The goal of this study was to add a predictive modeling approach to the meta-analysis of continuing medical education curricula to determine whether this technique can be used to better understand clinical decision making. Using the education of rheumatologists on rheumatoid arthritis management as a model, this study demonstrates how the combined methodology has the ability to not only characterize learning gaps but also identify those proficiency areas that have the greatest impact on clinical behavior.Methods:
The meta-analysis included seven curricula with 25 activities. Learners who identified as rheumatologists were evaluated across multiple learning domains, using a uniform methodology to characterize learning gains and gaps. A performance composite variable (called the treatment individualization and optimization score) was then established as a target upon which predictive analytics were conducted.Results:
Significant predictors of the target included items related to the knowledge of rheumatologists and confidence concerning 1) treatment guidelines and 2) tests that measure disease activity. In addition, a striking demographic predictor related to geographic practice setting was also identified.Discussion:
The results demonstrate the power of advanced analytics to identify key predictors that influence clinical behaviors. Furthermore, the ability to provide an expected magnitude of change if these predictors are addressed has the potential to substantially refine educational priorities to those drivers that, if targeted, will most effectively overcome clinical barriers and lead to the greatest success in achieving treatment goals.