We present Clinical Asset Mapping Program for FHIR (CAMP FHIR), a Java application that transforms clinical data stored in a relational database (RDBMS) to HL7 Fast Healthcare Interoperability Resources (FHIR). CAMP FHIR aims to improve interoperability, standardization, and semantic harmonization by enabling transformation of data from any RDBMS, including but not limited to common data models (CDM), such as PCORnet, to FHIR with a single, source-agnostic tool. CAMP FHIR provides scripts that give CDMs relational FHIR views that the application maps and validates into valid FHIR specification. CAMP FHIR is a clinical research data pipeline that enhances the reusability of data by adding HL7 FHIR as a bidirectional interoperability layer.
Our team formed over the course of the application life cycle, with members of different skills and levels entering the project the University of North Carolina - Chapel Hill. The members form a development pod lead by James Champion as the primary software engineer with over 12 years of experience. His career have been focused on building and supporting a variety of applications in clinical research informatics. He has expertise in Perl, Java, JavaScript, PHP and SQL. Paul Kovach has expertise in clinical research informatics and software development. He has a Master's in Public Health from UNC-Chapel Hill. Asiyah Ahmad is a second year PhD student with a passion for programming and health care. Asiyah specializes in the continuous integration pipeline. Adam Lee is a 5th year PhD student in Health Informatics and serves the project in interoperability mappings. Anna Jojic is an experienced project manager with the Informatics and Data Science group who has worked on several complex data infrastructure projects; she has a Master’s degree in Library Science from the University of Illinois at Urbana-Champaign.
The team collaborates remotely each week and pursues enhancements that will promote adoption and use within the community.
Development of the Clinical Asset Mapping Program for FHIR (CAMP FHIR) began in 2018 with the goal of providing a Common Data Model (CDM) agnostic tool that leverages HL7’s Fast Health Interoperability Resources to enable data reuse and sharing between different institutions. In a multisite clinical research collaboration, institutions may or may not use the same CDM to store clinical data. Moreover, many community hospitals do not have research CDMs implemented at all. To overcome these challenges and allow sites with different CDMs to work together, we developed CAMP FHIR with a generic FHIR relational model provided that can accept any source clinical data model. This allows users to convert any relational database (RDB) to FHIR resources using CAMP FHIR as the intermediate application. This assists in data harmonization efforts across sites, supports data sharing opportunities between academic medical centers and community hospitals, and allows for easier integration between clinical and non-clinical datasets. For example, CAMP FHIR has been used to integrate clinical data with public exposure data and is used in the creation of data for other applications including the Clinical Annotation Research Kit and the FHIR Patient data Integration Tool.
Bidirectional support for converting FHIR resources into RDB is currently in development and would further open up the potential for conversions between CDMs, i.e. PCORnet to OMOP. To date, support has been provided for twelve of the most commonly used clinical research domains, with support for additional domains being added incrementally.
CAMP FHIR adheres to FAIR data principles and data reuse best practices. The application is hosted in a findable and accessible manner through a public facing GitHub page where all scripts needed to generate a relational FHIR schema are included, as well as documentation of the process. Using FHIR as our data reuse standard also promotes data interoperability and reusability in our project as the FHIR standard provides interoperability readily and base resources can be used as-is or adapted to fit the needs of individual organizations. We recommend that researchers adhere to these same principles when using and extending CAMP FHIR. CAMP FHIR is open source and is intended to be used across institutional and departmental boundaries. Therefore, adherence to FAIR principles is the only way to ensure that CAMP FHIR remains usable and interoperable between different stakeholders
Fast Health Interoperability Resources (FHIR) is a popular data exchange standard that represents data using a hierarchal schema in various formats (XML, JSON, and Turtle). CAMP FHIR is an open-source application developed primarily in Java. At the core of the application, we utilize two popular Java-based open-source platforms. The first platform we utilize is Hibernate, an open-source object relational mapping tool. The second is HAPI-FHIR, a Java-based platform that provides a complete implementation of the HL7 FHIR standard for healthcare interoperability within our application. The combination of these two technologies provides us with the necessary technical architecture to develop code that allows us to simply extract data from a relational database and generate an output in HL7 FHIR format.
To replicate this approach, users will need to prepare their source data and download the CAMP FHIR application from our Git repo into a location that can execute a Java application. In our CAMP FHIR Git repo, we also provide prepackaged CAMP FHIR scripts to create a transformation layer for their CDM, in the form of database views. These scripts can be used to execute the field and value set mappings that enable CAMP FHIR to generate FHIR resources from relational data. During transformation, CAMP FHIR uses the HAPI FHIR API to ensure that FHIR files output by CAMP FHIR are valid and properly conform to the R4 FHIR specification.
Too often research data extract, transformation, and loading is performed in an uncoordinated, inconsistent manner across studies, fragmenting and obfuscating data processes. Reusability of data is easiest and best when both data and the processing of data is transparent and unfettered by layers of transformation. CAMP FHIR, while performing transformations, keeps the data as raw and intentional as possible. Clear, timely and relevant documentation of field and value set mapping is provided, allowing others to easily consume data for reuse. The overall benefit to data reuse is to garner additional insights and explore new perspectives of data outside its original intention or use. CAMP FHIR empowers the community to reuse data from various data sources by offering transparent translation from well-documented, pre-existing data sources to a well-documented, transport standard. This translation is key to reducing and removing barriers for researchers needing data in a non-native syntax, or to exploring methodologies on data in a non-relational format. Creating an open-source research data pipeline that leverages other highly adopted technologies and standards opens the door for community engagement.
