Preparing for CPHQ..

an overview of concepts
Ghada Al-Barakati
Consultant Quality and Accreditation Department of Health and Medical Services Dubai, UAE

Information Management the

JCIA View Providing patient care is a complex endeavor that is highly dependant on information,.. To provide, coordinate and integrated services, healthcare organizations rely on information about the science of care, individual patients, care provided, results of care, and their own performance. Like Human, material and financial resources, information is a resource that must be managed effectively by organizations leaders. Every organization seeks to obtain, manage, and use information to improve patient outcomes and individual and overall organization performance.

Information Management the

JCIA View Overtime the organization becomes more effective in

Identifying information needs Designing and information Management system. Defining and capturing data and information. Analyzing data and transforming into information Integrating and using Information

Information Management the

JCIA View Although computerization and other technologies improve efficiency, the principles of good information management apply to all methods, whether paper based or electronic. These standards are designed to be equally compatible with non computerized systems and future technologies

Integrating Data Analysis into Performance Improvement

The quality professional need to analyze data & accurately integrate the information driven from them into the performance improvement process. He/she can use outcomes data, customer satisfaction data , risk management data, performance measurement data, utilization management data.

Information Management
The Goal of Information Management in healthcare To obtain, manage & use information to improve patient outcomes and individual performance in patient care, governance, management & support processes. JCAHO

Basic Concepts of IM

Data about the quality of care are very important to make informed decisions.

Basic Concepts of IM

Healthcare organizations must identify:

Who needs the information. What information they need.

Basic Concepts of IM
Healthcare organizations must develop a system that the right people receive the right information, at the right time, in the right way.

Basic Concepts of IM

Data must be carefully

defined, analyzed interpreted before usage.

Basic Concepts of IM

Not all data are useful in QM. It is patterns of care, not individual cases, that can help for QI. You need to integrate multiple measures of QUALITY to get a clear picture. Outcome information are insufficient without monitoring the process of care. Cost & Quality are inseparable.

Information Management Definitions

Data: Un-interpreted observations or Facts.

Information: Data transformed through analysis & interpretation into a form useful for decision making.

Why Do We Need Data

To know exactly where are we standing? What do we want to achieve? Are we on the right track? What did we achieve?... For all quality activities. Insufficient, incorrect, or differently interpreted data can easily be the major cause of flawed decisions

Why Information Management?

Proof of Quality: Quality decisions & actions based on facts. Demands of patients / purchasers. Clinical/Financial Data Mix: For integration of QM/UM/RM information. Reporting & Benchmarking: Performance measures reporting & comparison with similar organizations. Reporting responsibility to higher authorities.

Decision Making Process

Differences in Decision making basis
Traditionally Intuition Opinion Logic Rationalization Hearsay TQM Data collection Data analysis & interpretation Get information Lead to knowledge

What is a Good Decision?

A good decision is: taken and supported by the people affected by it. (Participatory management) based on facts and data (not just opinions). made quickly enough to meet deadlines, but not so quickly that important information are ignored. (Is timely) made knowing what the consequences will be and dealt with ahead of time (contingency plans)

Fact-Based Decision Making Information Management Steps

Identify critical information needed. Define data elements. Determine data collection plan and tools. Collect data. Aggregate & display data. Analyze data. Interpret data to get information. Information lead to knowledge. Take decision, implement, then monitor it

Common Sources of Information

Analytic / Qualitative Data Sources: Customer satisfaction surveys. Customer complaints. External sources /surveys. Numerical / Quantitative Data Sources: Common indices & registers, e.g. master pt. Index, surgery index, ER log, national registers... Severity indices.

Potential Quality Data Sources

Internal Sources: Patient records. Quality measurement summaries. Customer surveys. Clinical reviews: medication use, operative... Utilization-, risk- & case-management. External Sources: Reference database. Identified best practice. National rates.

Types of Information available

Qualitative: Mission, vision, values, regulations, community needs & expectations...etc. Quantitative: Costs, performance measures, case-mix, productivity, encounters, length of stay...etc Both Quantitative & Qualitative: Patient / employee satisfaction surveys.

