Development of the HMO Research Network Insights and

Development of the HMO Research Network: Insights and Lessons Learned Texas Medicaid Managed Care Quality Forum February 23, 2011 Sarah M. Greene, MPH Group Health Research Institute

Overview v What is the HMO Research Network? v How does it work? v What insights from our research network experience might be applicable in other settings?

In the beginning…(1994) v Frequent research partnerships among a small group of scientists based at a handful (n = 10) of integrated delivery systems v “Let’s form a Network!” v Coalition of the willing v v Shared commitment to non-proprietary, public domain research Objective: To improve population health and health care through collaborative research, dissemination, and sharing of best practices and methodologies

Health Systems + Research Centers =

Hallmarks of Member Sites v Stable presence in community v Mature v Comprehensive data v Defined population base v Commitment to research in the public’s interest v Combined population ~11 M v Mix of healthcare delivery arrangements

Characteristics of our Population

Then and Now: Milestones v CDC funds Vaccine Safety Datalink (mid-1990 s) v Founding member’s sabbatical at NIH (1997) v Establishment of the HMO Cancer Research Network v Spawned development of other topical networks in pharmacoepidemiology, CVD, mental health v Infrastructure building contract in 2005 – codified processes, created resources v Surfeit of ARRA funding, still more topic-specific networks (asthma, diabetes, hepatitis C) v Today: Invited by Francis Collins to propose a “Collaboratory” to facilitate broad platform of population-based research

370 researchers What do they do?

Many Networks within the Network Asthma CRN HMO Research Network Diabetes TBD RN PGRN Hep C CVRN DRN CERT MHRN SPAN

How Does It Work? Governance v Fairly informal v Has functioned largely as an alliance of willing collaborators with little formal organizational structure v Board of Governors (1 Center Director/site) provides overall leadership; Chair rotates every 2 years v Committees oversee data development, stewardship of HMORN’s key assets (scientists, infrastructure, etc. ) v Financial contributions from each support some centralized administration functions (accounting, communication, web) v Topical Research Networks are linked but still autonomous v Adolescence – changes are on the horizon

How Does It Work? Data v Model of choice: the “Virtual Data Warehouse” (VDW) v Could also call it a Distributed Data Warehouse - the data are real enough § but each site maintains its native data at its own site until needed for approved research activities v The VDW is… …a series of dataset standards and automated processes at each of 11 sites, § that allow a SAS program written at one site to be run against all the others quickly § and with a minimum of site-specific customization §

Schematic of How the VDW works Leadership through VDW Oversight Committee – technical and scientific experts who prioritize, set milestones, identify quality assurance issues

VDW Successes, Persistent Challenges v Assuages concerns related to centralized data repository v v v All sites have functional version of VDW with 6 data areas Highly dependent on health system; no control over health system migrating to a new EHR v Enables rapid turnaround on feasibility requests Site-to-site VDW implementation varies (for many reasons) v Significant investment of time and resources at site level v Federated data model defies easy explanation v Misperceptions are common – assume it’s automated (it’s NOT!) v Two-way communication as anomalies are identified v v “How may ____ cases were there at sites X, Y, Z in 2001? ” Ongoing opportunity to improve data quality Aggregation across dozens of underlying legacy datasets Yes, we do use it for research

The HMORN Today Dedicated contract to build HMORNwide infrastructure Membership growth and scientific synergies Health reform debate, ARRA and CER priorities Scientific, And Operational, growing & Financial pains Growth! Member $$ assessments initiated Demonstrated success of HMORN projects = visibility

Insights, Implications, Lessons

Unique features of delivery system research Why do we do this? Advantages of scientific evidence from large, integrated health plans: – Reflects care in real-world, everyday settings – Follows large, stable, diverse populations over time; generalizable – Provides access to data on health outcomes and the cost of care – “the whole picture”

Ideal: Learning Health Care Systems

“Meaningful Use” of Research Findings v The HMORN enables reciprocal or bidirectional learning between the health research enterprise and national priorities, as well as between research and health systems

Hard, messy work v Knowledge translation = holy grail of biomedicine v MDs: 75 studies and 12 journals/day just to keep up

Ingredients for Research Network Success v Shared vision and purpose v Commitment to dissemination and diffusion of findings v Sustainable financial model v Vigilance about collaboration’s “burden to benefit ratio” v Alignment of approaches to data use/aggregation/sharing v Evaluation – check-in systematically w/research teams v v Bring researchers & practitioners/providers together as often as possible to share views, ideas Recognize and leverage diversity within the Network (of the population, the research disciplines, the practice settings)

Thank you! www. hmorn. org greene. sm@ghc. org