Methods: 260 candidates were grouped into two separate geographical groups - urban and suburban/rural. Descriptive analysis, skewness and kurtosis were performed for normality assessment, whereas Cronbach's alpha, McDonald's omega, and Greatest lower bound assessed internal consistency. For validity measures, correlations were calculated between scores for separate stations, overall scores, urban and suburban/rural status. Also, exploratory factor analysis was performed on the five stations as validity measures. Difficulty and discrimination indices were calculated as quality measures. Qualitative analysis was performed on "red flag" comments detailing grossly unsuitable candidates.
Results: Roleplayer-driven stations yielded more red flags than examiner-driven stations. The three examiner-driven stations were significantly and moderately correlated (rho between 0.602 and 0.609, p 0.530), whereas the stations were distributed equally in difficulty index.
Conclusion: The UMS MMI has identified specific skillsets that may be in short supply in our incoming medical students. Also, it illustrates the yawning gap between academic knowledge and 'translational' scientific knowledge and communication skills.