELINE BASELINE BASELINE BASELINE BASELINE BASELINE BASELINE BASELINE BASELINE BASELINE BASELINE

Y Y

Y Y

CDISC, 2005. All rights reserved FINAL

Page 153 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

Example 2:
This example show data from one subject collected at one visit for a questionnaire with standard text answers. These answers cannot be converted to numeric type, so only QSSTRESC is populated. The derived record, however, does have a derived value in QSSTRESN. Notice how QSTPTNUM is used to organize the same question being asked at various time points on the same date where no time was collected. In rows 1-3 QSTEST is ARM and QSTPTNUM is incremented 1,2,3. When QSTEST changes to BUTTER the number in QSTPTNUM start again at 1. Row 11 - Shows derived records. Notice how QSORRES is blank for derived records because there is no corresponding text value for the numeric value shown (see section 4.1.5.1). STUDYID Row 1 Row 2 Row 3 Row 4 Row 5 Row 6 Row 7 Row 8 Row 9 Row 10 Row 11 STUDYX STUDYX STUDYX STUDYX STUDYX STUDYX STUDYX STUDYX STUDYX STUDYX STUDYX DOMAIN QS QS QS QS QS QS QS QS QS QS QS USUBJID P0001 P0001 P0001 P0001 P0001 P0001 P0001 P0001 P0001 P0001 P0001 QSBLFL QSDRVFL QSSEQ 1 2 3 4 5 6 7 8 9 10 11 QSTESTCD COG01T02 COG01T02 COG01T02 COG01T03 COG01T03 COG01T03 COG01T04 COG01T04 COG01T04 COG01T09 COG01X VISIT SCREENING SCREENING SCREENING SCREENING SCREENING SCREENING SCREENING SCREENING SCREENING SCREENING SCREENING QSTEST ARM ARM ARM BUTTER BUTTER BUTTER CABIN CABIN CABIN GRASS WORD RECALL VISITYDY -14 -14 -14 -14 -14 -14 -14 -14 -14 -14 -14 QSCAT ADAS ADAS ADAS ADAS ADAS ADAS ADAS ADAS ADAS ADAS ADAS QSSCAT WORD RECALL WORD RECALL WORD RECALL WORD RECALL WORD RECALL WORD RECALL WORD RECALL WORD RECALL WORD RECALL WORD RECALL WORD RECALL QSDY -10 -10 -10 -10 -10 -10 -10 -10 -10 -10 -10 QSORRE S NO NO NO NO NO NO NO NO NO NO QSSTRES C NO NO NO NO NO NO NO NO NO NO 9

QSSTRESN Row 1 (cont) Row 2 (cont) Row 3 (cont) Row 4 (cont) Row 5 (cont) Row 6 (cont) Row 7 (cont) Row 8 (cont) Row 9 (cont) Row 10 (cont) Row 11 (cont)

VISITNUM 1 1 1 1 1 1 1 1 1 1 1

QSDTC 2001-03-20 2001-03-20 2001-03-20 2001-03-20 2001-03-20 2001-03-20 2001-03-20 2001-03-20 2001-03-20 2001-03-20 2001-03-20

QSTPTNUM 1 2 3 1 2 3 1 2 3 1 99

Page 154 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

9.4.6

SC Example

The example below shows data that is collected once per subject that does not change during the trial and does not fit into the demographics domain. For this example the eye color and initials were collected. Rows 1, 4 Illustrate use of an additional Race response. In this case, the value of RACE in DM would be 'OTHER'.

STUDYID DOMAIN Row 1 Row 2 Row 3 Row 4 Row 5 Row 6 Row 7 Row 8 Row 9 Row 10 ABC ABC ABC ABC ABC ABC ABC ABC ABC ABC SC SC SC SC SC SC SC SC SC SC

USUBJID ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-002 ABC-001-002 ABC-001-002 ABC-001-003 ABC-001-003 ABC-002-001 ABC-002-001

SCSEQ SCTESTCD 1 2 3 1 2 3 2 3 2 3

SCTEST

SCORRES SCSTRESC JAPANESE JAPANESE BROWN HLT ALASKA NATIVE BLUE BAM GREEN ALM HAZEL CQH BROWN HLT ALASKA NATIVE BLUE BAM GREEN ALM HAZEL CQH

VISIT Baseline Baseline Baseline Baseline Baseline Baseline Baseline Baseline Baseline Baseline

VISITNUM 1 1 1 1 1 1 1 1 1 1

SCDTC 1999-06-19 1999-06-19 1999-06-19 1999-03-19 1999-03-19 1999-03-19 1999-05-03 1999-05-03 1999-06-14 1999-06-14

RACEOTH Race Other EYECD Eye Color SUBJINIT Subject Initials RACEOTH Race Other EYECD EYECD EYECD Eye Color Eye Color Eye Color

SUBJINIT Subject Initials SUBJINIT Subject Initials SUBJINIT Subject Initials

CDISC, 2005. All rights reserved FINAL

Page 155 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

9.4.7

VS Example

The example below shows one subject with two visits, Baseline and Visit 1, including examples of both collected and derived baseline measurements. Rows 1-2, 4-5, 8, 9 -VSTPT and VSTPTNUM are populated since more than one measurement was taken at this visit. Rows 3, 6 - Show an example of a derived value that was not considered to be an original result. In this case the sponsor derived the value in a different variable in the operational database. VSTPT and VSTPTNUM are not populated for these derived records. Rows 8, 9 - Show two temperatures taken at the baseline visit. Row 9 has a "Y" in the VSBLFL to indicate it was used as the baseline measurement. Row 14 Row 15 - Shows a value collected in one unit, but converted to selected standard unit. - Shows the proper use of the STAT variable to indicate "NOT DONE" where a reason was collected when a test was not done.

Page 156 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

STUDYIDDOMAIN Row 1 Row 2 Row 3 Row 4 Row 5 Row 6 Row 7 Row 8 Row 9 Row 10 Row 11 Row 12 Row 13 Row 14 Row 15 ABC ABC ABC ABC ABC ABC ABC ABC ABC ABC ABC ABC ABC ABC ABC VS VS VS VS VS VS VS VS VS VS VS VS VS VS VS

USUBJID ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001 ABC-001-001

VSSEQ VSTESTCD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 SYSBP SYSBP SYSBP DIABP DIABP DIABP PULSE TEMP TEMP WEIGHT HEIGHT SYSBP DIABP TEMP WEIGHT

VSTEST

VSPOS

VSLOC

VSORRES VSORRESU VSSTRESC 154 152 44 48 72 34.7 36.2 90.5 157 95 44 97.16 mmHg mmHg mmHg mmHg bpm C C kg cm mmHg mmHg F 154 152 153 44 48 46 72 34.7 36.2 90.5 157 95 44 36.2

Systolic Blood Pressure SITTING LEFT ARM Systolic Blood Pressure SITTING LEFT ARM Systolic Blood Pressure SITTING LEFT ARM Diastolic Blood Pressure SITTING LEFT ARM Diastolic Blood Pressure SITTING LEFT ARM Diastolic Blood Pressure SITTING LEFT ARM Pulse SITTING LEFT ARM Temperature MOUTH Temperature MOUTH Weight STANDING Height STANDING Systolic Blood Pressure SITTING LEFT ARM Diastolic Blood Pressure SITTING LEFT ARM Temperature MOUTH Weight VISIT BASELINE BASELINE BASELINE BASELINE BASELINE BASELINE BASELINE BASELINE BASELINE BASELINE BASELINE VISIT 1 VISIT 1 VISIT 1 VISIT 1 VISITNUM 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2

VSSTRESN VSSTRESU VSSTAT VSREASND VSBLFL VSDRVFL Row 1 (cont) Row 2 (cont) Row 3 (cont) Row 4 (cont) Row 5 (cont) Row 6 (cont) Row 7 (cont) Row 8 (cont) Row 9 (cont) Row 10 (cont) Row 11 (cont) Row 12 (cont) Row 13 (cont) Row 14 (cont) Row 15 (cont) 154 152 153 44 48 46 72 34.7 36.2 90.5 157 95 44 36.2 mmHg mmHg mmHg mmHg mmHg mmHg bpm C C kg cm mmHg mmHg C NOT DONE Subject refused

