Clinical, technical, and implementation characteristics of real-world health applications using FHIR

Lay Summary The Fast Healthcare Interoperability Resources (FHIR) standard is increasingly being adopted for healthcare data exchange in the United States and globally, but an overall picture of the applications (apps) that implement the standard is not well known. This study aimed to provide a preliminary assessment of real-world FHIR apps' clinical, technical, and implementation characteristics. To achieve this aim, we (1) searched digital repositories and libraries containing FHIR apps, (2) extracted the app name, source, and developer contact information, and (3) collected information about the apps via an electronic survey of FHIR app developers and implementers. Respondents reported a variety of health domains, terminologies, FHIR specifications, and types of developing organizations for the 112 apps about which we received information. Apps were increasingly introduced from 2012 to 2022, and most focused on clinical care (n = 74) or research (n = 45). The majority were stand-alone web-based (n = 67) or embedded within an electronic health record (n = 51). Of the many health domains covered, cardiovascular care was the most common (n = 28). Half (n = 56) were implemented across multiple health settings, highlighting FHIR's substitutable nature. Overall, this study demonstrates the momentum around FHIR, the diversity of FHIR apps, and the need for a robust global FHIR app repository. Objective Understanding the current state of real-world Fast Healthcare Interoperability Resources (FHIR) applications (apps) will benefit biomedical research and clinical care and facilitate advancement of the standard. This study aimed to provide a preliminary assessment of these apps' clinical, technical, and implementation characteristics. Materials and Methods We searched public repositories for potentially eligible FHIR apps and surveyed app implementers and other stakeholders. Results Of the 112 apps surveyed, most focused on clinical care (74) or research (45); were implemented across multiple sites (56); and used SMART-on-FHIR (55) and FHIR version R4 (69). Apps were primarily stand-alone web-based (67) or electronic health record (EHR)-embedded (51), although 49 were not listed in an EHR app gallery. Discussion Though limited in scope, our results show FHIR apps encompass various domains and characteristics. Conclusion As FHIR use expands, this study-one of the first to characterize FHIR apps at large-highlights the need for systematic, comprehensive methods to assess their characteristics.