Computers as Tools for IM

Hardware:

System unit (CPU). Input devices: keyboard, mouse, scanner, D. camera. Output devices: monitor, printer & speakers. Peripheral devices e.g. modems. Storage devices: RAM, ROM.

Computers as Tools for IM

Software: (Operating system software: controls input & output, handles file e.g. copying, deleting) Application software: word processor, spreadsheets, graphics Software to support QI.

How Can the MIS Help Leaders?

Assess achievements (by measures). Follow changes in financial resources. Evaluate current policies & processes effectiveness. Identify the need for system & policy changes. Provide information for prioritization. Monitor aspects of organization performance & take corrective actions. Judge progress toward strategic goals. Appointing, credentialing & re-credentialing. Determine goals for health improvement in the community. Define the organizations economy, efficiency & effectiveness. Help the governing body to evaluate & improve its performance.

Clinical Information Systems

Designed to support direct patient care. Have the potential for analyzing information. Examples:

medical records & their retrieval, computer-assessed medical decision making. Health risk programs. HMO encounter data.

Administrative Support IS
Aid the day-to-day operations. Include:

Financial information systems. Human resources information systems. Office automation systems,

Example: electronic mail, scheduling, facsimile transmitting, electronic spreadsheets ...

Decision Support IS
Deal with strategic planning functions. Include: Strategic planning. Resource allocation. Performance evaluation & monitoring. Services evaluation. Medical management.

Data Inventory Process

The Quality Council: Review the teams report. Establish improvement priorities. Monitor measurable objectives. Establish mechanisms for IM education.

Data Inventory Process

The Quality Council should select a cross-functional team to collect data. Train all team-members in team processes, sites to collect data, and what (defined) data to collect. The team will collect & document data.

Data Inventory Process

The data collection team then starts to analyze he data collected: Is it useful or (DRIP)? is there duplication? Is it valid for decision making? Who is using it? Is it shared.

Data Inventory Process

The team will report which data is/isnt: Based on established PI measures. Trended over time. Tracked by the organization.

Information & Knowledge

Data Collection Data Collection is a time consuming & costly process. Therefore one must Try to prevent duplicate data collection. Avoid taking actions on data that might not be reliable. Always pilot any new data collection process.

Data Collection Plan

Determine the 6 Ws: why, what, who, when, where, how. Structure the design. Choose & develop the sampling method Conduct the necessary training. Delegate responsibilities. Facilitate interdepartmental coordination Determine the budget required. Conduct a pilot procedure.

Ten Steps for Interpretation and Utilization of Information

1.

Plan & Organize for data collection & use:

2. 3. 4. 5. 6. 7. 8.

Identify responsibilities, Multidisciplinary collaboration, Anticipate barriers & put solutions. Verify & Correct collected data: Identify its limitations. Review & correct data. Identify systems that led to data errors.

Ten Steps for Interpretation and Utilization of Information

Identify important findings: Perform a preliminary data analysis Compare data with previous ones (internally). Compare with external data for similar organizations. Find out any trend over time. Translate data into meaningful information. Study further: Analyze variation. Review additional data e.g. case-mix, diagnostic categories ...etc. Conduct retrospective medical reviews. Perform process analysis. Present important findings: In a clear & concise manner.

Use tables & graphs

Ten Steps for Interpretation and Utilization of Information

Recommend Change: based on facts. Take Actions to implement change: Staff education & training. Change policies & processes. Monitor the Impact of the Change: Have change actually been implemented? Could compliance with it be enhanced? What effects it had on patient outcomes? Should change need modification? Do it need to be tested for a longer period or, on a wider scale?

Ten Steps for Interpretation and Utilization of Information

Communicate Results of Change: QI begins with communicating where we are & where we want to be. Then communicate, with others, what did we achieve & how did we achieve it? Record learned lessons for future, Celebrate Success & Continuously Improve.

Management Information Systems (MIS)

Can contain both, the manual & the automated methods that provide information for decision making. The quality of decisions is directly correlated with the availability of data & information.

