Electronic Health Records and 

Clinical Data Interchange Standards

Stephen E. Wilson, Dr.P.H., CAPT USPHS
Director
Division of Biostatistics III
Center for Drug Evaluation and Research
US Food and Drug Administration
IPPCR – February 23, 2016
 Disclaimer
This presentation reflects the views of the 
author and should not be construed to 
represent FDA’s views or policies.
 Acknowledgements
• Becky Kush, CDISC
• Chuck Cooper, (formerly FDA/CDER)
• Ron Fitzmartin, FDA/CDER
• Shannon Labout, CDISC
• Lilliam Rosario, FDA/OTS
 Science, Statistics and 
Experimental Design
Science is concerned with understanding 
variability in nature
Statistics is concerned with making decisions
about nature in the presence of variability
Experimental design is concerned with 
reducing and controlling variability in ways 
which make statistical theory applicable to 
decisions about nature.
Winer, et.al., Statistical Principles in Experimental Design, 
New York, 1962, 1971, 1991
 Data Transparency
• Patients and their physicians depend on 
clinical trials for reliable evidence on 
what therapies are effective and safe. 
• Responsible sharing of the data gleaned 
from clinical trials will increase the 
validity and extent of this evidence.

Institute of Medicine, Sharing Clinical Trial Data – Maximizing Benefits, Minimizing Risk, 2015 
 My Assumptions – What You Have  
Already Learned in this IPPCR Course
• Good Clinical Research Practice
– Design
– Conduct
– Statistical Analysis 
– Legal and Ethical Principles
• Data Management
• Drug Development
• FDA Product Regulation
 Steve’s Black Box Warning
• FDA/CDER‐Centric  ‐‐ There are regulation/guidance 
differences between  Centers at FDA
• FDA  regulation /guidance codifies Good Research 
Practice 
• Primary focus: Confirmatory (Phase III) Trials
• A CDER statistical reviewer perspective
• If you feel that  your clinical research will contribute 
to the development of marketed medical products 
… please pay  special attention
• Good clinical research practice is a global concern
 Outline
• Substantial Evidence and FDA/CDER  Review of 
Confirmatory  (Phase 3) Clinical Trials
• CDISC Data Standards
– New CDER/CBER Requirements
– CDISC Basics
• Using eHealth Records in Clinical Research for 
Drug Development
 FDA Mission
• FDA is responsible for protecting the public health by 
assuring the safety, efficacy and security of human 
and veterinary drugs, biological products, medical 
devices, our nation’s food supply, cosmetics, and 
products that emit radiation.
• FDA is responsible for advancing the public health by 
helping to speed innovations that make medicines 
more effective, safer, and more affordable and by 
helping the public get the accurate, science‐based 
information they need to use medicines and foods to 
maintain and improve their health.
 FDA Mission
• FDA is responsible for protecting the public health by 
assuring the safety, efficacy and security of human 
and veterinary drugs, biological products, medical 
devices, our nation’s food supply, cosmetics, and 
products that emit radiation.
• FDA is responsible for advancing the public health by 
helping to speed innovations that make medicines 
more effective, safer, and more affordable and by 
helping the public get the accurate, science‐based 
information they need to use medicines and foods to 
maintain and improve their health.
 FDA
Food and Drug Administration

11
 NDA/BLA Review at CDER
(Center for Drug Evaluation and Research)

12
 CDER
• Almost all clinical research associated with New 
Drug Applications (NDAs) and Biologics License 
Applications (BLAs) are reviewed by CDER
• Within CDER ‐‐ Primary offices for planning and 
review of clinical trials
– Office of New Drugs (OND)
– Office of Biostatistics (OB)
Office of Biostatistics (OB)
Statistical Review Staff
We Worry About Bias That Will 
Affect Our Decisions –
Due Decision Diligence (DDD) 
Bias is a  preconceived personal 
preference or inclination that 
influences the way in which a 
– measurement
– analysis
– assessment 
– or procedure 
…is performed or reported
Always a Risk‐Benefit Decision

BENEFIT RISK
 We Worry About Bias

BENEFIT RISK
 We Worry About Bias

BENEFIT

RISK
 Substantial Evidence / Adequate 
and Well‐Controlled Studies
• The Food and Drug Administration considers 
these characteristics in determining whether an 
investigation is adequate and well‐controlled 
for purposes of section 505 of the act. 
• Reports of adequate and well‐controlled 
investigations provide the primary basis for 
determining whether there is "substantial 
evidence" to support the claims of effectiveness 
for new drugs.
21 CFR Part 314 Sec 314.126
ICH E9 – Confirmatory Trials
FDA’s “Gold Standard” for Approval

