The Last Mile of Interoperability: Integrating Outside Data Into Clinical Workflows to Improve Care - PROJECT SUMMARY Building a nationally interoperable health care delivery system remains a top priority in US health care, with early studies theorizing that robust data exchange could reduce health care costs by over $80 billion annually. Despite recent efforts to enhance interoperable health information exchange, there is mixed or insufficient evidence that increased connectivity between health care delivery organizations is improving outcomes or reducing costs. One potential reason for this is the “last mile” of interoperability; busy clinicians lack the time to search for, view, and process patient data delivered via interoperability, which is often not integrated into the local electronic health record (EHR). Delivering on the promise of interoperability requires a better understanding how clinicians use data generated outside their institution, and the value of that data for improving outcomes such as utilization, quality, and patient satisfaction. Finally, developing actionable guidelines for when and where outside records data (ORD) is valuable within the context of a given encounter is critical to building interoperability solutions that facilitate high-value use by clinicians. To address this, our study – The Last Mile of Interoperability: Integrating Outside Data Into Clinical Workflows to Improve Care – will leverage a natural experiment and detailed EHR metadata from two institutions to identify the causal impact of HIE use on a range of patient outcomes including costs and utilization, health, and patient satisfaction. We will use an innovative combination of robust econometric instrumental variables to identify the impact of ORD review and novel machine learning methods to reveal the clinical scenarios where HIE use is likely to be valuable within the context of a specific clinical encounter. We will use these results to generate public goods that will guide the development and evaluation of tools designed to streamline clinician information consumption and optimize the review of outside records data with a focus on new artificial intelligence and large language model applications. Our study will generate actionable evidence on the impact of outside record integration, the value of viewing outside records data by clinicians, and inform the development and evaluation of new technologies towards the goal of realizing the benefits of the national investment in interoperability.
