A Reusable Framework for Automated Record Creation and Population ARC: Beyond the API A Reusable Framework for Automated Record Creation and Population Liz Moyer, BA; Erik Benton, BA; Alison Grossblatt-Wait, MA; Christiane Winter, BS; Matthew Lawhead, BS; Timothy Kilgore, BA; James Scott, MBA; Julie Mitchell, BS; Robert Schuff, MS Oregon Clinical & Translational Research Institute, Oregon Health & Science University Introduction Projects that prospectively assemble cohorts from external systems, such as research data warehouses, require a cost-effective way to streamline the creation and population of records in REDCap. Use Case A study on reproductive health required identification of a cohort of women in our research data warehouse, which could then be imported into REDCap. The initial cohort count was over 15,000 and approximately 100 new patients would be added monthly. Challenges with existing methods Manual import and refresh of the patient data set would be cumbersome and error prone. DDP provides a way to update data but doesn't handle new record creation. The API supports record creation but requires custom development for each project. Objectives Automate a recurring process for new record creation and population without user intervention. Transform and encapsulate into a re-usable web service the existing REDCap processes to import data. Integrate regulatory checks and codify the process and workflow for setup and management of the service. Methods ARC (Automated Record Creation) is a feature of our DDP service, enabling the initial creation of records by ARC, followed by record population via DDP. The workflows for ARC setup and run are: Discussion ARC shares much of the same code with DDP for regulatory checks and general architectural principles. ARC can be used with both classic and longitudinal projects. For longitudinal projects, the unique REDCap event name is embedded in the cohort query and returned with the results. Limitations For performance considerations, ARC is not currently recommended for use with datasets greater than 20,000. Every time ARC is run for a project, the full dataset is loaded, as opposed to importing only new records. Future Opportunities Packaging ARC and developing a plug-in, coupled with some enhancements listed below, would position the service for dissemination within the consortium. Notify project user(s) regarding the status of ARC runs. Add mechanism for users to create explicit variable mapping between REDCap and the data repository. Improve performance by adding incremental loading. Generalize the interface for connecting to the data repository. Setup 1 Build patient cohort query 2 Collect project info to run ARC 3 Add ARC API user to REDCap project 4 Build data collection form(s) Run 5 Scheduled process invokes ARC 6 Run compliance check 7 Request patient data 8 Import data into research project Set Up The cohort query returns unique IDs for each record for ARC to use at import. A separate “Admin” project stores data about each REDCap project using ARC: REDCap project ID IRB number API token for ARC user Name of cohort query associated with project The user will need API import rights and an API token (ARC uses API to create new records). REDCap project variables must exactly match cohort query field names. 1 2 3 4 Run ARC can be started with a cron job or other authenticated web service. Using the API, ARC exports project details from the Admin project and sends a query to check IRB status. If the IRB check passes, ARC executes the project’s cohort query to acquire data. ARC returns data as paged results, which are imported into the research project via the API. 5 6 7 8 Acknowledgements The work reported in this poster was supported by National Center for Advancing Translational Sciences of the National Institutes of Health under award number UL1TR000128. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.