The NIH convened a group of senior scientists, administrators, and policy makers from the HMO Research Network (HMORN), supported by the National Cancer Institute (NCI) and National Heart, Lung and Blood Institute (NHLBI), as well as the Centers for Medicare and Medicaid Services (CMS) and Agency for Healthcare Research and Quality (AHRQ), to identify opportunities and strategies to enhance coordination and research infrastructure of the HMORN to benefit research programs of many ICs.
The NIH Director, Dr. Francis Collins, opened the meeting and gave the charge to participants. Senior HMORN representatives followed with brief presentations on the successes, challenges, and opportunities in three critical areas that could benefit from network enhancement:
- Rapid scale-up of mega-epidemiologic studies to enhance understanding of etiology, diagnostics, therapeutics, personalized medicine and disease burden;
- Transformation of the clinical trial enterprise; and
- Development of an evidentiary basis for improved health care delivery.
Drs. Michael Lauer (NHLBI) and Robert Croyle (NCI) then led a group discussion which culminated in the identification of research opportunities and next steps. Workshop Highlights:
- Developing a more synergistic partnership between HMORN members and the NIH, especially with smaller institutes, could maximize the efficiencies and benefits of a HMO Collaboratory.
- HMORN representatives suggested expanding comparative effectiveness research (CER) capabilities and developing joint research positions among NIH, HMO-affiliated academic health centers (AHCs) and training programs.
- A thoughtful organizational structure would help forward research projects more quickly and generate new research hypotheses especially in the areas of mega-epidemiology, clinical trials, and health care delivery research.
- HMORN representatives stated that many of their faculty members are "at the limit of bandwidth for taking on more collaborative research,"; and that infrastructure expansion would entail additional resources, including for central core staff.
NIH Perspective and Vision
Dr. Collins, NIH Director, charged the participants with finding possible solutions for creating a more vigorous collaborative network which many NIH ICs could access in order to get rapid responses to important questions about healthcare. Dr. Collins highlighted five major opportunities for the NIH, several of which could be addressed by strengthening the HMORN and NIH collaborations. The themes include:
- Applications for high-throughput technologies to better understand fundamental biology and to uncover the causes of specific diseases;
- Translating basic science discoveries into new and better treatments;
- Putting science to work for the benefit of health care reform;
- Encouraging a greater focus on global health;
- Reinvigorating and empowering the biomedical research community by encouraging young investigators and stimulating innovation.
The idea of an HMO Collaboratory has been partially realized by the existence of the HMORN, a consortium of 15 practice-based organizations. This network is geographically distributed, and each of the participating health systems has an advanced system of electronic medical records, access to thousands of patients and biospecimens, and a track record of significant research with many NIH institutes, as well as AHRQ. Developing a more synergistic partnership between HMORN members and the NIH, especially with smaller institutes, could maximize the efficiencies and benefits of a HMO Collaboratory.
A thoughtful organizational structure would help accelerate research projects and generate new research hypotheses. Dr. Collins suggested that this could be accomplished by focusing on three priority areas:
- Mega-epidemiology studies that employ high-throughput genetic and laboratory analyses, biobanking, record linkage, standardized data collection and reporting, and high quality information technology systems to study disease burden, etiology, course, treatment and prevention, to empower personalized medicine,
- Transformation of government-sponsored clinical trials to be efficient, practical, and applicable to patients with common and rare diseases, Improving health care delivery research by creating a shared research platform to enable health systems analyses (including CER) to develop and deliver practical evidence on health care delivery, including provider incentives, and inform health care reform.
Dr. Collins emphasized that the goal of the meeting is to establish previously unrecognized opportunities, set priorities, and develop the next steps that would catalyze a more effective network. He mentioned that such a trans-NIH enterprise could be supported by the NIH Common Fund.
Dr. Larson, Executive Director of Group Health Research Institute and Chair of the HMORN Board of Governors, described the history of the HMORN. The HMORN sites consists of 15 U.S. research centers (along with one international member in Israel), all of which are embedded in integrated delivery systems; they are highly interdisciplinary and collaborative across sites, have long-standing partnerships with academic institutions, and are different enough to ensure diversity among patient populations, patterns of care, and delivery system designs. The culture is one of evidence-based practice and supports the concept of a learning healthcare system.
