Innovative and Inexpensive Approaches to Health Research on Women and Disadvantaged Populations (Text Version)
On September 16, 2009, Carol Mangione made this presentation at the 2009 Annual Conference. Select to access the PowerPoint® presentation (374 KB).
Research to Reform:
Achieving Health System Change, September 13-16, 2009
Innovative and Inexpensive Approaches to Health Research on Women and Disadvantaged Populations
Carol M. Mangione, M.D., M.S.P.H.
Professor of Medicine and Public Health
David Geffen School of Medicine at UCLA
UCLA School of Public Health
- An example of a multi-center academic/governmental/health plan collaboration
- Key TRIAD findings for women with diabetes
- Advantages of an Academic/Health Plan Partnership
- Data related challenges and limitations
- An academician's perspective on partnered research with health plans
- Development of a Virtual Lab for Women's Health Research in California - a new project in the planning phase
Translating Research Into Action for Diabetes
A Multi-Center Study of Diabetes Care in Managed Care Settings
Rationale for TRIAD - 1998
- Diabetes is a large, growing, costly and complicated challenge for the U.S. health care system.
- Many effective interventions are not being optimally implemented, indicating missed opportunities to reduce the burden of diabetes.
- Systems approaches (e.g., disease management) offer possibilities for improving diabetes care and outcomes.
- Managed care was an important setting in which to study the system-level barriers and facilitators to better care and outcomes.
TRIAD Sites and Sponsoring Agencies
Pacific Health Research Institute
Translational Research Centers / TRIAD Study Sites
Centers for Disease Control - Sponsor
Veterans Health Administration / TRIAD Study Sites (2000-2004)
National Institute of Digestive and Diabetes and Kidney Disorders - Sponsor
TRIAD Study Group: Principal Investigators and Sponsors
- Indiana University
David Marrero, PhD
- Kaiser Permanente, N. California
Joe Selby, MD, MPH
- University of Michigan
William Herman, MD, MPH
- Pacific Health Research Institute
David Curb, MD, MPH
- University of Medicine and Dentistry of New Jersey
Jesse C. Crosson, PhD
(Formerly Norman Lasser, MD, PhD)
- UCLA School of Medicine
Carol Mangione, MD, MSPH
Arleen F. Brown, MD, PhD
- Centers for Disease Control & Prevention - Sponsor
Ed Gregg, PhD
Ted Thompson, MS
- National Institute of Diabetes and Digestive and Kidney Diseases - Sponsor
Sanford Garfield, PhD
- VA TRIAD Study (5 sites)
Eve Kerr, MD
TRIAD Design and Structure
- Multi-center study: 10 managed care health plans.
- Multi-level study: diabetes outcomes study with assessment of 11,927 diabetic patients, linked to measurement of 68 provider groups, 10 health plans, and hundreds of communities that serve them.
- Multi-design study: Unified, multi-center cohort study with focused evaluations and natural experiments overlaid on broader structure.
- Diverse in age, gender, race/ethnicity, socioeconomic status, geography, and type of system.
TRIAD Nested Sampling Scheme
Timeline of Original TRIAD Cohort and CVD Studies
Original TRIAD Focus
- "System-level" analyses of managed care organizational characteristics and services:
- Impact of diabetes interventions in everyday practice
- Possibility that managed care creates barriers to care
- "Patient-level" factors in this large and diverse cohort of diabetic patients:
- Variation in quality of care across diverse populations
- role of socioeconomic position on health status, health behaviors, diabetes complications, comorbidities
CVD Risk Factor Management by Gender
Ferrara et al., Diabetes Care, 2004
Key TRIAD Findings Relevant to Women's Health
- Women with diabetes are less likely to be on aspirin and statins and to have LDL tested - Ferrara et al., Diabetes Care, 2004
- Slightly worse levels of several cardiovascular disease care processes and intermediate outcomes were found in diabetic women compared with men - Ferrara et al., Diabetes Care, 2008.
TRIAD Gestational Diabetes Findings
- More than 80% of women with a history of gestational diabetes (GDM) reported receiving counseling on lifestyle modification but less than one-third reported receiving postpartum diabetes screening - Kim et al., Diabetes Care, 2007.
- 90% of women with a history of GDM recognized that GDM was a risk factor for future diabetes, but only 16% believed that they themselves had a high chance of developing diabetes - Kim et al., Diabetes Care, 2007.
- In women with diabetes of childbearing age, older age, higher BMI, and no insulin use were associated with lower likelihood of pre-conception counseling regarding glucose control and family planning - Kim et al., Am J Ob Gyn, 2005.
- First multi-agency collaboration in diabetes multi-disciplinary translation research: CDC, NIDDK, VA.
- Major training ground for junior and minority public health researchers.
- Creation of the TRIAD project required collaboration from 10 health plans and a willingness to execute data use agreements, share claims over multiple years and participate in the recruitment of beneficiaries and providers for targeted surveys and medical record reviews.
- Major influence on health policy and public health response to diabetes:
- 41 peer-reviewed articles published or in press, with dozens more in preparation.
- Has influenced disease management practices in participating managed care health plans.