• NDA/BLA
• Two adequate and well‐controlled, 
confirmatory clinical trials (AWCCCT)
• Pre‐specified endpoint(s), sample size & 
analysis
• Two‐sided, p‐value < 0.05
• Clinical benefit
21st Century Review of NDAs/BLAs
Answering Some Review Questions
• Assess compliance with protocol / blinded analysis plans
• Assess traceability and quality of submitted data
• Verify results reported in the NDA.
• Check appropriateness of statistical models and conclusions.
• Assist DSI in planning investigator audits
• Modify models and assess robustness/sensitivity of the results.
• Evaluate Patient entry errors (cancelled patients, ineligible patients)
• Determine impact of “non-evaluable” patients
• Assess extent and potential impact of missing data
• Check consistency of results within and across studies
• Examine the trial and data for bias:
– Results by center
– Treatment assessment
– Baseline predictors
– Important subgroups (sex, age, race, etc,)
• Assess impact of audits
The Call from a CDER Reviewer

SHOW ME 
THE DATA!
Review Tools: Patient Profiles

Korvick & Szarfman, 2000

 Review Tools:
OCS/CSC JumpStart Service
• Provides a recommended sequence for using the outputs 
• Allows reviewer to follow a safety signal from a high‐level to the 
specific patient details with complementary tools

MAED: System Organ  MedDRA at a
Class Glance Report
Identifies a difference  Shows same signal
between treatment  across multiple
arms for both risk  levels of the
difference and relative  hierarchy for the
risk. treatment arm.

JReview: Risk JReview: Graphical

Assessment Patient Profile
Magnifies the safety Shows which
signal when viewing patients experienced
patients that were the Adverse Event
not treated with the shortly after taking a
study drug. specific concomitant
medication.
Rosario, 2014
 OCS/CSC JumpStart Service
Starts a review by performing many standard analyses and 
identifying key information

Rosario, 2014
 Required Submission of CDISC
Standardized Data Standards
Published Final  Published Final 
Dec 17, 2014 May 5, 2015

http://www.fda.gov/forindustry/datastandards/studydatastandards/default.htm
Fitzmartin, 2015
28
 Required Submission of CDISC
Standardized Data Standards
Published Final  Published Final 
Dec 17, 2014 May 5, 2015

24 months after
final guidance ,
sponsors must use
standards
identified by FDA
(NDAs, ANDAs,
BLAs).

http://www.fda.gov/forindustry/datastandards/studydatastandards/default.htm
Fitzmartin, 2015
29
The New Call from a CDER 
Reviewer

SHOW ME THE 
HIGH QUALITY, 
CDISC‐
STANDARDIZED 
DATA!
 CDISC
Clinical Data Interchange Standards Consortium

• What is CDISC?
• Submission Data Standards to CDER
• Developing Ebola TAs
 What is CDISC?
www.cdisc.org
• Global Standards Development
Organization (SDO)
• Founded in 1997 (all volunteers)
• Incorporated in 2000 as a non-profit
organization
• Over 300 member organizations
• 90 countries; Coordinating Committees in
Europe, Japan, China, Asia-Pacific; 20
user networks
 CDISC Vision and Mission
• Vision: informing patient care and safety
through higher quality medical research.
• Mission: The CDISC mission is to develop
and support global, platform-independent
data standards that enable information
system interoperability to improve medical
research and related areas of healthcare.

http://www.cdisc.org/CDISC-Vision-and-Mission
 CDISC Standards …
• Streamline research processes and
enable data sharing/aggregation
• Support all types of research from protocol
through analysis and reporting
• Include link to healthcare through EHRs
 Do you need Data Standards?

Do you have  Do you do  Do you annotate, 

limited time to  protocol‐based  acquire, aggregate, 
complete  clinical research? analyze, archive 
research  data?
programs?

Do you have 
limited 
Do you want  resources? Do you want to 
high quality  save time and 
data? costs?
 Do you need Data Standards?

Do you have  Do you do  Do you annotate, 

limited time to  protocol‐based  acquire, aggregate, 
complete  If you said Yes!
clinical research? analyze, archive 
research  data?
programs? to any of these –
you need
Do you have 
standards!
limited 
Do you want  resources? Do you want to 
high quality  save time and 
data? costs?
 Do you need CDISC?