Dr. Larson addressed the goal of this meeting to explore what it would take to scale up and augment the HMORN's capabilities as a research collaboratory with the NIH and other external partners. He identified two major areas that present barriers â€” data/infrastructure and capacity. He emphasized that it will take significant efforts to align the data systems of all HMORN members. In addition, the HMORN is near its capacity with regard to the amount of work that can be accomplished with the current system. Several HMORN leaders spoke about the barriers, what it would take, and what could be done to scale up the HMO collaboratory.
Dr. Joe Selby and Dr. Raymond Baxter from Kaiser Permanente addressed the barriers and facilitators for the HMORN becoming a leader in mega-epidemiology, with regard to resources, science and technology. Although many HMORN sites have been doing large-scale epidemiological studies for years, the challenge will be to bring these projects to a network-wide, membership-wide level. They identified several barriers to accomplishing this, including data availability, comparability and completeness, a scarcity of behavioral, environmental and patient reported outcomes data, the learning curve for using new EMR data, the time limits of existing HMORN investigators, and data access by non-HMORN scientists with concerns for patient privacy, data interpretability, and proprietary interests. They emphasized that it will take advancement in several key areas to scale up the collaboratory: funding to enhance the virtual data warehouse; standardization and validation of data definitions, endpoints, and collection of patient-reported data; a systematic and integrated approach to collection and storage of biospecimens that can be linked to EMR data (consent for research samples built into routine care); re-thinking the national research training strategies; and a thoughtful approach to expanding collaborations with non-population-based researchers.
Clinical Trial Enterprise
Drs. Eric Larson and Alan Go from Kaiser Permanente and Dr. James Hereford from Group Cooperative Health discussed the barriers and potential for improving the clinical trial enterprise to a network-wide level. They highlighted a need for better trial coordination and prioritization based on scientific evidence and patients' needs, standardization of procedures and integration of clinical trial protocols in EMRs, increased investigator-level research capacities, a more efficient infrastructure to streamline IRB and regulatory issues and integrate clinical trials into routine patient care, a cross-disciplinary approach to trials, and a better health care plan that aligns reimbursement with true costs of care. Several HMORN sites already have large, productive clinical trials groups that could be used as a guide for leveraging HMORN assets.
Health Care Delivery
Drs. Walter Stewart and Ronald Paulus from Geisinger Health System discussed how the health care delivery system could be leveraged to improve research and enhance the relationships between the infrastructure of the network, scientific resources, and delivery of care. However, there are several barriers. There is a lack of resources for enhancing infrastructure, such as centralized management and data standardization. There are organizational and socio-cultural barriers to collaboration and constant changes in heath care, which can affect patient enrollment and benefit design. Improvements would take a substantial investment to develop a central management system and should include development of more harmonized policies for data use and IRB review. The key issue will be to establish stronger ties between research centers and their delivery systems.
Discussion Session 1
Dr. Collins asked how representative of the U.S. population is the HMORN membership and; are there any limitations in using the network for research? The HMORN panel noted that the HMORN population is representative of families with health insurance, and consequently, patients in clinical trials tend to be wealthier individuals on average. Rural populations were underrepresented until Geisinger and Scott and White Health System (based in central Texas) joined the network. The panel commented that the patient population is very stable over time; however, the population composition often changes when there is a change in health benefits. The key would be to develop a system by which different HMORN members could access patient stability data from other members. Individuals will need to be placed within the context of their environment and communities. Some HMORN members have developed metrics around activity, behavior, diet, and other lifestyle factors to better track patient populations, and potential risk factors, and predict outcomes.
The NIH asked what the biggest barriers were for the HMORN in conducting clinical trials. The panel commented that two components of clinical trials have undergone great improvement. The first is streamlining the institutional review board (IRB) process and the second is the integration of electronic medical record (EMR) systems such as EPIC. The HMORN has undergone many efforts to improve access to various databases, similar to CMS and Veterans Administration (VA) databases. This enhanced data access capability can help to increase patient enrollment in trials. CMS has been trying to integrate all the databases to make data more available and easier to use. The panel also commented that cluster trials, while not trivial, are valuable for learning details that could not have been done with a traditional trial design. The NIH provided examples of their own successful collaborations such as Food and Drug Administration (FDA), Centers for Disease Control and Prevention (CDC), and ARHQ. DEcIDE (Developing Evidence to Inform Decisions about Effectiveness) is an AHRQ program that has been essential for a shared vision between NIH and AHRQ. It employs a coordinated approach and provides networking tools.
