SMART on FHIR
The SMART system, or Substitutable Medical Apps and Reusable Technologies is a new health information technologies system that gives application developers the tools needed to build applications that directly utilize electronic health record (EHR) data. It has long been a goal of the health IT community to establish third party applications that harness EHR data akin to smartphone-based applications. Through a multitude of steps, including standardization of healthcare information exchange and SMART, this is now possible.
Smartphones, and the third party applications they run, have shifted the paradigm of healthcare delivery and use (1). This new type of healthcare, e-health has been defined as healthcare built upon the 10 e’s— efficient, evidence-based, ethics driven, equitable, empowering, educating, enabling informational exchange, expansive, and encouraging both consumers and patients (2). There has been a marked uptake in smartphones by both healthcare professionals and consumers (1). While e-health had wide ranging effects, one area that has been slow to embrace this trend is electronic health records. Healthcare, as an industry, has been slow to adopt an open standard which has thus stimmed innovation and an equivalent expansion of innovative tools. While mobile health has seen expansive growth with third party applications, and large consumer uptake—electronic health record (EHR) data has largely remained siloed (1, 3). Fortunately, this has begun to shift with the widespread adoption of Health Level Seven standards and an uptake in a modern approach based extensible specification known as Fast Healthcare Interoperability Resources (FHIR). With this adoption comes the desire for an integrated digital healthcare platform that can harness the power of EHRs. On top of FHIR, an application programming interface or API specification platform was created to simplify the process of building applications that can use EHR data. The subsequently named SMART platform helps bridge application developers and EHR data through FHIR. Though still in its infancy, multiple proof of concept applications have been created that have both shown technical feasibility and clinical uptake (3, 4).
With the rise of the consumer web, there has been a widespread adoption in APIs, or application programming interfaces. APIs represent a serviced oriented architecture and allow a set of services to be used by a third party for its own gain (5). These APIs allow deep integration within supported environments without reliance on project specific custom reverse engineering (6). This in turn allows substitutable applications that can still utilize EHR data. Uptake by EHR vendors has allowed a new level of clinical interoperability now allowing third parties to tightly integrate into the clinical data previously locked into vendor specific systems.
History of SMART on FHIR
In 2009, two members of the Computational Health Informatics Program at Boston Children’s Hospital and Department of Biomedical Informatics at Harvard developed an open API specification platform that leveraged FHIR. The soon to be called, SMART app platform or first bridged the gap between EHR data and an open health app platform (3).
The first iteration released in 2011, called SMART Classic, was clinically focused and built upon web standards and common healthcare terminologies such as Logical Observation Identifier Names and Codes or LOINC. This model was built upon a Resource Description Framework or RDF which represents a standard data interchange model on the web. RFD was chosen given its flexibility and adoption in prior web applications as a model to represent common clinical statements. From this, an API was developed to offer data access via hypertext transfer protocol or HTTP. While the developers were able to launch an initial product, there was criticism that the scope was limited and that standards groups were not involved.
In 2013, the SMART team pivoted. FHIR’s API provided a more robust version of the SMART API and was built with vendor and the greater healthcare IT community’s support. The team worked with the core FHIR team to help design resources, infrastructure, and tools. After multiple iterations the SMART team found FHIR to provide a robust methodology to deliver sematic interoperability while isolating developers from EHR specific extraneous details. Using prior knowledge from SMART Classic, the team adopted conventions from Meaningful Use Stage 2, and thus used terminologies such as LOINC for observations, SNOMED Clinical Terms for problems, and RxNorm for medications. Given the necessary security measures linked to healthcare data, SMART on FHIR built in OAuth token authorization. Similarly, authentication was handled via OpenID Connect to allow apps the necessary security measures to ensure proper usage. This SMART on FHIR starter kit allows close to plug and play capabilities to bring app developers EHR data in a simplified fashion. This allows both web-based applications via HTML5, desktop applications, and mobile applications—including Apple’s ios and Alphabet’s Android platforms.
- Mosa ASM, Yoo I, Sheets L. A Systematic Review of Healthcare Applications for Smartphones. BMC Medical Informatics and Decision Making. 2012;12(1):1-31.
- Eysenbach G. What is e-health? Journal of medical Internet research. 2001;3(2).
- Mandel JC, Kreda DA, Mandl KD, Kohane IS, Ramoni RB. SMART on FHIR: a standards-based, interoperable apps platform for electronic health records. Journal of the American Medical Informatics Association. 2016;23(5):899-908-.
- Bloomfield RA, Polo-Wood F, Mandel JC, Mandl KD. Opening the Duke electronic health record to apps: Implementing SMART on FHIR. International Journal of Medical Informatics. 2017;99:1-10-.
- Benson T, Grieve G. Principles of Health Interoperability. Third ed: Springer; 2016.
- Mandl K. Connected Apps in Healthcare 2017. 2017.
Submitted by Steven McGaughey