VSTPT BASELINE 1 BASELINE 2 BASELINE 1 BASELINE 2 BASELINE 1 BASELINE 2

VSTPTNUM VISITDY 1 2 1 2 1 2 1 1 1 1 1 1 1 1 1 1 1 35 35 35 35

Y Y Y Y Y Y

Y Y

CDISC, 2005. All rights reserved FINAL

Page 157 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

9.5
9.5.1

TRIAL DESIGN EXAMPLES

Defining Epochs - a How-to Example Using the Example Crossover Study

This example is based on the crossover study shown in Figure 4 of Section 7.4.2. Users may find it useful to construct a "study design matrix" like the following. Trial Screen P-5-10 5-P-10 5-10-P Screen Screen Screen 1st Treatment Placebo 5 mg 5 mg 1st Rest Rest Rest Rest 2nd Treatment 5 mg Placebo 10 mg 2nd Rest Rest Rest Rest 3rd Treatment 10 mg 10 mg Placebo Trial Follow-up Follow-up Follow-up Follow-up

This table is not submitted as part of the SDTM. It, like the figure in Section 7, is purely an aid to the user in working out the trial design datasets. Note that the row labels (in bold type) are the names of the Arms. The column headers (also in bold type) are the names of the trial Epochs. The contents of the cells are Elements (plain type). In many cases, the user will fill in the column headers first, then place the Elements in the cells of the table. In other cases, the user may find it easier to work out the Elements in each Arm, then to fill in the names of the Epochs. The Trial Arms table contains a record for each Element in each Arm (row of the study design matrix). TAETORD gives the order of the Elements within the Arm. The study design matrix table does not explicitly record the branch a diagram is better for displaying that feature of the trial design.

Page 158 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) Example of a full Trial Arms dataset for a crossover study:
ARM P-5-10 P-5-10 P-5-10 P-5-10 P-5-10 P-5-10 P-5-10 5-P-10 5-P-10 5-P-10 5-P-10 5-P-10 5-P-10 5-P-10 5-10-P 5-10-P 5-10-P 5-10-P 5-10-P 5-10-P 5-10-P TAETORD 1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 ELEMENT Screen Placebo Rest 5 mg Rest 10 mg Follow-up Screen 5 mg Rest Placebo Rest 10 mg Follow-up Screen 5 mg Rest 10 mg Rest Placebo Follow-up TABRANCH Randomized to Placebo - 5 mg - 10 mg

Randomized to 5 mg/Placebo/10 mg

Randomized to 5 mg/10 mg/Placebo

CDISC, 2005. All rights reserved FINAL

Page 159 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1) The Trial Elements dataset has a record for each unique Element. The properties of the Elements are not recorded in the trial design matrix, but the trial design matrix can be used to check that all Elements have been included in the Trial Elements dataset. The following information may be helpful in working out the start and end rules of the Elements: TESTRL answers the question, "What happens that marks the start of the Element?" or, "How do you know the subject has entered this Element?" TEENRL answers the question, "When should the subject leave this Element?" Note that this rule concerns when the subject should leave the Element. The subject actually leaves the Element when they enter the next Element. TEDUR is used when the element end rule says the subject should leave the Element after a fixed period of time. TEDUR records this fixed period of time as a duration using ISO 8601 notation. If the element end rule is of some different form, for example if it involves satisfying certain conditions, then TEDUR is left blank.
TESTRL Informed consent First dose of placebo First dose of 5 mg drug First dose of 10 mg drug 48 hrs after last dose drug 48 hrs after last dose drug TEENRL 2 weeks after start of element 2 weeks after start of element 2 weeks after start of element 2 weeks after start of element 1 week after start of element 3 weeks after start of element TEDUR P14D P14D P14D P14D P7D P21D

Example of the Trial Elements dataset for the same crossover study:
ELEMENT Screen Placebo 5 mg 10 mg Rest Follow-up

The Trial Elements dataset contains information on timing (the start and end rules), while the Trial Arms dataset contains information on how the Elements are assembled to make the Arms. Users may find it useful to join the two datasets (merge on Element) to display all this information in one table. The table below shows such a join of Trial Arms and Trial Elements. A column for EPOCH has been included in this table. Sorting this dataset by TAETORD, then ARM, makes it easy to check for consistency of the start and end rules of the Elements within an Epoch, across different Arms. Note that in this example, for each group of records with the same value of TAETORD and EPOCH, the Start Rules are the same or similar. For the Treatment Epochs, the start rules are "similar" in that they differ only in the specific dose (Placebo, 5 mg or 10 mg) that starts the element. The end rules for the Treatment Epochs are the same. For all the other Epochs in this trial, all the Elements within the Epoch are identical, so the start and end rules are also identical. If a group of Elements that have been labeled as belonging to the same Epoch do not have similar start and end rules, then they probably do not really constitute a valid Epoch. For a blinded trial, it may be useful to think about a "blinded start rule" or "epoch start rule." While the trial is still blinded, participants will not really know whether they have entered a particular treatment Element, only that they have entered a particular Epoch. For example, the element start rules for the 1st Trt Epoch specify receiving a particular dose (placebo, 5 mg or 10 mg), but during the trial, participants will know only that they have received the dose for that Epoch. Un-blinding the study reveals which treatment the subject received, and therefore which Element they passed through.

Page 160 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) Join of Trial Arms and Trial Elements datasets (merged by Element), sorted by TAETORD, then ARM:
ARM 5-P-10 P-5-10 5-10-P 5-P-10 P-5-10 5-10-P 5-P-10 P-5-10 5-10-P 5-P-10 P-5-10 5-10-P 5-P-10 P-5-10 5-10-P 5-P-10 P-5-10 5-10-P 5-P-10 P-5-10 5-10-P TAETORD 1 1 1 2 2 2 3 3 3 4 4 4 5 5 5 6 6 6 7 7 7 ELEMENT Screen Screen Screen 5 mg Placebo 5 mg Rest Rest Rest Placebo 5 mg 10 mg Rest Rest Rest 10 mg 10 mg Placebo Follow-up Follow-up Follow-up TASTRL Informed consent Informed consent Informed consent First dose of 5 mg drug First dose of placebo First dose of 5 mg drug 48 hrs after last dose drug 48 hrs after last dose drug 48 hrs after last dose drug First dose of placebo First dose of 5 mg drug First dose of 10 mg drug 48 hrs after last dose drug 48 hrs after last dose drug 48 hrs after last dose drug First dose of 10 mg drug First dose of 10 mg drug First dose of placebo 48 hrs after last dose drug 48 hrs after last dose drug 48 hrs after last dose drug TAENRL 2 weeks after start of element 2 weeks after start of element 2 weeks after start of element 2 weeks after start of element 2 weeks after start of element 2 weeks after start of element 1 week after start of element 1 week after start of element 1 week after start of element 2 weeks after start of element 2 weeks after start of element 2 weeks after start of element 1 week after start of element 1 week after start of element 1 week after start of element 2 weeks after start of element 2 weeks after start of element 2 weeks after start of element 3 weeks after start of element 3 weeks after start of element 3 weeks after start of element TADUR P14D P14D P14D P14D P14D P14D P7D P7D P7D P14D P14D P14D P7D P7D P7D P14D P14D P14D P21D P21D P21D EPOCH Trial Screen Trial Screen Trial Screen 1st Trt 1st Trt 1st Trt 1st Rest 1st Rest 1st Rest 2nd Trt 2nd Trt 2nd Trt 2nd Rest 2nd Rest 2nd Rest 3rd Trt 3rd Trt 3rd Trt Trial FU Trial FU Trial FU TABRANCH Randomized to 5 mg/Placebo/10 mg Randomized to Placebo/5 mg/10 mg Randomized to 5 mg/10 mg/Placebo

