This article is part of a series about obstacles in academic research. In this article we're discussing the challenges of complex studies or multi-site studies, particularly when carried out with a single database tool such as REDCap.
Why study complexity needs adequate planning and setup
Data-rich studies require a fully relational database. While REDCap is an excellent data collection tool for studies with simpler data models, it doesn’t offer a fully-relational platform that can link instruments and tables to meet any degree of complexity.
Consider a study in which participants are tracked through (potentially multiple) emergency room visits over a five-year period, with an arbitrary number of labs taken during each visit. This example, simple as it is, involves a one-to-many-to-many relationship, as shown below.
An effective study database would allow users to select a participant from a list, select an ER visit record related to that participant, and then enter labs specifically related to that visit.
Unfortunately, REDCap can't manage the relational structure required for this model. Entering such data into REDCap demands awkward work-arounds, which are difficult to use and invite errors.
REDCap has similar limitations for many-to-many relationships. Consider a study that offers various ongoing support groups as an intervention for behavioral health outcomes. One would expect the study database to allow users to enter as many groups as needed, then record participants' attendance in individual group sessions. Entering and managing this kind of data is easy with a fully relational database, but there's no effective way to model it in REDCap.
Real Life Example:
The limits of REDCap were instrumental in a study on traumatic brain injuries in college athletes, conducted by the NCAA and DoD.
A subject might sustain multiple injuries during the study, with each injury involving multiple outcomes and instruments. Although the relationships in this data model were not especially complex, the researchers found they could not manage them adequately in REDCap. They searched for a more flexible database option, which they found in QuesGen.
Depending on your data model, REDCap may be perfectly adequate for your study's needs. Or it may present obstacles to good data entry and unpleasant surprises at the end of the study, with data that's difficult or impossible to analyze as expected.
If your institution has questions about its studies, data models, or data management tools, we encourage you to reach out to our team to talk further.
But that's not the only obstacle . . .
To learn more about the next obstacles in clinical research, keep reading: Overcoming Academic Research Obstacles: Custom Integrations with EHRs