An academician's perspective on partnered research with health plans
- The data set is a resource that no one investigator could have created
- IRB, HIPAA, regulatory, and legal considerations
- Resources required to execute the Data Use Agreement, data manipulation and transfer
- Risks in terms of findings and publication rites, identification of the health plans, and data security issues - value added by having the merged dataset inside the CDC.
- The wider use of EMRs hold great promise for lowering the cost and time needed to conduct research on clinical and therapeutic approaches if the privacy and IT hurdles can be addressed.
- Could studies such as TRIAD scale up to the State level and have a multi-condition focus?
State of California Virtual Laboratory for Women's Health Research
Carol M. Mangione, M.D., M.S.P.H.
Janet Pregler, M.D.
Allison Diamant, M.D., M.S.
University of California, Los Angeles (UCLA) National Center of Excellence in Women's Health
- Studying existing data is a cost-efficient way to perform research to improve quality, safety, efficiency and effectiveness of healthcare for all Americans, including women and minorities.
- EMR and administrative claims data are under-used resources for real time evaluations of quality of care, and variation in care by gender, race/ethnicity, and socioeconomic status.
- "Easy" access to multiple merged databases representing different geographical areas, healthcare systems, and racial, ethnic, and socioeconomic groups potentially could provide an efficient way to study comparative effectiveness, variation in access to healthcare and health equity.
Rationale for California
- California's racial and ethnic minority populations are socioeconomically and geographically diverse
- The State has been a leader in using survey and claims research to guide public policy. The California Health Interview Survey (CHIS) provides extensive data on the population of California as a whole. The Office of Statewide Health Planning and Development (OSHPD) gathers a wide range of health data
- California is home to some of the most experienced providers of managed care in the nation, including Kaiser Permanente and others. These entities have extensive longitudinal data on their patient populations, and experience leading and participating in research and evaluation projects using secondary data analysis.
- To create a coalition of governmental, academic, and private entities within the State of California to provide access to a diverse set of existing data to answer research questions relating to the health of women and minorities
- Coalition members will contribute to a data warehouse that contains de-identified public use files that can be accessed using an interface that is easy to use, and that is available for a wide range of public and private investigations and to a broad segment of the research community.
- To have a data warehouse that greatly streamlines the use of existing data to answer policy relevant questions by eliminating the need for each individual researcher to negotiate data acquisition, legal and confidentiality arrangements with one or more of the partners in the coalition.
- The project will provide a "roadmap" for other states to follow to create additional state wide resources in other regions of the U.S.
Women's Health Research in California
- California is home to 2 Centers of Excellence in Women's Health, at UCSF and UCLA, funded by the US Department of Health and Human Services. These 2 Centers of Excellence have a long history of collaboration, as well as extensive ties to governmental and community entities across California.
- The UCLA CoE has an ongoing collaboration moving forward the women's health agenda in California with the State of California Office of Women's Health and the Los Angeles County Office of Women's Health, as well as the National Center of Excellence in Women's Health.
Goals for the Planning Year
- The project will build upon existing relationships with the COEs, the UCLA Division of Health Services Research, and the RCI to develop a coalition of public and private partnerships with the major holders of data in the State. Inclusion of data sets with detailed medical, laboratory, and pharmaceutical claims will greatly enhance our ability to compare therapeutic approaches.
- Leaders from state and local governments, health plans, healthcare provider groups, pharmacy management companies, academia, and RAND will be convened at a meeting to develop a plan for creating a database warehouse and to tackle the barriers.
- The plan will be informed by previous but smaller efforts to merge data from diverse entities to conduct health services research
Issues to be Addressed
- IRB, HIPAA, Regulatory, and Market Competition Considerations
- Consensus on key elements in a Data Use Agreements
- Protection of proprietary information
- Protection of Publication Rites
- A multi-layered approach to the creation of the data warehouse will be considered, ranging from a relatively limited store (claims for service use and medications) to more expansive data store (inclusion of specific data elements most relevant to women's health from EMRs). Costs associated with each level will be estimated.
- In addition to geocoding data, strategies for the collection of patient-level graphic data will be considered, including collaboration and linkage with CHIS and OSHPD.
- In parallel the project will engage IT expertise and a variety of platforms, systems, and partnerships will be considered with the goal of achieving harmonization of data from multiple sources. Costs of overcoming the IT issues with be estimated.
Additional Issues to be addressed
- The quality and content of available data will be assessed.
- A work group in the coalition will determine the type of research questions that could be answered by the proposed data warehouse with an eye toward creating the needed information to conduct comparative effectiveness research and studies of health and health care disparities across a wide array of health systems and diverse populations.
- The process of negotiating the basis for the multicenter database warehouse will be documented using qualitative research methods. IRB approval will be obtained for this portion of the project.
Next Steps After the Planning Year
- Once this initial report has been submitted, the Virtual Lab Coordinating Center will try to identify funding for a limited solicitation to our collaborators to pilot test the data resource.
- The participating entities will be encouraged to design specific projects that are likely to undergo a separate, peer-reviewed process at AHRQ, NIH, state government, foundations and For-profit entities.