Data Analysis

Interpretation of Data
Data Analysis It is the reporting, analysis, and ultimate interpretation of data that give them importance & vitality. Good presentation of data creates interest and enhances understanding. Data should be reported & analyzed on regular basis.

Statistical Analysis

Fundamental Concepts
Data Collection Tools/Instrument The devices (sheets) that researchers use to obtain & record their research / study data. They can be in the form of questionnaires, rating scales, interview transcripts...etc. The tool / instrument should be reliable & valid.

Fundamental Concepts
Reliability:

The extent to which a measuring tool or procedure yields the same results on repeated trials. The degree to which an instrument measures what it is intended to measure.

Validity:

Types of Quantitative Data

Count Data:

They are not continuous (discrete), they are whole numbers (digital), with no fractions, so they have gapes. They are 2 subtypes:

Nominal, Ordinal.

Types of Quantitative Data

Count Data: Data which consist of counts of observations or incidents fall into these two categories. Examples:

N: N: O: O:

Appointment no-show rates: males & females. Surgery Patients: pre-op. & post-op. Educational level. Likert scale.

Types of Quantitative Data

Continuous Data:

They are measured on scales (Measurement) that theoretically have no gaps (continuous), so they can have fractions. They are 2 subtypes:

Interval, Ratio.

Types of Quantitative Data

Continuous Data:

Data which give a measurement or number for each observation or unit, which can be subdivided. So, it can have fractions. Examples:

Weight, height & temp. Blood glucose. Infection rate. C-S rate.

Sampling
Population: A group of cases/people that meet certain criteria for inclusion in a study group.

e.g. all physicians, all utilization review nurses, all ladies who did give birth at a particular hospital.

Sampling: Getting a portion of the population to represent the entire population.

Purposes of Sampling

Provides a logical way of making statements (study) about a large group based on a small group. Allows us to generalize (apply) from the sample to the population if the selection process has been random & systematic.

Sampling Designs
Probability Sampling

Sampling which gives an equal / random chance for every person in that population to be selected. It can be one of 4 ways:

Simple random sampling. Systematic sampling. Stratified random sampling. Cluster sampling.

Simple Random Sampling

Each individual in the population have an equal chance to be chosen. Like pulling a name out of a hat containing all names of the concerned population

Systematic Sampling

The first case / individual will be selected by simple random sampling. The following ones will be selected by drawing every nth element.

Example: picking every fifth name of a list of that populations names

Stratified Random Sampling

The population are divided into strata (homogenous subgroups). Then, each member of a stratum got an equal chance to be selected. Examples:

Dividing the population by gender (males & females) then randomly selecting from each group separately. Dividing patients according to their residency (living in which area), or according to their economic class.

Cluster Sampling

The population are divided into groups/ clusters Then, randomly pick some of these groups/ clusters (pick the group as a whole, not members from each group). Example:

If you want to study children behavior at school, but you cant get all the students names in the area, you can list the names of schools and randomly pick

Non-probability Sampling

With this kind of sampling there is no way of estimating the probability that each member will be included in the sample. So, the result will be representative of the sample only, and cant be generalized to the population.

Convenience sampling. Purposive / judgment sampling. Expert sampling. Quota sampling.

Convenience Sampling

It is the use of any available group of subjects. Not randomized. Not representing the whole population. This type is used if subjects are difficult to identify, but are known to others (e.g. snowball sampling). Selection bias is likely to be present.

Judgment/Purposive Sampling

It is the selection of a particular group or groups based on certain criteria. The researcher uses his /her judgment to decide who can represent the population he/ she needs to study. It is a subjective method.

Example: using a group of nurses as to represent a cross section of women (working).

Expert Sampling

This is a type of purposive sampling where the researcher select experts in a given area related to the study. Used in the Delphi technique: Questionnaires on the study area are sent to experts to get their opinions.

Example: An organizer of a symposium or program who uses Delphi method to identify potential demands of the audiences.