Do you need to be 
Do you intend  Do you want to 
transparent and 
to or have you  retrieve data 
compliant?
acquired  from EHRs?
another 
company or  Do you track and 
organization? Do you work 
with partners  report safety data?
(CROs, 
technology 
providers,  Do you need 
Do you submit  patients and 
research 
data to the US  investigators?
partners)?
FDA or to PMDA?
 Do you need CDISC?
Do you need to be 
Do you intend to  Do you want to  transparent and 
or have you  retrieve data  compliant?
acquired  If you said Yes!
from EHRs?
another 
company or  to any of theseDo you track and 
–
organization? Do you work 
youwith partners 
need global report safety data?

community-
(CROs, 
technology 

Do you submit 
wide standards!
providers, 
research 
Do you need 
patients and 
data to the US  partners)? investigators?
FDA or to PMDA?
 CDISC Data Standards 
in the Clinical Research Process
Planning   Data Collection   Data Tabulations   Analysis  Reporting

Protocol Case Data Data Analysis Submission/ Review

Report Capture Mgmt Reporting /
Form Archiving
PRECLINICAL

CDISC Standards support the process from end to end

Adapted from Kush and Howard, CDISC, 2014 
Adapted from Kush and Howard, CDISC, 2014 
 CDISC – End to End 
(quality, speed, provenance)

Adapted from Kush and Howard, CDISC, 2014 
 CDISC Data Standards for 
Submission to CDER
• Study Data Tabulation Model (SDTM)
• Analysis Data Model (ADaM)
• Therapeutic Area (TA) Standards 
Study Data Tabulation Model (SDTM)
Protocol Case Data Data Analysis Submission/ Review
Report Capture Mgmt Reporting /
Form Archiving
PRECLINICAL

Collected/Observed
Data
SDTM

Adapted from Kush and Howard, CDISC, 2014 
 Basic Concepts of SDTM
One model – multiple implementations
• Pre‐clinical (SEND)
• Clinical (SDTMIG)
• Therapeutic areas / Devices / PGx

Model concepts
• Standard variable names
• Standard list of values (CT)
• Standard sets of data (domains) with standard names
• Standard data types, formats and other attributes
• Standard assumptions for implementation
• Standard way to submit “non‐standard” variables

Designed to hold anything you collect

Backward compatible, but evolving
Adapted from Kush and Howard, CDISC, 2014 
 Value of SDTM
Originally designed as a standard for data submissions to regulators 
to support
• Initial review of the data
• Warehousing of the data
• Data mining

Potential to improve the review / approval process

Supports data warehousing /data mining for organizations

Creates the opportunity of re‐usability for organizations

Facilitates collaborations, acquisitions, mergers

Adapted from Kush and Howard, CDISC, 2014 
 Without SDTM …
Name for Variability… Name for
demographics
Subject ID Study #2 – dmg.xpt
dataset varies
is never the ID GENDER
same A1 Male
Study #1 – demog.xpt A2 Male
A3 Female Study #4 – axd222.xpt
SUBJID SEX
A4 identifiers
Subject Female USUBID SEX
0001 M look different
A5 in Male 00011 0

0002 F
every study…not 00012 1
clear if the same 00013 1
0003 F person was in 00014 0
Study
more than#3one
– dmgph.xpt
0004 M 00015 1
study.PTID GENDER
0005 F 0001 1
0002 1
Gender or
0003 2
Is Sex Male or Sex, which
0004 2
Female, M or F, one will this
0005 1
1 or 2? study use?

Adapted from slide courtesy of Armando Oliva, M.D. and Amy Malla, FDA
 With SDTM
Column Header
(Variable) for No Variability! Name for
Subject ID is Study #2 – DM.xpt
always same demographics
USUBJID SEX dataset always
DEF-001 M the same!
Study #1 – DM.xpt DEF-002 M
ABC-001 M Study #4 – DM.xpt
USUBJID SEX
DEF-004 F USUBJID SEX
ABC-001 M DEF-005 M JKL-011 M
Consistent format and standard
ABC-002 F definition for USUBJID JKL-012
makes it F
GHI-003
clear which subjects were in F
ABC-003 F
more
Study #3 than one study. JKL-014
– DM.xpt M
ABC-004 M JKL-015 F
USUBJID SEX
ABC-005 F GHI-001 M
GHI-002 M
Sex is always
GHI-003 F
Sex always uses reported using
GHI-004 F
same terminology the same
GHI-005 M
(codelist) variable