Dr. Collins commented that data access is critical for discovery. Several barriers such as academic conflicts and the informed consent process can slow down progress. The panel felt it was important for the HMORN to involve non-HMOs in this process to help fill gaps in patient populations. More analytic resources are needed; the HMORN has more data than can be analyzed. The panel expressed that there is a mismatch between where the researchers are and where the data are. Most researchers want to share data, however many are concerned about patient privacy, especially in regards to biospecimens, as well as data interpretability. The underlying data sources are typically collected for clinical care or financial reimbursement purposes, and therefore must be rendered useful for research. This data transformation process is often complex and nuanced, and studies benefit tremendously from having HMORN scientists involved rather than simply supplying data and subjecting it to misinterpretation. Patients are generally supportive of sharing their data for research, however, autonomy is a fundamental principle; they want to be asked, rather than told. Patients are also duly concerned about privacy and whether a profit is being made from their data. Data sharing in the context of the collaboratory would require thoughtful scrutiny and deliberation to maximize the value while respecting legal provisions that vary by site.
The interaction between research enterprise/data collection and quality improvement enterprise was addressed. The panel commented that joint planning groups take place to set priorities and forecast outcomes and quality improvements.
The HMORN leaders commented on effectiveness trials. It is important to examine which patterns of care are more beneficial than what is considered care as usual. The HMORN has the capacity to more efficiently identify logical targets for intervention for effectiveness trials. Using variation in practice and outcomes has been successfully demonstrated. However, it is very difficult to integrate health services interventions into this type of system. There is a consensus that there is a difference in quality of data from research studies versus routine health data from diagnostic studies in the clinic. These data need to be better integrated.
The NIH asked about the ease of participating in NIH studies and whether HMORN members work on NIH funded trials. The panel commented that the network looks at NIH proposals as they are submitted. Most collaborative trials are conducted in the institutes where the network PI has interest. The panel felt that more direct interaction and bidirectional communication between the HMORN and smaller NIH institutes would help foster collaboration among these groups.
The panel suggested that in order for the network to be strengthened, it would require some type of coordination function. It is not clear where this coordinating body would be located, what type of staff is required, or how it would be governed. Currently there is a rich network of centers, but no "headquarters"; and no clear answer for an optimal organizational model. However, one example comes from the Kaiser institutions, which have been conducting an experiment through The Center for Effectiveness and Safety Research (CESR). The Kaiser community benefit has provided all of the 8 Kaiser regions, which are not all HMORN members, resources to address these exact issues. A national director is being recruited that could be located in any of the Kaiser regions. However, they are still struggling with issues of governance and establishing administrative and investigator networks, a data-coordinating center, and a steering committee. It has taken a tremendous amount of time to prioritize opportunities, considering investigator capacity. The key question is how a coordinated data system across sites can be developed. For the Kaisers, this has taken substantial time and resources.
Discussion Session 2
The group addressed the next steps for creating an HMORN collaboratory with NIH. The funding required for such a network was discussed. The panel felt that the traditional NIH model for supporting a network was not appropriate in this case. There was a perceived advantage to having active participants that are embedded in the organizations and not staff that are located at a distant coordinating center. The key question is how this will become operational and high-functioning, given the distributed nature of the research. What could the formal structure look like? There are 2 steps that could be taken. One is for NIH to facilitate the HMORN to develop planning activities and another is to propose a funding model to support a coordinated entity.
One suggested approach for developing the coordinating center is a milestone driven model. It includes four primary levels of engagement:
- the legal and organizational nature of the entity, i.e. what does the network do versus the individual organization,
- a staffing model to make the collaboratory functional; this should correspond with the goals of the network,
- a central group to ensure progress at all 15 centers; NIH could be this central group or candidate sites could be proposed, an
- a funding model to sustain the organization.
The approach and resources needed to scale up the collaboratory need to be established. The NIH common fund could help. Overall, the panel felt it was best to first consider the short-term goals, and then establish what the collaboratory will look like in 5-10 years. The structure should not be centrally heavy, but should promote success with a work plan for broader partnerships, especially smaller NIH institutes. HMORN needs assistance in connecting with outside research networks, such as academics or NIH researchers where HMORN has no principal investigator who is interested in that research project. The hope is to develop the collaboratory as an NIH supported enterprise.