Quota Sampling

The researcher identifies strata of the population & determines the proportions of elements needed to be studied from each. Example:

the sampling of 5% or 30 cases, whichever is greater, for reviewing blood-usage, surgery / proceduresfor quality reviews.

Sampling Designs
Sample Size The larger the sample, the more valid & accurate the study. A large sample is more likely to represent the population. Continuous data require smaller sample than categorical ones e.g. for a study about hypertension, if you use mild, moderate & sever youll need a larger sample than if you use 140/90159/99, 160/100179/109

Data Analysis
Data Analysis is the process of studying & interpreting
aggregated & displayed data, and drawing valid conclusions that lead to the proper decision making (based on facts).

Initial Analysis:

Done by those closest to the process being measured / studied, or by the team that is charted to improve the process. Include: aggregation, summarization, validating, comparing to other data, separating issues for P.R.

Intensive Analysis:

Done by those persons with knowledge & experience to study the process issue in-depth, and find out the type & causes of variation.

Analysis Process Questions

Does this data adequately represent the group? Is the sample size large enough? Is the accumulated data accurate? For each indicator, has the trigger for intensive analysis been reached? Is there a gap between the demonstrated performance & the expected outcome? Is the gap based on an isolated case or on a pattern or trend?

Conclusions about data

Conclusions about data may:

Identify unexpected patterns of care or clinical competence ( positive or negative). Focus attention on high priority issues for improvement. Identify differences in patterns of care between groups. Clarify the range of acceptable variation.

Statistical Handling of Numbers

Ordering Averaging Finding Variability Comparing

Statistical Handling of Numbers

Ordering: To see characteristics of group performance more easily. To compare groups of data. e.g. frequency distribution.

Statistical Handling of Numbers

Averaging: Arithmetic mean. Mode. Median.

Statistical Handling of Numbers

Finding Variability: The variation or scattering of data around an average value. e.g. range, standard deviation.

Statistical Handling of Numbers

Comparing: Comparing 2 distributions (scatter diagram) Comparing means Comparing rates.

Statistical Techniques
Measures of Central Tendency:

Measure of variability:

Mean, Median, Mode.

Tests of Statistical Significance: Parametric Tests: t-test, Regression analysis.

Range, Standard deviation.

Nonparametric Tests: Chi-square (X2) test. Related Concepts: Confidence intervals, Levels of significance.

Measures of Central Tendency:

Mean: The arithmetic average. The sum of all values divided by the number of scores. If the ages of a group of participants are: 49, 27,57, 33, 50 years, their average age will be: (49+27+57+33+50)/5

Median:

The middle score. If the total number of your scores is odd, it is the meddle value, e.g. the Median of: 1, 3, 4, 6, 7, 9, 9 is 6 . If the total number of your scores is even, it will be the average of the middle 2 scores, the Median of: 1, 3, 4, 6, 7, 9 is (4+6) / 2 = 5 . It doesnt take into consideration the quantitative values of the individual scores. It is not sensitive to extreme scores.

Mode:

The value that occurs most frequently, e.g. the mode of these values: 16, 21, 27, 33, 33, 33, 44, 60 is 33 . It tend to fluctuate widely from sample to sample in the same population. So, it is reported infrequently & seldom used alone.

Measure of variability
Range: The difference between the highest & the lowest values. It is sensitive to extreme values It is a quick estimate to variability. It is best reported as the values themselves (as a maximum & a minimum), not the distance between them, e.g. Ages of employees range from 22 to 60 years, it is better to represent the range as : 22 to 60 years, rather than as 38 years (60-22=38).

Standard Deviation:

An average of the deviations from/ scatter around the mean. The average spread of a group of values around their mean. All values in the distribution are taken into consideration.

Tests of Statistical Significance

t-test: To determine if the difference between 2 group means is significant. The 2 groups may be dependent (e.g. pre- and post-treatment) or independent (a group using the new treatment & a placebo group).

Regression Analysis:

It explains the inter-relationship among variables, So, it allows you to predict outcomes. If the correlation between 2 variables is perfect ( r = 1 or r = - 1), you can make a perfect prediction about the outcome. Multiple Regression Analysis: estimates the effects of 2 or more independent variables.

