Data Analysis

All statistical analyses for the quantitative data were performed using Stata Software Package, version 14.2 (StataCorp LLC, Texas, USA). Significance was defined as having a p-value of < 0.05, all p-values were double sided, and 95% confidence intervals were computed.

For the first study arm, the results of the MCQ pre-, mid-, and post-rotation quizzes were compared in a one-way repeated measure analysis of variance (ANOVA) to understand whether there was a difference in test scores amongst students before, during, and after the innovative clerkship curriculum. In addition, the mean scores, and standard errors were computed for each group’s MCQ score. Tukey’s HSD post hoc methodology was utilized to indicate where the significant effects occurred between the three test groups.

To compare student performance in the innovation group to the previous class, exit exam scores were evaluated using a one-way ANOVA with Group (Pre-intervention clerks, Post-intervention clerks) as the only factor. These analyses determined whether there was a significant difference in mean exit exam scores between the two groups. Summary statistics of the exit exam scores for both groups were also estimated, providing mean scores and standard errors.

In order to determine and understand the perceptions of the new clerkship curriculum, semi-structured interviews were conducted with both students and maestros. Interviews were analyzed using an inductive content analysis approach [25]. The intent of this analysis was to develop a rich description of learner and maestro views, which offers both a summative and an integrative element across participants. Initial coding began by analysts coding independently, then meeting to discuss and create a refined coding schema, (Appendix 2) which was then applied to all interviews. N-Vivo 12 was used to manage qualitative data analysis.