Adapted from slide courtesy of Armando Oliva, M.D. and Amy Malla, FDA
 SDTM Logically Organizes Data
Trial Design  Subject Data Associated 
Data              Special Purpose  General Observation Class Domains Persons Data 
Domains (Standard Variables)       (Non‐Standard 
TE
Variables) Mimics Subject 
(Trial Elements) VS (Vital Signs)
Data structure, 
TA LB (Lab Results) SUPPLB but used for data 
(Trial Arms) SE (Subject about persons 
Elements) EG (ECG Results) SUPPEG who are not 
TV DM (Demographics**)
subject in the 
(Trial Visits) FA-- (Findings about…) SUPPFA-- trial, but whose 
SV (Subject data is useful for 
Visits) AE (Adverse Events) SUPPAE
TD the study, such as 
(Trial Disease MH (Medical History) parents, siblings, 
Assessments) device operator 
DS (Disposition**) who experiences 
TI CO an Adverse Event 
(Trial Inclusion/ (Comments) EX (Exposure**) SUPPEX with the device.
Exclusion
CM (Con Meds) SUPPCM
Criteria)
SU (Substance Use)
TS
(Trial Summary RELREC 
Information) (Related observations, stored in separate 
datasets. are identified and linked),

Gary Walker, Quintiles
 Analysis Data Model (ADaM)
Protocol Case Data Data Analysis Submission/ Review
Report Capture Mgmt Reporting /
Form Archiving
PRECLINICAL

Analysis Data

Adapted from Kush and Howard, CDISC, 2014 
 Basic Concepts of ADaM

Standard presentation of analysis data

• Supports the analyses in the CSR

“One PROC away from TLFs”

With the ADaM Define‐XML, provides 
traceability  back to SDTM
Adapted from Kush and Howard, CDISC, 2014 
 Value of ADaM
Describes consistent presentation of 
analysis data
• Flexible enough to support the analysis
• Standardized enough to support regulatory review 
needs
Provides part of traceability from analysis 
results back to CRFs 

Describes standardized approach to 
common analyses, e.g., AE, Time to Event
Adapted from Kush and Howard, CDISC, 2014 
 OCS/CSC JumpStart Service
Starts a review by performing many standard analyses and 
identifying key information

Rosario, 2014
 CDASH – Data Collection
Protocol Case Data Data Analysis Submission/ Review
Report Capture Mgmt Reporting /
Form Archiving
PRECLINICAL

CDASH

Adapted from Kush and Howard, CDISC, 2014 
Ebolavirus Disease Case Study - Global
Harmonization to Increase Power and
Accelerate Outcomes in Clinical Research
Data
Facilitator: Shannon Labout
Vice President Education @CDISC
CDISC
54
 A Case Study: A Randomized,
Controlled Trial To Evaluate
Ebola Therapies

Dionne L. Price, Ph.D.

Director, Division of Biometrics IV
OB/OTS/CDER/FDA

55
55
Setting the Stage and Understanding
the Challenges of Ebolavirus Disease
Clinical Research

Laura Merson, Clinical Trialist, Oxford

University
Consulting Expert to CDISC Ebolavirus
Therapeutic Area standards development team

56
 Leveraging CDISC standards to
support Ebola research
Maura Kush, Pharmastat consultant
Co-lead of CDISC Ebolavirus Therapeutic Area
standards development team

57
 WHO CORE CRF Annotations

58
Leveraging CDISC Standards to
Support Ebola Research
Efforts to standardize data collection and reporting
began last October
Started to annotate CRFs we received from the Oxford
University Clinical Research Unit (OUCRU) using
CDASH and SDTM variables:
– EVD CORE CRF (World Health Organization)
– RAPIDE BCV (Chimerix)
– Plasma Convalescence
Ended up with a finalized four-page CRF the WHO
shortened to only collect the most important, pertinent
information

59
Development of an EVD-specific
Standard
Project Team includes representatives from
Oxford, CDISC teams, government, and
pharma
Face-to-face meetings at CDISC Interchange
Conferences in November 2014 (Washington,
DC) and May 2015 (Basel, Switzerland)
Following the CDISC Standards Development
Process (COP-001) for TA standards
Anticipated draft EVD User Guide for Public
Review ~ November 2015

60
 Expanding CDISC Standards
Most variables and terms can be taken from the
existing SDTMIG and TAUGs (TB/Virology/Hepatitis
C)
Epidemiological data presents an opportunity for
growth
TB TAUG contains some ‘contact investigation’
criteria, controlled terms in SC domain:
– CNTCINV, CTRYDDTC, CTRYDEXP, EDLEVEL, EMPJOB,
JOBCLAS, NATORIG, PRICON, RISKPOP, RISKSOC,
SETCON, SRCCSINV, TYPCON
Terms may need to be added
– Discussion in Basel of a new ‘Contact Investigation’
domain, or a set of implementation rules to assist the
collection and reporting of epidemiological data
61
 What’s next?
Broadening use of CDISC standards from the
traditional clinical trial setting to outbreak and public
health situations
For Ebola, this includes:
– Names, surnames, addresses using GPS coordinates
– Death occurring pre-’exposure’ or pre-’admission’
– Contact investigation information
• Friends/family (names and addresses using GPS coordinates),
healthcare workers, animals
– Missing information
Ensuring broad public review of the EVD draft
standard will ensure that it will be useful for
research on future outbreaks of EVD and related
viruses
62
EHRs in Clinical Research