Chi-square ( X2 ) test:

Measures the statistical significance of a difference in proportions (or percentages). Commonly used in the medical literature.

Concepts Related to Tests of Significance

Confidence Interval (CI): CI provides a range of possible values around a sample estimate. It reflects the uncertainty which is always present when working with samples. e.g. we are 95% confident that(the result of the study).

Concepts Related to Tests of Significance

Level of Significance: Identifies the probability of observing a difference as large as the one found in a study when, in fact, there is no true difference between the groups. Indicated by the p value. A p < .05 means the study results are statistically significant.

Interpretation of Data

Data Reporting

It is important to first validate whether the data were collected accurately. Then, display them in an easily understood format, that data will tell its own story. Together with the highly representative chart or graph, a brief summary should be provided. It is essential to provide a contextual background of data, to aid effective interpretation of them.

Basic rules for data presentation:

Dr. Walter Shewart put 2 basic rules for data presentation: A graph should be used to display data & a table of values should accompany, with specifics regarding data collection (e.g. how, when, where & from whom) The time-order of data collection.

Outcomes Management

The use of information about the care process & outcomes through monitoring, measuring & data analysis) to achieve optimal patient outcome.

Comparable Data Measures

HEDIS : Health Plan Employer Data & Information Set: The NCQA developed a set of performance measure for the managed care organizations, to provide information on the quality of their care to purchasers & the community. NAQAs Quality Compass: A national database of comparative information about the quality of the nations managed care plans, including HEDIS & accreditation information.

Comparable Data Measures

ORYX :
A set of performance measures developed by the Joint Commission to integrate the use of performance measures into the accreditation process . And, to monitor the performance of accredited organizations on a continuous basis. Is designed to help healthcare organizations strengthen their efforts to identify the issues that require attention and, verify the effectiveness of their corrective actions.

Benchmarking

It is a process of measuring services & practices against competitors in order to find out & implement the best practice. It has been long used in industry. Engineer Joseph Juran urged organizations to ask themselves what other organizations do to get better results.

Benchmarking
The ideal assessment process should incorporate 4 basic comparisons: With self, over time. With others, With standards, and With best practices. The last 3 constitute Benchmarking.

Why Benchmarking?

To gain insight by comparing an organizations clinical outcomes & processes to those of the benchmark for the purpose of achieving innovative breakthroughs in improvement.

Steps of Benchmarking

Select a clinical disease or a process that your organization has identified as needing improvement in quality or financial outcomes. Identify another organization that has achieved the desired outcomes. Specify the outcomes to be measured. Prepare your organizations data for that specific outcome. Compare you data to those of the benchmark organization. Analyze data and determine areas for improvement. Develop improvement plans to address those areas.

Steps of Benchmarking

Implement your improvement plan on a small scale ( pilot study). Monitor & measure its outcomes. If the desired improvement has achieved, Implement your plan widely. Continue to monitor your process & measure outcomes over time, to ensure maintaining the desired outcomes. If the desired improvement has not occurred, repeat the improvement cycle : (plan, pilot, implement & measure). When achieving the desired outcome Celebrate success, and Find another process to improve.

Interpreting Benchmarking Data

The quality management professional will assist the organization & staff with the interpretation of benchmarking results. The goal to identify how to improve the organizations outcomes, not to identify the differences. The QM professional should provide the pertinent results to the right audience.

Interpreting Benchmarking Data

When variance is identified, the departments / services involved musk ask:

What are we doing? How are we doing it? What is the measure of how well we do? Why are we looking for improvement?

Benchmarking & Clinical Pathways

Benchmarking will identify the most problematic & the highest cost DRGs for which your staff should develop clinical pathways. After developing & implementing the clinical pathways: Assess internal variability. Benchmark against external sources.

Incident Reporting

It provides a practical method for risk identification, and allows the organization & the risk manager to track untoward events. By trending IRs , your organization can determine theKey Exposure Areas & Identify potentially compensable events while they are occurring, or as soon as possible thereafter. Healthcare organization should develop its policy for reporting incidents.