CDISC, 2005. All rights reserved FINAL

Page 161 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

9.5.2

Trial Summary Example

Trial Summary Example1: Rows 1 and 2: This trial is of two different types so there are 2 records for the type of trial Rows 1,2,5,6,7 and 12 use controlled terminology for TSVAL STUDYID Row 1 Row 2 Row 3 Row 4 Row 5 Row 6 Row 7 Row 8 Row 9 Row 10 Row 11 Row 12 Row 13
XYZ XYZ XYZ

DOMAIN
TS TS TS

TSSEQ
1 2 1

TSPARMCD
TYPE TYPE TITLE

TSPARM
Type Of Trial Type Of Trial Trial Title

TSVAL
EFFICACY SAFETY A 24 Week Study of Daily Oral Investigational Drug vs. Placebo in Subjects with Asthma Reduce the incidence of exacerbations of asthma Y DOUBLE BLIND PARALLEL Investigational New Drug Asthma 18 70 BOTH Placebo

XYZ XYZ XYZ XYZ XYZ XYZ XYZ XYZ XYZ XYZ

TS TS TS TS TS TS TS TS TS TS

1 1 1 1 1 1 1 1 1 1

OBJPRIM RANDOM MASK DESIGN TRT INDIC AGEMIN AGEMAX SEX COMPTRT

Primary Objective Of Trial Trial Is Randomized Trial Blinding Schema Description of Trial Design Reported Name Of Test Product Trial Indications Planned Minimum Age Of Subjects Planned Maximum Age Of Subjects Sex Of Participants Comparative Treatment Name

Page 162 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

10 Appendices
10.1 CDISC SDS TEAM
Role Team Lead Team Co-Leads Name Wayne Kubick Fred Wood Mary Lenzen Tom Guinter Technical Manager Sub-team Leads Julie Evans Randall Austin Chris Tolk Madhavi Vemuri Diane Wold Team Participants Karen Alexander Gayle Amato Jane Boone Gary Cunningham Kathleen Giordano Dan Godoy Andreas Gromen Joyce Hernandez Sandy Lei Musa Nsereko Bill Qubeck Robert Sarate Sarah Shiue Gail Stoner Mike Walega Gary Walker Yuguang Zhao FDA Participants Randy Levin Norman Stockbridge Steve Wilson Company Lincoln Technologies, CDISC Procter and Gamble Pfizer Octagon Research CDISC GlaxoSmithKline Bayer Healthcare Johnson and Johnson PRD GlaxoSmithKline Boehringer-Ingelheim Bayer Healthcare Webclin AstraZeneca Merck AstraZeneca Schering AG IBM Johnson and Johnson PRD Cephalon Pfizer Merck Merck Centocor Covance Quintiles Sanofi-Aventis FDA FDA FDA

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

10.2

GLOSSARY OF TERMS

The following abbreviations and terms are used in this document. Additional definitions can be found in the CDISC Glossary available at http://www.cdisc.org/glossary/index.html. CDISC CRT Dataset Domain FDA HL7 SDS SDTMIG SEND SDSIG SDTM V3.x Clinical Data Interchange Standards Consortium Case Report Tabulation A collection of structured data in a single file A collection of observations with a topic-specific commonality about a subject Food and Drug Administration Health Level 7 Submission Data Standards (Team) "Study Data Tabulation Model Implementation Guide for Human Clinical Trials (this document) Standard for Exchange of Non-clinical Data SDS Implementation Guide V. 3.1, now referred to as SDTMIG. Study Data Tabulation Model Refers to Version 3.1 of the SDSIG and all subsequent versions of the SDTMIG

Page 164 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

10.3

STANDARDIZED AND RESERVED CODES

This appendix includes a preliminary list of standard codes that have been defined by the CDISC SDS Team for use with the SDTM. It includes a set of reserved domain codes, test codes for two findings domains, and an initial set of name codes for use in the Supplemental Qualifiers (SUPPQUAL) special-purpose dataset. Since a robust set of controlled terminology is so essential for effective use of the SDTM and other CDISC models, CDISC has assigned a new "Controlled Terminology" team. Beginning in March 2005, CDISC will be publishing sets of proposals for controlled terminology in separate documentation packages that will be available through the CDISC website for review and comment. The set of controlled terminology will be expanded over time and will eventually be available through online search tools provided by the National Cancer Institute of the United States National Institutes of Health.

10.3.1

Reserved Domain Codes

The following domain codes have been reserved for use with the domain topicalities listed. CDISC will be preparing additional domain models to describe many of these over time. Code AE AU BM BR CM CO DA DC DM Domain Adverse Events Autopsy Bone Mineral Density (BMD) Data Biopsy Concomitant Meds Comments Drug Accountability Disease Characteristics Demographics Class Events Findings Findings Findings Interventions Special Purpose Findings Findings Special Purpose Description See 6.2.1.1, assumption 1. Additional data about deaths, specific to findings from autopsies. Findings resulting from evaluations of bone mineral density. Findings resulting from biopsies. See 6.1.1.1, assumption 1. See 5.1.2. Data regarding the accountability of study drug, such as information on the receipt, dispensing, return, and packaging. Data related to the characteristics of a specific disease state under study for a subject. Demographics includes a set of essential standard variables that describe each subject in a clinical study. It is the parent domain for all other observations for human clinical subjects. See SDTM 2.2.6. See 6.2.2.1, assumption 1. See separate DV domain model, assumption 1. Findings resulting from electroencephalogram (EEG) tests. Findings related to the collection of ECG data, including position of the subject, method of evaluation, all cycle measurements and all findings from the ECG including an overall interpretation if collected or derived. See 6.1.2.1, assumption 1. Healthcare resource utilization data such as subject visits to physicians, hospitalizations, and nursing home stays. See 6.3.2.1, assumption 1.
Page 165 August 26, 2005

DS DV EE EG

Disposition Protocol Deviations EEG ECG

Events Events Findings Findings

EX HU IE

Exposure Healthcare Resource Utilization Inclusion / Exclusion

Interventions Findings Findings

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) Code IM LB Imaging Laboratory Data Domain Class Findings Findings Description Findings resulting from diagnostic medical imaging tests such as X-ray or MRI. Laboratory test findings including, but is not limited to hematology, clinical chemistry and urinalysis data. Does not include microbiology or PK data, which are stored in separate domains. Microbiology Specimen findings, including gram stain results, and organisms found. See 6.2.3.1, assumption 1 Information regarding the subject's meal consumption, such as fluid intake, amounts, form (solid or liquid state), frequency, etc., typically used for PK analysis. Microbiology Susceptibility Test results, plus results of any other organism-related tests. Findings from organ measurement evaluations. Concentrations of drugs/metabolites in fluids or tissues as a function of time. Findings collected during a physical examination of the subject. It has findings that are discovered that are related to body systems. Does not include vital signs measurements, which are stored in the VS domain. Pharmacokinetic parameters derived from pharmacokinetic concentration-time (PC) data. Pharmacogenomics findings initially focusing on Genotype and Gene Expression data. See 6.3.5.1, assumption 1 See 6.3.6.1, assumption 1 See SDTM 3.3. Surgical information. Findings from diagnostic skin tests. Findings from diagnostic sleep tests (polysomnography). Evidence of disease or condition, including objective signs and subjective symptoms that are typically observed by a physician or described to the investigator by the subject. Findings from exercise stress tests. See 6.1.3.1, assumption 1 See SDTM 3.3. See SDTM 3.2. See SDTM 3.2. See SDTM 3.4. See SDTM 3.5. See SDTM 3.2. Measurements including but not limited to blood pressure, temperature, respiration, body surface area, BMI, height and weight.