&
How can this be done?
Adapted from CDISC, 2015 
 CDER e‐Source Guidance 
“…promotes capturing source
data in electronic form…,”

Final
September, 2013  [assists] “in ensuring the
reliability, quality, integrity,
and traceability of electronic
source data.”

http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/UCM328691.pdf
64
Fitzmartin, 2015
 “Real World Data” 
Rob Califf and Rachel Sherman, FDA 12/10/2015
• Networked systems, electronic health records, 
electronic insurance claims databases, social 
media, patient registries, and smartphones 
and other personal devices together comprise 
an immense new set of sources for data about 
health and healthcare. 
• … these “real‐world” sources can provide data 
about patients in the setting of their 
environments—whether at home or at work—
and in the social context of their lives
http://blogs.fda.gov/fdavoice/index.php/2015/12/what‐we‐mean‐when‐we‐talk‐about‐data/
 “Electronic health records: new 
opportunities for clinical research”
J Intern Med. 2013 Dec;274(6):547‐60. doi: 10.1111/joim.12119. Epub 2013 Oct 18

• … emerging research infrastructures are being developed 
to ensure that EHRs can be used for secondary purposes 
such as clinical research, including the design and 
execution of clinical trials for new medicines. 
• …EHR systems should be able to exchange information 
through the use of recently published international 
standards for their interoperability and clinically 
validated information structures (such as archetypes and 
international health terminologies), to ensure consistent 
and more complete recording and sharing of data for 
various patient groups. 
Coorevits , et al., 2013
 Electronic Health Records (EHRs)

• Digital versions of a 
medical chart
• Real‐time, patient‐
centered records

67
CDISC, 2015 
 EHR  in Clinical Research
• EHR – EDC ‐ eSource 
• eCRF 
• Use of remote data capture (RDC) is increasing
– Oracle Clinical, Clintrial, Macro, Rave, eClinical 
Suite
• Electronic CRFs

CDISC, 2015 
 What can EHRs do?
• Make patient information 
easier to find
• Automate providers' 
workflows
• Provide evidence‐based 
tools
• Support changes in 
insurance requirements

69
CDISC, 2015 
 What are the benefits of EHRs?
Improved
•quality of patient care
•accuracy of diagnosis
•care coordination

Increased
• patient participation
• practice efficiencies
• cost savings
70
CDISC, 2015 
 EHR vs EMR
“An EHR is an EMR with interoperability.”

• Electronic Medical Record
– Medical and treatment 
history of patients in one 
practice
• Electronic Health Record
– Information can be shared
– Information from all 
clinicians involved
71 http://www.greenwayhealth.com/solutions/primesuite/emr-vs-ehr/
CDISC, 2015 
 Example ‐‐ Weight  
Patient’s weight EDC:
1. Pull up the patient in EHR and EDC (one system on each monitor)
2. Print out EHR form
3. Write the cycle and week on the document
4. Hole punch and insert the form into the patient “Chart” (manila 
folder)
5. Locate patient’s weight on the form and enter it into EDC
6. Redact the patient health information (PHI) on the form with 
adobe 
7. Add a header (trial, id, week/cycle, title, date) to the form
8. Save the form (format: date, ID, cycle #, form name) to a Shared 
folder
9. Attach the document in EDC as source data in the comments 
section
72
CDISC, 2015 
 FR Notice: CDER Promoting EHRs

https://www.federalregister.gov/articles/2015/06/26/2015‐15644/source‐data‐capture‐from‐electronic‐health‐
records‐using‐standardized‐clinical‐research‐data
 Source Data Capture from 
Electronic Health Records (EHRs)

http://www.fda.gov/Drugs/DevelopmentApprovalProcess/FormsSubmission
Requirements/ElectronicSubmissions/ucm464653.htm
Proposed Demonstration Solutions
 In Conclusion
• Substantial Evidence and FDA/CDER  Review of 
Confirmatory  (Phase 3) Clinical Trials
• CDISC Data Standards
– New CDER/CBER Requirements
– CDISC Basics
• Using eHealth Records in Clinical Research for 
Drug Development
[email protected]
Adapted from Oliva,, 2007