New investments in clinical programs at Vanderbilt University Medical Center increasingly are contingent upon studies to validate claims of benefit and the potential for scalability and spread.
To support this pragmatic approach, VUMC is leveraging translational, multidisciplinary team science to support clinical and operational decision-making across the health system.
The principles and practices involved in this strategy, and lessons learned along the way at VUMC, are set out by Christopher Lindsell, PhD, and colleagues in a paper in Academic Medicine, “Learning from What We Do, and Doing What We Learn: a Learning Health Care System in Action.”
Authors of the paper note that the learning health care system (LHS) concept has gained currency in recent years and has been interpreted differently at different centers. One approach taken at VUMC has been to embed pragmatic clinical trials into routine processes of care to generate evidence about the effectiveness of different therapies. To support this work, an LHS platform has been developed.
“Our LHS platform at VUMC embraces all of our missions. In creating and disseminating generalizable knowledge that improves what care is provided and how care is provided, a true team science approach is needed. By partnering our research, clinical and learning experts, we are able to evaluate innovations and their real-world effects at scale and in a way that will inform clinical practice everywhere,” said Lindsell, professor of Biostatistics and director of the methods program at the Vanderbilt Institute for Clinical and Translational Research (VICTR), where centralized resources for the LHS platform are housed.
The paper lists eight pragmatic effectiveness trials already completed in the platform, among them safe removal of penicillin allergy labels in the ICU and early introduction of palliative care for patients with advanced liver disease. According to Lindsell, more than 100 projects have been evaluated by the LHS platform. As described in the paper, key elements of the platform include:
- Engagement of patient and family representatives, front-line health care providers and clinical and operational leaders;
- Weekly workshops where proposed projects are explored, with representatives from affected clinical areas assisting project design, ensuring that pragmatic research activities fit manageably into clinical workflow;
- Use of randomization to reduce bias, with outcomes data readily available from routine electronic health records and other information systems.
Support offered through the platform includes project management; biostatistics for study design and analysis; clinical informatics for data collection and automated clinical decision support; and regulatory and bioethics guidance. The governance structure for the LHS platform combines research leadership with health system operations and administrative leadership.
As outlined in the paper, among the next steps for the LHS platform are a new focus on implementation and dissemination of study findings and measuring the cost-effectiveness of clinical interventions.
In a related paper, “Embedding Learning in a Learning Health Care System to Improve Clinical Practice,” also in Academic Medicine, Matthew McEvoy, MD, professor of Anesthesiology, and colleagues describe how they’re testing the effectiveness of a smartphone application developed at VUMC, QuizTime, for supporting dissemination and uptake by clinicians of evidence generated through the LHS platform.
As described in the paper, QuizTime supports practice change by sending clinicians who opt in to a multiple-choice question each workday, with the modality (text or email) and time of delivery determined by the individual clinician; clinicians receive immediate feedback about the topic, including whether their answer was correct, a summary of the implications for practice and a longer rationale for the topic being tested.
In all, 17 authors from VUMC joined Lindsell and McEvoy in these two papers. The LHS platform is supported by the National Institutes of Health (TR002243).