MB MH ML

Microbiology Specimens Medical History Meal Data

Findings Events Interventions

MS OM PC PE

Microbiology Susceptibility Organ Measurements PK Concentration Physical Exam

Findings Findings Findings Findings

PP PG QS SC SE SG SK SL SS

PK Parameters Pharmacogenomics Questionnaires Subject Characteristics Subject Elements Surgery Skin Test Sleep (Polysomnography) Data Signs and Symptoms

Findings Findings Findings Findings Study Design Interventions Findings Findings Findings

ST SU SV TA TE TI TS TV VS

Stress (Exercise) Test Data Substance Use Subject Visits Trial Arms Trial Elements Trial Inclusion/Exclusion Criteria Trial Summary Trial Visits Vital Signs

Findings Interventions Trial Design Trial Design Trial Design Trial Design Trial Design Trial Design Findings

Page 166 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

10.3.2

Electrocardiogram Test Codes (for measured or calculated parameters)

EGTESTCD EGTEST Heart Rate ECG Interpretation PR Interval QRS Interval QT Interval QTc Interval Bazett QTc Interval Fridericia RR Interval Ventricular Rate

The following table contains the test codes suggested by CDISC for use in Electrocardiogram data domains. HR INTP PR QRS QT QTCB QTCF RR VRATE

10.3.3

Vital Signs Test Codes

The following table contains the test codes suggested by CDISC for use in Vital Signs domains.

VSTESTCD BMI BODYFAT BSA DIABP FRMSIZE HEIGHT HR MAP PULSE RESP SYSBP TEMP WEIGHT

VSTEST Body Mass Index Body Fat Body Surface Area Diastolic Blood Pressure Frame Size Height Heart Rate Mean Arterial Pressure Pulse Rate Respiratory Rate Systolic Blood Pressure Temperature Weight

CDISC, 2005. All rights reserved FINAL

Page 167 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

10.3.4

Supplemental Qualifiers Name Codes

The following table contains an initial set of standard name codes suggested by CDISC for use in the Supplemental Qualifiers (SUPPQUAL) special-purpose dataset.

QNAM AESOSP AETRTEM COMPLT FULLSET ITT PPROT SAFETY

QLABEL Other Medically Important SAE (includes the descriptive text, not a flag) Treatment Emergent Flag Completers Population Flag Full Analysis Set Flag Intent to Treat Population Flag Per Protocol Set Flag Safety Population Flag

Page 168 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

10.3.5

Trial Summary Codes

The following table contains proposed controlled terminology for use in the Trial Summary domain. Although this is the most current set of terminology available as of this writing, it is likely to grow or change as it undergoes further review by the CDISC Controlled Terminology team. TSPARMCD ADDON AEDICT TSPARM TEST PRODUCT IS ADDED ON TO EXISTING TREATMENT ADVERSE EVENT DICTIONARY TSVAL Y N MedDRA WHOART SPONSOR DEFINED IN UTERO PRETERM NEWBORN INFANTS NEWBORN (0-27 DAYS) INFANT AND TODDLER (28 DAYS - 23 MONTHS) CHILDREN (2-11 YEARS) ADOLESCENT (12-17 YEARS) ADULT (18-65) ELDERLY (> 65) No controlled terminology. No controlled terminology. OPEN LABEL SINGLE BLIND DOUBLE BLIND No controlled terminology. PLACEBO ACTIVE NONE HEALTHY SUBJECTS Use controlled terminology for interventions (DOSE). Use controlled terminology for interventions (DOSFRQ). Use controlled terminology for interventions (DOSU). WHODRUG No controlled terminology. DIAGNOSIS CURE MITIGATION [Note: used narrowly to mean mitigate adverse effect of another treatment] TREATMENT PREVENTION
Page 169 August 26, 2005

AGESPAN

AGE SPAN

AGEMAX AGEMIN BLIND

PLANNED MAXIMUM AGE OF SUBJECTS PLANNED MINIMUM AGE OF SUBJECTS TRIAL BLINDING SCHEMA

COMPTRT CONTROL

COMPARATIVE TREATMENT NAME TYPE OF CONTROL

DIAGGRP DOSE DOSFRQ DOSU DRUGDICT INDIC INDICTYP

DIAGNOSIS GROUP TEST PRODUCT DOSE PER ADMINISTRATION TEST PRODUCT DOSING FREQUENCY TEST PRODUCT DOSE UNITS DRUG DICTIONARY TRIAL INDICATIONS TRIAL INDICATION TYPE

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) TSPARMCD LENGTH MHDICT OBJPRIM OBJSEC PHASE TSPARM LENGTH OF TRIAL MEDICAL HISTORY DICTIONARY TRIAL PRIMARY OBJECTIVE TRIAL SECONDARY OBJECTIVE TRIAL PHASE TSVAL No controlled terminology. MedDRA SPONSOR DEFINED No controlled terminology however, the objective should be described in terms of the desired statement in labeling. No controlled terminology however the objective should be described in terms of the desired statement in labeling. 1 1-2 2 2A 2B 2-3 3 3A 3B 3-4 4 5 NA No controlled terminology. Y N Use controlled terminology for interventions (ROUTE). M F BOTH No controlled terminology. No controlled terminology. No controlled terminology. SAFETY EFFICACY BIO-EQUIVALENCE BIO-AVAILABILITY CONFIRMATORY EXPLORATORY PHARMACOECONOMIC PHARMACOGENOMICS PHARMACOKINETICS PHARMACODYNAMICS

PLANSUB RANDOM

PLANNED NUMBER OF SUBJECTS TRIAL IS RANDOMIZED

ROUTE SEXPOP

TEST PRODUCT ROUTE OF ADMINISTRATION SEX OF PARTICIPANTS

SPONSOR TITLE TRT TYPE

SPONSORING ORGANIZATION TRIAL TITLE REPORTED NAME OF TEST PRODUCT TYPE OF TRIAL

Page 170 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

10.4

CDISC VARIABLE-NAMING FRAGMENTS

The CDISC SDS group has defined a standard list of fragments to use as a guide when naming variables in SUPPQUAL (QNAM) or assigning TESTCD values that could conceivably be treated as variables in a horizontal listing derived from a V3.x dataset. In some cases, more than one fragment is used for a given keyword. This is necessary when a shorter fragment must be used for a TESTCD or QNAM that incorporates several keywords that must be combined while still meeting the 8-character variable naming limit of SAS transport files. When using fragments, the general rule is to use the fragment(s) that best conveys the meaning of the variable within the 8character limit; thus, the longer fragment should be used when space allows. If the combination of fragments still exceeds 8 characters, a character should be dropped where most appropriate (while avoiding naming conflicts if possible) to fit within the 8-character limit. In other cases the same fragment may be used for more than one meaning, but these would not normally overlap for the same variable.
Keyword(s) ACTION ADJUSTMENT ANALYSIS DATASET BASELINE BIRTH BODY CANCER CATEGORY CHARACTER CONDITION CLASS CODE COMMENT CONCOMITANT CONGENITAL DATE TIME - CHARACTER DAY DEATH DECODE DERIVED DESCRIPTION DISABILITY DOSE, DOSAGE DURATION ELAPSED ELEMENT EMERGENT END
CDISC, 2005. All rights reserved FINAL

Fragment ACN ADJ AD BL BRTH BOD CAN CAT C CND CLAS CD COM CON CONG DTC DY DTH DECOD DRV DESC DISAB DOS, DOSE DUR EL ET EM END, EN ETHNICITY EXTERNAL EVALUATOR

Keyword(s)

Fragment ETHNIC X EVAL EVL FAST FN FL FRM FREQ GR GRP HI HOSP ID INDC IND INT INTP INV LIFE LOC LOINC LO MIE NAM NST NR ND
Page 171 August 26, 2005

EVALUATION FASTING FILENAME FLAG FORMULATION, FORM FREQUENCY GRADE GROUP HIGHER LIMIT HOSPITALIZATION IDENTIFIER INDICATION INDICATOR INTERVAL INTERPRETATION INVESTIGATOR LIFE-THREATENING LOCATION LOINC CODE LOWER LIMIT MEDICALLY-IMPORTANT EVENT NAME NON-STUDY THERAPY NORMAL RANGE NOT DONE

CDISC SDTM Implementation Guide (Version 3.1.1)

