Use of Learning Management System Data to Predict Student Success in a Pharmacy Capstone Course

Objective. Learning management system (LMS) data from online classes may provide opportunities to identify students at risk of failure. Previous LMS studies have not addressed the possibility of change in student engagement over time. The purpose of this study was to apply a novel statistical technique, group-based trajectory modeling (GBTM) to LMS data in an online course to identify predictors of successful course completion.Methods. Exploratory GBTM was used to assess the association of LMS activity (total activity time, dates of activity, and pages viewed) and attendance at virtual synchronous learning sessions with examination performance in a capstone disease-management course delivered in the final didactic quarter of a three-year Doctor of Pharmacy program. Groups were assigned based on trajectories of weekly page view counts using structural-equation modeling.Results. Group-based trajectory modeling identified three page view engagement groups (median total page views, n): group 1, high (1,818, n=24): group 2, moderate (1,029, n=74), and group 3, low (441 views, n=35). Group assignment alone was somewhat associated with final grade. Stratification based on consistent virtual synchronous learning session attendance improved predictive accuracy; for example, a top (A or A-) grade was earned by 49.0% and 24.0%, respectively, of group 2 students with and without consistent synchronous engagement.Conclusion. Application of GBTM to LMS data, including information about synchronous engagement, could provide data that allow educators to identify early warning signs that a student may fail a course and target interventions to those at-risk students. The technique should be further tested with alternative LMS data and obtained early in the didactic curriculum, before patterns of engagement are established.