REDCap (Research Electronic Data Capture) is a browser-based, metadata-driven EDC software solution and workflow methodology for designing clinical and translational research databases.[2] It is widely used in the academic research community: the REDCap Consortium is a collaborative, international network of more than 1700 institutional partners in over 95 countries, with more than 300,000 total end-users employing the software for more than 200,000 ongoing research studies.[3]

History

REDCap was developed by an informatics core at Vanderbilt University with ongoing support from NCRR and NIH grants. REDCap was designed to address common problems for academic biomedical researchers hoping to use electronic databases. First, major vendor EDC and CDMS solutions are targeted and priced for large clinical trials, and can be prohibitively expensive for investigator-initiated studies or other such studies at a smaller scale.[4] Secondly, the independent research environment often lacks the informatics and other multidisciplinary support necessary for effective IT integration into research protocols.[5] The REDCap software as distributed through the REDCap consortium attempts to facilitate informatics support for clinical researchers and foster a collaborative network of institutional researchers who share and support REDCap as a common research tool.[2]

Software licensing and intended use

Although REDCap is available at no charge to institutional partners – discounting the cost of internal IT support staffing – REDCap is expressly not open-source software. Certain end-user license agreements distinguish it from a typical open-source license. Namely, the software is restricted in use, permitted only for non-commercial research purposes. REDCap is also restricted in redistribution because Vanderbilt is the only entity that can distribute it. Furthermore, any and all derived works – such as innovations or programmatic features added on by the user – are essentially owned by Vanderbilt. Vanderbilt catalogues such derived works in their REDCap Consortium library, which is available to all consortium members. The REDCap End-User License Agreement[6] also encompasses control by Vanderbilt over its licensees' publications on or about REDCap, specifying that Vanderbilt shall coordinate and have editorial control over any "publications created by CONSORTIUM MEMBERS which discuss the SOFTWARE and its methodologies, functionality, and/or abilities." Publications that describe scientific studies which have utilized REDCap are excepted from these editorial restrictions.[7]

The REDCap software is distributed from Vanderbilt to institutional consortium partners, who in turn give research teams access to REDCap. REDCap project design has an intended workflow outlined by the developers.[2] Upon request, the informatics core gives the research team a demonstration of REDCap, highlighting the most relevant user interface features. The researchers then fill out a Microsoft Excel spreadsheet file with key metadata (i.e. field name, data-type, data range, etc.) about each measurement in their CRF. The informatics core converts this spreadsheet template into study-specific database tables which are linked to a working web-based EDC forms and applications environment. The researchers then test this prototype web application by filling out dummy data, and the metadata spreadsheet is revised and refined in an iterative process. Once the REDCap project design is finalized, the application is deployed from development to production mode, all dummy data is lost, and researchers begin committing actual patient data.

The REDCap design workflow has important limitations. When a project is deployed into production mode, further revisions in database design need to be approved by the informatics core. Furthermore, certain production mode revisions are simply not allowed. For instance, when applying the REDCap calendar tool, production mode revisions to calendar metadata is prohibited, so researchers are advised to be very careful in formulating metadata and event-handlers before committing to the inflexibility of production mode.[8]