Keyword(s) NUMBER NUMERIC ONGOING ORDER ORIGIN ORIGINAL OTHER OUTCOME OVERDOSE PARAMETER PATTERN POPULATION POSITION QUALIFIER REASON REFERENCE REGIMEN RELATED RELATIONSHIP RESULT RULE SEQUENCE SERIOUS SEVERITY Fragment NUM N ONGO ORD ORIG OR OTH, O OUT OD PARM PATT POP POS QUAL REAS REF, RF RGM REL, R REL RES RL SEQ S, SER SEV SPECIMEN SPONSOR STANDARD START STATUS SUBCATEGORY SUBJECT SUPPLEMENTAL SYSTEM TEXT TIME TIMEPOINT TOTAL TOXICITY TRANSITION TREATMENT UNIT UNIQUE UNPLANNED VARIABLE VALUE VEHICLE Keyword(s) Fragment SPEC, SPC SP ST, STD ST STAT SCAT SUBJ SUPP SYS TXT TM TPT TOT TOX TRANS TRT U U UP VAR VAL V

Page 172 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

10.5

LESSONS LEARNED FROM THE PILOT

After reviewing the data structures submitted for the pilot, the CDISC SDS Team overwhelmingly agreed that there were more commonalities than differences in the data structures submitted by the different companies. However, the team also concluded that the Version 3 concepts require additional documentation and training to minimize possible different interpretations as well as to highlight the model subtleties. Thus the principal conclusion from the pilot was the need for a much more detailed and complete implementation guide. The CDISC SDS Team presented these lessons at an FDA Public Meeting on October 2, 2003. The three main objectives of the pilot were to stress the conceptual models by including as many different types of therapeutic areas and data as possible, to test new model components (e.g., Trial Design), and to summarize sponsor and FDA experiences. A broad variety of data was submitted from eight different therapeutic classes and four different trial designs. The therapeutic classes included Oncology, Metabolic, Cardiovascular, Anti-infective, Hormone Replacement Therapy, Central Nervous System, Respiratory, and Endocrine. The study designs included 1) Double-blind, parallel group, 2) Double-blind followed by open label, 3) Multiple treatment phases/arms, and 4) Infusion and oral dosing. All of the predefined Version 3.0 CDISC SDS domain models were tested as well as more than twelve new domains (e.g., Microbiology, Questionnaire, Study Drug Compliance, Mammogram, Tumor Identification and Measurement, Disease Response, Health Care Utilization). Although most of the participants used their legacy database, one participant tested the conversion of Version 2.0 datsets to Version 3.0. The testing of new model components included the inclusion of new variables, concepts and datasets. An analysis of the variables used in the pilot was shared with the CDISC SDS Team. This analysis highlighted the variables actually used in the submission datasets, the new Version 3.0 variables, plus those that were not described within the specifications. It was through this analysis that the team deliberated over the cost/benefits of the inclusion of new variables or changing existing ones. Furthermore, the team discovered that new variables such as --GRPID, --SEQ, --SPID, --CAT, and --SCAT were beneficial but sponsors would need additional guidance to ensure consistent application. One especially significant conclusion was the reexamination of the date-time precision variable combination. The team concluded that the ISO 8601 date/time specification was more efficient and complete and leveraged a preexisting industry standard compared to the Version 3.0 variable pairing. Therefore, after much debate the team endorsed and included the ISO 8601 date format into version 3.1. Additionally, new data domains and concepts were tested, for example the trial design datasets. The findings and feedback from sponsors as well as from FDA increased the importance of including these concepts in the Version 3.1 release. The last major objective was to capture both sponsor participant and FDA experiences with the pilot. On the sponsor side, the team looked into the ease and consistency of use and categorization of data into the observation classes (Findings, Events and Interventions) as well as the identification of challenging areas. On the FDA side, the team was interested in reviewer experiences with the data and applications used in the pilot. A number of challenges were highlighted, such as re-conceptualizing how the electronic submission data is organized. For example, data that was traditionally collected within one CRF module may now go into multiple V3 domains and data that was collected on multiple CRF modules may go into one V3 domain. The second challenge faced by the participants was the physical mapping between horizontal data structures to a vertical or more normalized paradigm. Like the industry at large, pilot participants needed to be educated on new terms (e.g., epochs, elements and arms) that are used to replace common terms (e.g., periods, phases, and treatment groups) and concepts. Lastly, it was not always clear cut where data should be placed if the data did not easily fit into a specific observation class; this was in part due to the fact that the Version 3 specifications were still in a state of flux during the pilot period. In conclusion, the number-one learning from this pilot was that additional guidance and specifications were needed to reduce inconsistencies and enhance comprehension of the models. Specifically, a detailed implementation guide was necessary to more clearly communicate the specifications and the rules, as well as to provide additional guidance through examples is necessary. Also, the team learned that the vertical nature of the Findings datasets highlights the importance of and the need for specific controlled terminology to be able to provide record level metadata (e.g., via define.xml). Implementing Version 3 for the safety domains based on Version 2 was
CDISC, 2005. All rights reserved FINAL Page 173 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1) straightforward. Lastly, complex data (e.g., Microbiology, where multiple relationships (one to many to many) exist within and between variables and records) and study designs (e.g., Oncology) could be represented in Version 3. The SDTM is not meant to mirror data as collected on CRFs; that is, while CRFs will continue to be optimized for data collection, the SDTM models are optimized for electronic data exchange with the FDA. The V3 structures seemed to handle all of the many variations in data provided from different sponsors, and seemed to provide a very effective foundation for a common data standard.

Page 174 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

10.6

REVISION HISTORY

As stated above, this document, when approved, together with the SDTM, represents the most recent version of the CDISC Submission Data Domain Models, previously known as Version 3.1. The V3.x Study Data Tabulation Model Implementation Guide for Human Clinical Trials was the first implementation-ready version of the CDISC Submission Data Domain Models.

10.6.1

Changes from CDISC Version 2 Models to Version 3

V3 represented a major change from the CDISC Version 2.0 (V2) domain models for two primary reasons. First, it included a general model for representing all types of tabulation data (the general model has now been published separately as the SDTM). Second, the general vertical record structures used in V3 were more normalized than the more horizontal V2 record structures. Standard horizontal listings of data, such as those described in the V2 horizontal representations of ECGs and Vitals by visit, will be produced by FDA standard review tools. V3 was initially released for comment in March 2003, and finalized as an approved HL7 informative document that addressed all prior comments on June 9, 2003. V3 was then tested in Summer 2003 by individuals from nine sponsor companies who participated in an FDA pilot project. The results of the pilot (which are discussed in Section 10.5) were shared with industry at an FDA public meeting held on October 2, 2003, and feedback from the pilot was a primary input to V3.x. Another key input was a list of comments that had to be deferred for the June 9, 2003 publication, but which were addressed in V3.1 and later versions.