Incident Reporting
The policy should clearly state that: Reporting of incidents is not a punishable offense.

Incident Reporting
The policy should include: A format to help protect IR from discovery, no copying is allowed, nor to document it in the medical record. Not reporting an incident, can have serious consequences.

Occurrence Screening

It is a concurrent review process of the medical records, to determine if any adverse event has occurred that is a result of healthcare management and not of the disease process. It is another method to help the risk manager with risk identification within the organization. The screening process is integrated with QM & UM record review, the 3 processes are often done simultaneously.

Advantages of Occurrence Screening Method

All patient records are reviewed. Reviews are concurrent, not retrospective ( an immediate benefit ). Reviews are based on objective criteria. Department-specific criteria can be developed. Healthcare professionals are the cornerstone of the process, they are the ones to develop & approve criteria for reviews and review the findings. The screening process is integrated with QM & UM record review.

Analysis of IR & Occurrence Data

The risk manager, with the cooperation of the quality manager, perform the analysis & report of the incident / occurrence reports. This analysis provides insight into potential problem areas. Risk management results should be reported to the PI coordinating Council & the safety committee. A multidisciplinary forum will be formed from the members of the council & committee. The forum will review & analyze the report. After a thorough analysis, the forum will recommend actions to be implemented to decrease ( hopefully prevent) future occurrence of the incident, or at least minimize its impact on the outcome. Continuous monitoring & trending of occurrence / IRs are essential to test the effectiveness of the actions.

The Medical Record

The Medical Records

The medical record is the primary legal document, as well as the primary data source for recording & ascertaining the quality of healthcare delivery to patients. Its contents must be sufficiently detailed & organized to help practitioners, other consultants, and utilization & quality management.

Purposes of The Medical Records

Basis for planning & continuity of care. Communication among practitioners. Support for the diagnosis (history, examination & investigations). Justification for treatment (consent, orders & observations). Documentation of the course & results of: procedures, tests & therapy . Legal protection for: patient, practitioner & the organization. Provide Data for use in continuing education, research, risk & utilization management, performance appraisal, and quality measurement, assessment & improvement Identity of the patient: identification data & psychosocial needs

Confidentiality of Information

Although timely and easy access to data & information is required, organization must also ensure that data & information are safeguarded. The organization is responsible for protecting records against: loss, defacement, tampering & any unauthorized use. Healthcare organizations must keep confidential all information related to peer review and patient care. Access to each category of information should be appropriate to the users title & job function.

Confidentiality P&P s

Healthcare organizations should have a clear policy to safeguard the medical records, whether they are written or computerizes, as well as parts of the records e.g.

Identification of individuals with access to information. Specifying information to which individuals have access. Requirements for those individuals to keep the information confidential. Requirements for release of health information. Securing information against damage.

videotaping of procedures.

Examples of Security Systems

Portions of medical records may be stored separately (e.g. certain types of psychiatric information). Restrict access to computer files or portions of them, by the use of security codes. Organizations should have an adequate backup plan for computer application.

Release of Information

A written consent is required to release pt. Information to anyone outside the organization. The consent should contain:

Patients name. Name of the person / organization requesting he information. Exact material to be released & at what time Reason & anticipated use of information. Signature (dated) of the patient.

Authorized Release of Information

Patient information may be released without written authorization to:

Governing body. Organization director. Healthcare professionals involved in the care of the patient. Those who are responsible for quality improvement activities. The medical records personnel.

Confidentiality & Security

Patient records Peer review Utilization management

Confidentiality of Patient Records

What are the legal basis of medical records confidentiality? The physician-patient privilege is recognized as a special relationship by law, and confidentiality is essential for its maintenance.

Information Covered by the Privilege:

Oral, written, X-rays, cardiograph strips, lab. Results & any other information concerning a patients condition

which have been communicated in the context of Physician- or practitioner-patient relationship. Which have been given with the expectation that it remains confidential.