Local Accountability for Research Protection in

Local Accountability for Research Protection in

Health Services Research

Seth Eisen, MD, MScDirector, HSR&D

• What is Health Services Research?

• Why is HSR important to VA?

• Why are VA’s electronic medical databases important to HSR?

• What are some examples of HSR?

• What are some unintended potential harms that might result from HSR?

• How do we reduce the likelihood of unintended harm from HSR?

Some Questions

Research Mission

“To discover knowledge and create innovations that advance the health and

care of veterans and the nation.”

Health services research (HSR) is a multidisciplinary field of basic and

applied inquiry that examines health care services regarding:

• access • use • patient preferences• costs• quality

• delivery• organization• financing • clinical practice• outcomes

Fundamental Goal of VHA: Provide the Highest Quality Care to Veterans

Health Services Research is a vital component of the bench-

to-bedside research team.

Health Services Research Involves Collaboration among:

• Clinical epidemiologists• Clinical researchers• Organization researchers• Implementation

researchers• Statisticians

• Informatists• Psychologists• Sociologists• Economists• Public health experts• Clinical ethicists• Care providers and managers

Health Services Researchers Love Data

• Outpatient Encounters: 450k 129 M • Inpatient Discharges: 2k 0.6 M • Pharmacy Fills: 600k 172 M• Chemistry Lab Tests: 900k 250 M • Radiology Procedures: 475k 135 M

VA Workload: 2006

Characteristic Daily Annual

Provided by 1400+ Points of Care

Data Added to VHA Computer Database

• Progress Notes, Discharge Summaries +638k 874 M• Orders +955k 1.65 B• Images +884k 590 M• Vital Signs +729k 1.06 B• Inpt Meds Administered +607k 850 M

Statistics through December 2006

Information Characteristic Added per Workday Added in 2006

Categories of Health Services Research

• Analyses of published literature• Medical records based research• Patient observation research• Clinical trials research

Analysis of the published literature

Portion of a Table Describing Characteristics of Dementia Screening

Tests

Recommended Dementia Screening Test Use

Medical Records Based Research

Psychiatric Diagnosis Prevalence as a Function of Gender and Race/Ethnicity

Patient Observation Research

Change in Self-Reported Patient Physical Health During an 18 Month Period, by VISN

Actual better than expected Actual worse than expected

Clinical Trial Research

Impact of Reminder on Beta-Blocker Rx After 9 Months as a Function of Initial Referring Clinic

74% 66%

Clinical Trial Research

Impact of Falls, Incontinence, and Dementia Interventions on “Masked Conditions”

VA Health Services Research:

• Identified optimal dementia screening instruments for specific clinical needs

• Demonstrated that the prevalence of psychiatric disorders is common among OIF/OEF veterans seeking VA care

• Demonstrated that declines in patient reported health varies from VISN to VISN

VA Health Services Research:

• Demonstrated that a simple reminder mechanism improves MD prescribing

• Demonstrated that a preventive care reminder does not reduce MD attention to other preventive care

Summary of VA HSR Research Strengths

• Large national system• Outstanding epidemiologists & health services

researchers• Strong academic affiliations• Ready access to assistance through HSR&D Centers

of Excellence and Resource Centers• Electronic medical record• Accessible, centralized and regional clinical

databases• Organizational structure that facilitates

implementation

Just because there is minimal physical risk, that doesn’t mean that the risk from HSR is trivial

Risk Considerations

• Confidentiality and privacy• Psychological & other impacts on

participants (patient, staff subjects)• Impact on health care organization

Electronic Data is Particularly Vulnerable to Confidentiality Violations

• VA medical database is entirely electronic• Millions of items of confidential information

can be carried in a shirt pocket• VA database can be expanded through

merging with Medicare & other databases

Electronic Data is Particularly Vulnerable to Confidentiality Violations

• Merging 2 anonymized databases can produce identifiable information

• VA researchers have many collaborators• Text may be difficult to de-identify

Approaches to Reducing Vulnerability of Electronic Medical Record Data

• Exclude data – Only provide data required for analyses

• Transform data – Introduce data variation to reduce risk but not impact research quality

• Establish different datasets – Create a different dataset for each researcher

Approaches to Reducing Vulnerability of Electronic Medical Record Data

• Data obfuscation – replace specific information by more general information

• Evaluate data fields for limited variation• Establish physical security

Balance Confidentiality & Privacy Protection with Research Benefit

• Evaluate potential impact of research success on health care

• Compare risk of confidentiality violation from research to risk posed by usual care

• REMEMBER: Risk can never by completely eliminated, only reduced

Foster a “Culture of Concern” for Research Risks & Benefits

• Require that PI present research to IRB in person

• Invite investigators to attend an IRB meeting• Include information security professional on

IRB• Emphasize the importance of research

risk/benefit issues to HSR mentors

Focus Leadership on Issues Related to Health Services Research

Risks & Benefits• Department & division chairpersons & research

leaders regularly discuss research ethics & risk issues• Evaluate the commitment of leadership &

researchers to research risk/benefit issues• When possible ethics or risk issues are recognized,

discuss in a non-punitive environment

Educate in Health Services Research Issues

• Devote 5 minutes of every clinical journal club presentation to ethics of article

• Discuss at grand rounds & other conferences• Introduce into affiliated academic institution

coursework• Require as component of VA Career Development

Awards• Include in VA Research Day• Include research ethics ‘pop-up’ on HSR&D website

Summary

• Create a “Culture of Concern”• Enlist enthusiastic support from leadership• Educate and re-educate

No Single Approach Will be Effective.

Multiple Approaches, Varied and Repeated, are Required.

The Stakes Are HIGH

• Evaluate projects thoughtfully• Ask questions when you don’t understand• Accidents happen: evaluate the problem,

develop and implement solutions• Remember: Local issues may have an

immediate national impact• Continued thriving of VA research depends on

the success of our mission