10.6.2
1) 2) 3) 4) 5)

Changes from CDISC V3 to V3.1

The general underlying conceptual model described in V3 as the General Study Data Information Model, was published separately as the SDTM, a separate document from this implementation guide. Corrections and amendments applied in the SDTM were also applied throughout the domain models, assumptions, and examples provided in this document. A new Trial Design component was incorporated with examples to provide a standardized representation of study timing and treatment plans, and a way of representing actual subject visits. A more thorough solution was included for defining relationships between datasets, domains, and/or records. The representation of all date/time variables was changed to ISO 8601 character format. The concept of a separate Date/Time Precision variable for each date/time variable was eliminated in this version, because that purpose is met by the ISO 8601 standard. New domain variables were added to represent additional timing descriptions, flags, and descriptive attributes of an observation (e.g., --SCAT, --DOSRGM, --NRIND). Some variables were removed from the domain models (e.g., --INTP, --DESC, --BLRESC, --BLRESN), either because they were deprecated in the prior version or were inconsistent with the intent of the model. The core variable concept was expanded to categorize variables as required, expected, or permissible. The Qualifier variable role was sub-categorized into five more granular subcategories (See Section 2.1) to provide more detail on the use of variables. The SU, DS, and PE domains were significantly redesigned. Additional business rules were incorporated to address common data management situations, such as representing tests not done. Numerous changes to V3 variables, labels, formats, and notes were made to reduce ambiguity and improve consistency. New sections were included in the Implementation Guide to provide additional descriptive examples and additional guidance on proper usage.
Page 175 August 26, 2005

6) 7) 8) 9) 10) 11) 12) 13)

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

10.6.3
Major Changes Change

Changes from CDISC SDSIG V3.1 to SDTMIG V3.1.1 Draft

General

Change Add Add Add Add Add Add Change Add Minor Changes Correct Correct Correct Correct

2.5 4.1.4.8 Separate Document Separate Document Separate Document Separate Document 7.10 8 8.4.3

Changed abbreviation from SDSIG to SDTMIG. Updated references to SDTM v1.1. Introduced abbreviation of V3.x to refer to V3.1 IG and subsequent versions. Any valid variable included in an SDTM general class may now be used in a domain based on that class. New assumption on the use of date/time variables for findings collected over an interval. New DV Protocol Deviations domain model and example New DA Drug Accountability domain and example New Pharmacokinetics Concentrations domain model and example New Pharmacokinetics Parameters domain model and example New TS Trial Summary domain model and examples. Also added TS code to Appendix 10.3.1. Substantial revisions throughout this Section to clarify proper usage. Inserted new option for creating separate SUPPQUAL datasets (supp--) per domain.

1.3 2.1 2.5 2.6

Correct Correct

2.6.1 3.2

Correct Add Correct Correct Correct Correct Correct

3.2.1 4.1.1 4.1.2.2 4.1.2.4 4.1.3.4 4.1.3.5 4.1.4.1

Moved list of changes to appendix 10.6. Revised text to accommodate v. 3.11. Corrected spelling of --ORNRLO and --ORNRHI. Update list of domains with DV, DA, MB, PC, PP, TS Clarified step 11 to emphasize that labels should only be changed in new domains. Added statement about posting suggested new codes to CDISC web site. Corrected step number references in last 2 paragraphs. Updated documentation to clarify that the key order in metadata should reflect the natural sort order of the domain when viewed, since all physical keys are fixed within the SDTM (based on STUDYID, USUBJID, --SEQ). Clarified that roles are optional for standard domains. Changed dataset names to lowercase for consistency with FDA eCTD guidance requirements. Corrected Structure column for consistency with domain models. Added reference to CDISC Define.xml specification. Remove reference to RELDS and COMMENTS. Clarified when the use of mixed case text is allowable. Clarified that analysis variables may be included as supplemental qualifiers. Removed references to PE Corrected format of fourth bullet example by adding hyphen separators to year. Fixed link
CDISC. 2005. All rights reserved FINAL

Page 176 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1) Minor Changes Correct

4.1.4.2 4.1.4.3

Correct Correct

4.1.4.5 4.1.4.7

Correct Correct Correct Correct Correct/Add

4.1.5.1 4.1.5.2 4.1.5.4 5.1.1 5.1.1.1

Correct Correct Correct

5.1.2 5.1.2.1 6.1.1

Add/Remove Correct

6.1.1.1 6.1.2

Correct

6.1.3

Correct

6.2.1

Correct Correct Correct

6.2.1.1 6.2.2 6.2.3

Correct

6.2.3.1

Reworded first sentence. Updated footnote to use correct expression syntax (input(DTC, datetime18) Corrected examples and expanded documentation on use of ISO8601 for durations Updated to allow use in other domains. Updated diagram and description to explain how to address timepoints that either start before the study with an unknown end or start during a study with an unknown end. Clarified use of --ORRES in PE; minor text corrections to clarify. Used correct CT for SYSBP. Clarified that long text should be split for readability when > 200 characters. Clarified use of EVAL and that the name of the attribution should increment by 1 for each additional attribution on the same topic Specified use of ISO 3166 3-character country code. Added example of USUBJID convention. Clarified assumption 1 on necessity of including SITEID. Clarified assumption 7 on how to represent additional race information. Added assumption 8 on screen failures. Corrected reference to COVAL in paragraph 1. Corrected CDISC notes for DOMAIN, RDOMAIN, COREF and COVAL. Added reference to 8.5. Corrected references to IDVAR in assumption 1; correct example. Correct label, role and notes for CMSPID for consistency with other domain models. Corrected label for CMMODIFY for consistency with SDTM. Corrected role for CMCLASCD, CMDOSU, CMDOSFRM, CMDOSFRQ, CMROUTE to Variable Qualifier. Edited notes and added reference and new term to CMSTREF and CMENREF. Corrected CDISC Notes for CMOCCUR. Added new assumption 5 to discuss record structure. Removed old assumption 5 (which was redundant with Notes and assumption 4.1.4.7. Corrected role of EXSPID for consistency with SDTM and other domain models. Correct role for EXDOSU, EXDOSFRM, EXDOSFRQ, EXROUTE to Variable Qualifier. Updated CDISC notes for EXSTDTC and EXENDTC. Corrected role of SUSPID for consistency with SDTM and other domain models. Corrected role for SUCLASCD, SUDOSU, SUDOSFRM, SUDOSFRQ, SUROUTE to Variable Qualifier. Correct note to SUSTAT. Corrected Origin for SUROUTE for consistency. Corrected labels for AESTDY, AEENDY, AEDUR for consistency with SDTM. Corrected notes for AETOXGR and AEDUR. Corrected note and CT for AEENRF. Corrected text in 5a to allow use of both AEENDTC and AEENRF.. Corrected notes for DSREFID, DSSPID, DSSTDTC and DSSTDY. Corrected role for EPOCH to Timing. Corrected Label for DSTERM and DSDECOD Corrected labels for MHSEQ, MHTERM, MHCAT, MHSCAT, MHDTC for consistency. Corrected notes to MHTERM, MHDECOD, MHSCAT, MHDTC, MHENDTC. Changed Core to Permissible for MHSTDTC. Corrected note to MHENRF to allow use with MHENDTC. Updated Note 4d for clarity.

CDISC, 2005. All rights reserved FINAL

Page 177 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1) Minor Changes Correct

