More effective use of the vast quantities of health data generated every day is central to most conversations about how to improve patient care and outcomes. Beyond traditional research models, the thoughtful use of very large clinical datasets could support richer, far more detailed analyses. These would allow us to begin to study treatment effectiveness at the level of the individual and to provide more nuanced information to all major stakeholders, including individual patients who must make critical, personal health care decisions.
The challenge is how to effectively link the ever-expanding volume of structured and unstructured data in electronic health records (EHRs), insurance claims databases, patient registries, mobile devices, and other “silos” in ways that make them available for research and clinical care while protecting privacy and other legitimate interests of patients, providers, payers, and others who may own or manage such information.
The potential of harnessing these data for performance improvement, surveillance, and research has been apparent to researchers, clinicians, and policymakers for years. The efforts of the Patient-Centered Outcomes Research Institute (PCORI) in this area build on those of others, including the Food and Drug Administration (FDA), the National Institutes of Health, pharmaceutical and device manufacturers, and payer organizations.
PCORI’s efforts, manifest in PCORnet, the National Patient-Centered Clinical Research Network, are innovative in convening patients, health systems, clinicians, and payers as active participants in the enterprise, along with researchers and potential funders. It is an approach emblematic of the patient-centered and stakeholder-guided comparative clinical effectiveness research (CER) that PCORI was created to support.
The PCORnet Model
We launched PCORnet, a large, representative, national data resource, nearly two years ago. By harnessing the power of clinical data gathered in real-world settings, PCORnet is designed to allow the nation to conduct a range of comparative research studies more efficiently and less expensively.
PCORnet is made up of 11 Clinical Data Research Networks (CDRNs), which are based in health care systems, and 18 Patient-Powered Research Networks (PPRNs), which are operated and governed by patient-led groups. This set-up reflects much of the diversity of the repositories within which real-world data are found.
PCORnet is governed by a Steering Committee, comprised of the principal investigators from each network and the Coordinating Center that provides technical and logistical support for the initiative, along with representatives from PCORI, the patient community, federal agencies, and private industry. Additional task forces and working groups are organized according to subject matter expertise; these groups develop processes and products for the overall network.
One major challenge experienced by PCORI and others working in this area is that EHRs and other repositories of clinical data don’t follow global standards for how data are collected, stored, and transmitted. Moreover, individual systems and settings, such as hospitals and physician practices, enter data into and use their information systems differently to meet particular needs and interests.
PCORI doesn’t have the authority or leverage to impose data standards on EHR vendors, health systems, or physicians. So, instead, we meet the data where they are by using a Common Data Model (CDM). Nevertheless, a national movement toward greater agreement on and implementation of data standards for capture, storage, and exchange would not only benefit clinical care but also make secondary uses of clinical data for large-scale clinical research more efficient and secure.
Until then, PCORnet addresses the challenge of working with disparate data embedded in a variety of systems and locales by requiring that network members transform their local source data, along with related information (e.g., the standard used locally), into a common data structure. This CDM allows for a distributed approach to data analysis in which individual-level data remains in place but can be used by researchers across the network while maintaining local control, privacy, and proprietary interests. This distributed model is based on other major efforts, including the FDA’s Sentinel project, a well-established active surveillance system of FDA-regulated medical products with two dozen national collaborators.
Testing The Concept
We have planned a series of demonstration studies to test PCORnet’s technical capacity and functions. The first, approved by our Board of Governors in May 2015, is a randomized trial studying the optimal dose of daily aspirin needed to lower the risk of heart attack and stroke in patients with heart disease.
Subsequent demonstration projects will include observational studies comparing the effectiveness of different bariatric surgical procedures and the risk of childhood weight gain associated with various antibiotics given between infancy and age two. Another project will bring leaders of health systems together with CDRN informaticians and clinical researchers to explore PCORnet’s potential to address issues of critical interest to delivery systems.
These demonstration studies will evaluate many aspects of PCORnet’s function, from use of its CDM across multiple networks to organizational approval of oversight by central Institutional Review Boards, reduced times for contracting and subcontracting among PCORnet sites, and new and more efficient approaches to patient recruitment in large, simple clinical trials. We expect that PCORnet will be ready to engage, and function efficiently, in additional research studies by late 2015.
That 29 individual networks could reach agreement on a common data model reflects the vested interests they have in collaboration, efficiency in data collection and sharing, and building a reputation for conducting high-quality studies. They, and we, also recognize that patients, clinicians, health systems, and insurers all have a stake in making the data they own or manage more accessible for research, but that their time and resources are too limited and valuable to embrace this effort unless it is efficient and returns value.
Putting Patients At The Center
Beyond linking information, any successful health data research network must engender the trust, support, and active involvement of the patients whose data are being used, even with rigorous and appropriate privacy and security safeguards in place. As a central tenet of PCORnet’s evolving work, it requires that the patients, clinicians, and health care systems that provide the research data housed in each constituent network be actively involved in governance and use of that information.
There is growing evidence that patients and their advocates support use of their data for research when they feel they’ve been informed about the nature of the research and understand how their personally identifiable information is protected. We believe that comfort level is enhanced by their understanding that traditional research, for all of the advances it has produced, often doesn’t study many of the questions most important to them so that health care decisions they and their clinicians make are as well-informed as they could be.
This type of engagement is an extension of the same principle that drives all of our work at PCORI. We ensure that patients and other health care stakeholders are actively engaged in guiding what we do and are meaningful partners in all aspects of the CER projects we fund.
It’s an approach that we hope will ensure PCORnet’s sustainability, integrity, utility, and ability to generate the kind of information that improves health care and patient outcomes.
No comments:
Post a Comment