6.3.1

Correct Correct

6.3.1.1 6.3.2

Correct Correct

6.3.2.1 6.3.3

Add Correct

6.3.3.1 6.3.4

Correct Add Correct Add Correct

6.3.5 6.3.5.1 6.3.6 6.3.6.1 6.3.7

Add Correct

7.1 7.2.1

Correct

7.2.2

Correct

7.2.3

Correct Correct Correct Replace

7.3.1 7.3.2 7.4.2 7.4.4

Correct
Page 178 August 26, 2005

7.5

Added variables EGNRIND, EGMETHOD, EGTPTNUM to model. Corrected notes for EGGRPID, EGNRIND, EGSCAT, EGLOINC, EGEVAL. Corrected references for EGORRESU, EGSTRESN, EGSTRESU, EGDRVFL, EGELTM. Changed Core for EGPOS to Permissible. Added example in note to EGELTM. Corrected Assumption 1. Corrected Type of IESPID to Char for consistency. Corrected Core for IEORRES and IESTRESC to Required. Corrected Core for IEDTC to Permissible. Updated reference for IEDY. Corrected note to IETEST. Added new assumption 3 to clarify relationship to TI domain. Added variables LBSTRESU, LBNAM, and VISITNUM (all omitted in error). Corrected references for LBORRESU, LBREASND, LBDRVFL, LBELTM. Changed Role of LBNRIND to Expected. Corrected Note to LBNRIND. Changed labels and notes to use Reference Range instead of Normal Range. Added example to note for LBELTM. Added new assumption on use of reference range fields. Added PEORRESU, PESTRESN, PESTRESU, PEBLFL, PEEVAL, PESEV to domain model. Corrected note for PETEST. Corrected type of PESPID to Char. Changed ole of VISITNUM and PEDTC to Expected. Corrected reference for PEREASND. Moved PESPID to top of domain for consistency. Corrected note for QSCAT, Changed Core for VISITNUM to Expected, Corrected reference for QSDRVFL, QSELTM. Added new assumptions. Added new assumptions 1-3 to better explain use of QS. Added variables SCSTAT, SCREASND, SCDY for consistency. Corrected Role of SCSTRESC, SCSTRESN, SCSTAT, SCREASND to Result Qualifier. Added new assumption on use of SC vs. DM and SUPPQUAL. Added VSNRIND variable. Updated Core of VSSPID, VSPOS and VSNRIND to Permissible, VSTEST to Required, VSSTRESC and VSSTRESN to Result Qualifer, VSNRIND, VSSTAT and VSLOC to Permissible, VSBLFL and VSDTC to Expected. Corrected references for VSCAT, VSORRES, VSORRESU, VSSTRESC, VSSTAT, VSLOINC, VSBLFL, VSDRVFL, VSELTM. Corrected note for VSNRIND and VSDRVFL.. Added references to Epochs. Corrected structure in header. Corrected notes for ETCD, ELEMENT and TEENRL. Changed role of ELEMENT to Synonym Qualifier. Changed Core of TEENRL to Perm. Corrected structure in header. Corrected labels for ARM and ARMCD (removed "Treatment Group"). Changed EPOCH Role to Timing. Corrected notes for ARMCD and ELEMENT.. Corrected structure in header. Corrected TVENRL Origin to Sponsor Defined/ Protocol. Corrected label for ARMCD and VISITNUM. Corrected note for ARMCD and ARM. Changed Core of VISIT to Perm. Corrected DOMAIN code to SE. Corrected structure in header. Corrected note to ELEMENT. Corrected structure in header. Removed "Treatment" from note to SVENDTC Corrected ELEMENT values in table to match figure. Replaced second example with corrected version that adds EPOCH. Added clarification to last paragraph. Corrected table values for ELEMENT for consistency with figures. Corrected number of records in figure to 20.
CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) Minor Changes Correct Correct Correct Correct Correct Correct Correct Correct Add Correct Add

7.7.5 7.9 8.0 8.1.1 8.2 8.2.2 8.3 8.3.1 8.4 8.4.1 8.4.2

Correct Correct Add Add Correct Remove Replace Replace Replace Replace Add Replace Replace Replace Add Remove

8.5 8.5.1 8.6 8.6.1 9.1 9.3.3 9.4.1 9.4.2 9.4.3 9.4.4 9.4.5 9.4.6 9.4.7 9.5.1 9.5.2 10.3.1

Add Correct Correct Correct Correct

10.3.5 10.3.3 10.3.4 10.4 10.5

Clarified wording of second paragraph. Corrected Roles of IETEST to Synonym Qualifier and IECAT to Grouping Qualifier. Corrected notes for IETESTCD, IETEST, IECAT. Removed references to lesions in bullet 3 and made minor text edits. Clarified text in paragraph 1. Substantial text corrections to entire section for clarification. Corrected USUBJID reference in Example A Inserted clarifying test in paragraph 2. Replaced example of MB for split domains. Substantially revised text for clarification. Added new paragraphs 3 and 4 to clarify use of QNAM and QVAL. Substantially revised text for clarification. Substantial text revisions for clarification. Corrected notes for QNAM and QLABEL. Added new last paragraph. Substantially revised text for clarification. Added new paragraph at end to emphasize that null values for QVAL are not to be used. Adjusted values in example; changed PPE to PPROT. Added new foreign language example. Substantially revised text for clarification. Corrected COREF values in example. Substantially revised text for clarification. Added new paragraphs 3-5 and revised other sentences to clarify principle of domain topicality. Inserted new Stress Test multiple-dataset example. Removed last row in example Deleted statement beginning with "If terms are specified in the CRF" below note e. Updated EG Example with minor corrections. Updated IE Example with minor corrections. Updated LB Example; added second example to show a urine collection, including a negative reference period. Updated PE Example with PESEV and minor corrections Corrected QS example. Added explanations. Updated SC Example with minor corrections. Updated VS Example with minor corrections. Replaced Trial Design examples with new examples to better communicate core concepts; added new TS example. Added new TS example. Removed reserved codes for DR, DY, NE, PF, RF, SZ, VX as these may no longer be necessary as separate domains. Changed code for Biopsy from BX to BR. Added controlled terms for new TS domain. Added new code for BODYFAT Added new code for SCRNFAIL. Corrected several fragment names. Revised last paragraph. Other minor text revisions.

CDISC, 2005. All rights reserved FINAL

Page 179 August 26, 2005

CDISC SDTM Implementation Guide (Version 3.1.1)

10.6.4
Major Changes Add Change

Changes from CDISC SDTMIG V3.1.1 Draft to V3.1.1 Final

2.4 4.1.1.5

Add Change

4.1.4.1 4.1.4.2 4.1.4.7

Included additional information about choosing a general class and ability to use any qualifier from same general class. Corrected last bullet to allow use of any qualifier from same general class. Also clarified that a null column is acceptable for expected variables not collected. Reworded for improved clarity. Added examples of new ISO 8601-compliant representation for partial date/times with omitted components. Removed restriction that --STRF and --ENRF can only be used if start and/or end dates are missing. Also removed restrictions from domain models using those variables (CM, SU, AE, MH). New assumption on representing date results in findings. Substantially reworded assumption for clarity. Corrected EGTESTCD for RHYRATE in ECG example. Changed the header for all domain models to use lowercase xpt filenames for consistency with eCTD specifications. Revise notes and Core for RFSTDTC, RFENDTC and revise Assumption 8 to allow for representation of screen failures with ARMCD=SCRNFAIL rather than using SUPPQUAL. Changed label for all domain instances of --SPID from Sponsor ID to Sponsor-Defined Identifier so as not to be confused with the ID of a sponsor company. Reclassified --STAT as a Record Qualifier. Changed domain models for CM, SU, AE, MH to address possible use of new SuppQual variable --PRESP, reconsider use of --OCCUR, and emphasize importance of including data that represents actual interventions or events. Revised QS Assumption 1 to clarify domain contents. Added new variable SESEQ to uniquely identify each row and set the sequence of elements followed by a subject. The use of a single SUPPQUAL dataset is not recommended for users of V3.1.1 and later versions. The use of separate datasets has been found to be much more convenient for sponsors and is necessary for compatibility with the CDISC ODM model and define.xml. Changes to text on determining where to put data in the SDTM and removal of example, which is under discussion among the SDS Team. Removed example 1 from prior SU example, since proper use of the SUOCCUR variable is being reconsidered by the SDS Team; Revised example 2 (now1). Modified example to remove references to MHOCCUR variable, since use of this variable is being reconsidered by the SDS Team. Added description for each domain code; adjusted list of reserved codes.

Add Correct Change Change

4.1.4.9 4.1.5.1 Sections 5 &6 5.1

Change

Section 6

Change

Section 6

Change Add Change

6.3.5.1 7.3.1 8.4

Change Change Change Add

8.6 9.2.3 9.3.3 10.3.1

Add

10.7

New section 10.7 on Warranties, Limitation of Liability, Disclaimers.

Page 180 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) Minor Changes Add Add Correct Add Correct Correct Correct Correct Correct Correct Correct

1.1 1.3 2.1 2.2 6.1.1 6.1.3 6.2.1 6.2.1.1 6.2.2 6.2.2.1 6.2.3

Correct Add Correct Correct/Add Correct Correct Correct Correct Correct Correct Correct Correct Correct Correct Correct Correct Correct Correct Correct

6.2.3.1 6.3.1 6.3.2 6.3.3 6.3.3.1 6.3.4 6.3.5.1 6.3.6 6.3.6.1 7.2.1 7.2.2 7.3.2 7.4.4 7.5.2 7.9 7.10 8.3.1 8.4.2 8.6.1

Expanded definition of tabulation datasets. Included reference for FDA study data specifications. Modified bullet 6 to reference new appendix 10.6. Removed --SPEC, --LOT and --NAM as examples of Grouping Qualifiers, moved to Record Qualifiers. Referred to use of Result Qualifiers in DM. Additional explanatory text in paragraph 1. Corrected label for CMREASND to conform with 40 character limit. Added SUSTRF and SUENRF; modified label of SUTRT. Reclassified former Result Qualifiers as Record Qualifiers. Corrected examples in CDISC Note for AEREL. Added notes and ** to AETOXGR and reclassified former Result Qualifiers as Record Qualifiers. Reworded assumption 1 to clarify definition. Revised Assumption 4 to reduce ambiguity and discuss possible future use of AEPRESP and AEOCCUR. Corrected example list of values for DSDECOD. Reworded assumption 1 to clarify definition. Changed Core to Perm for MHBODSYS and MHDECOD. Corrected Origin for MHBODSYS to Derived. Deleted 2nd note for MHBODSYS reclassified former Result Qualifiers as Record Qualifiers. Reworded assumption 1 to clarify definition. Reworded assumption 3a to clarify use of MHCAT and MHSCAT. Added note to EGLOINC; corrected label for ESPID, corrected note to EGORES. Corrected labels to IETESTCD, IETEST and IECAT for consistency with SDTM TI labels. Corrected labels and notes for IEORES and IESTRESC. Corrected examples in CDISC NOTE for LBMETHOD. Added LBENDTC to model. Corrected wording of assumption 1 for LB. Corrected header to PE table. Removed "QSORRES" from assumption 4. Moved SCSPID after SCGRPID for consistency with other domains. Removed last 2 sentences from Assumption 2, which described an example that belonged in suppdm rather than suppsc. Added ISO 8601 to Format for TEDUR; Corrected note for ETCD (ETCD note also corrected in other trial design domains) Corrected note for ARMCD (ARMCD note also corrected in other trial design domains) Corrected note to VISITDY (to remove reference to use of decimals). Changed examples for TEDUR to ISO8601 format. Changed examples for TEDUR to ISO8601 format. Corrected labels to IETESTCD, IETEST and IECAT for consistency with SDTM TI labels. Corrected Notes to TSPARM and TSVAL; corrected core for TSVAL Updated example to use MB and MS microbiology domains. Clarified wording of bullet 2. Removed incorrect example; a better example will be provided in the next version of the SDTMIG.
Page 181 August 26, 2005

CDISC, 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1) Minor Changes Add Correct Correct Correct

9.3.3 9.4.1 9.4.4 9.4.6

Correct Correct Add Delete Correct Correct

9.4.7 9.5.1 10.3.1 10.3.2 10.3.3 10.3.4

Correct Correct

10.3.5.1 10.4

Revised example, eliminating MHOCCUR example and adding new note and new column to illustrate use of MHENRF. Corrected values of EGEVAL in Example 1. Corrected values of USUBJID in PE Example. Removed SCDY from example (not likely to be used in SC). Changed SCTESTCD values from "INITIALS" to "SUBJINIT" for consistency with domain examples. Corrected explanation of example and row references. Changed examples for TADUR and TEDUR to ISO8601 format Added SI Infection Site Measurements to domain codes. Removed FINDING from list of controlled terms for EG. Minor changes to VSTEST for TEMP and RESP Changed AGEGRP to AGESPAN, AGEGRP, CONTROL, DRUGDICT, OBJPRIM, OBJSEC; correction to code "3-4" in PHASE; added new values to TYPE. Changed MASK to BLIND; changed MEDDICT to MHDICT. Removed TSPARM = DESIGN. Corrected list of Trial Summary parameters; a more complete list of parameters will be defined separately by the CDISC Terminology team. Corrected list of fragments (Domain codes are no longer included in this fragment list, but appear only in 10.3.1).

Page 182 August 26, 2005

CDISC. 2005. All rights reserved FINAL

CDISC SDTM Implementation Guide (Version 3.1.1)

10.7

REPRESENTATIONS AND WARRANTIES; LIMITATIONS OF LIABILITY, AND DISCLAIMERS

CDISC Patent Disclaimers. It is possible that implementation of and compliance with this standard may require use of subject matter covered by patent rights. By publication of this standard, no position is taken with respect to the existence or validity of any claim or of any patent rights in connection therewith. CDISC, including the CDISC Board of Directors, shall not be responsible for identifying patent claims for which a license may be required in order to implement this standard or for conducting inquiries into the legal validity or scope of those patents or patent claims that are brought to its attention. Representations and Warranties. Each Participant in the development of this standard shall be deemed to represent, warrant, and covenant, at the time of a Contribution by such Participant (or by its Representative), that to the best of its knowledge and ability: (a) it holds or has the right to grant all relevant licenses to any of its Contributions in all jurisdictions or territories in which it holds relevant intellectual property rights; (b) there are no limits to the Participants ability to make the grants, acknowledgments, and agreements herein; and (c) the Contribution does not subject any Contribution, Draft Standard, Final Standard, or implementations thereof, in whole or in part, to licensing obligations with additional restrictions or requirements inconsistent with those set forth in this Policy, or that would require any such Contribution, Final Standard, or implementation, in whole or in part, to be either: (i) disclosed or distributed in source code form; (ii) licensed for the purpose of making derivative works (other than as set forth in Section 4.2 of the CDISC Intellectual Property Policy (the Policy)); or (iii) distributed at no charge, except as set forth in Sections 3, 5.1, and 4.2 of the Policy. If a Participant has knowledge that a Contribution made by any Participant or any other party may subject any Contribution, Draft Standard, Final Standard, or implementation, in whole or in part, to one or more of the licensing obligations listed in Section 9.3, such Participant shall give prompt notice of the same to the CDISC President who shall promptly notify all Participants. No Other Warranties/Disclaimers. ALL PARTICIPANTS ACKNOWLEDGE THAT, EXCEPT AS PROVIDED UNDER SECTION 9.3 OF THE CDISC INTELLECTUAL PROPERTY POLICY, ALL DRAFT STANDARDS AND FINAL STANDARDS, AND ALL CONTRIBUTIONS TO FINAL STANDARDS AND DRAFT STANDARDS, ARE PROVIDED AS IS WITH NO WARRANTIES WHATSOEVER, WHETHER EXPRESS, IMPLIED, STATUTORY, OR OTHERWISE, AND THE PARTICIPANTS, REPRESENTATIVES, THE CDISC PRESIDENT, THE CDISC BOARD OF DIRECTORS, AND CDISC EXPRESSLY DISCLAIM ANY WARRANTY OF MERCHANTABILITY, NONINFRINGEMENT, FITNESS FOR ANY PARTICULAR OR INTENDED PURPOSE, OR ANY OTHER WARRANTY OTHERWISE ARISING OUT OF ANY PROPOSAL, FINAL STANDARDS OR DRAFT STANDARDS, OR CONTRIBUTION. Limitation of Liability. IN NO EVENT WILL CDISC OR ANY OF ITS CONSTITUENT PARTS (INCLUDING, BUT NOT LIMITED TO, THE CDISC BOARD OF DIRECTORS, THE CDISC PRESIDENT, CDISC STAFF, AND CDISC MEMBERS) BE LIABLE TO ANY OTHER PERSON OR ENTITY FOR ANY LOSS OF PROFITS, LOSS OF USE, DIRECT, INDIRECT, INCIDENTAL, CONSEQUENTIAL, OR SPECIAL DAMAGES, WHETHER UNDER CONTRACT, TORT, WARRANTY, OR OTHERWISE, ARISING IN ANY WAY OUT OF THIS POLICY OR ANY RELATED AGREEMENT, WHETHER OR NOT SUCH PARTY HAD ADVANCE NOTICE OF THE POSSIBILITY OF SUCH DAMAGES. Note: The CDISC Intellectual Property Policy can be found at http://www.cdisc.org/about/bylaws_pdfs/CDISCIPPolicy-FINAL.pdf .

CDISC, 2005. All rights reserved FINAL

Page 183 August 26, 2005