Predictive Validity of Multiple Mini Interviews: A Multi-Site, Multi-cohort Study

Background: The Multiple Mini Interview (MMI) is used internationally as a selection tool for medical school admissions. The MMI is a series of short, one-on-one interviews that assess such attributes as communication, problem-solving, and teamwork skills [1,2]. This study investigated the predictive validity of the MMI for the following outcome measures: medical school performance (Educational Performance Measure [EPM], Situational Judgement Test [SJT], Prescribing Safety Assessment [PSA]) and passing professional membership exams in medicine (RCGP, MRCP, MRCS). The study included data from two medical schools in the UK.
Methods: Data from 4990 doctors who graduated from UK medical schools and sat the first part of professional membership exams in 2017-2019 was used. The UK Medical Education Database[3] provided linked data from different sources, including medical school admissions, assessments, and postgraduate training. Multinomial logistic regression analyses estimated the odds of passing college membership exam on first attempt.
Results and Conclusion: MMI was a significant predictor of medical school performance, even after controlling for other factors such as high school grades (UCAS scores) and clinical aptitude test (UCAT scores). The MMI was also a significant predictor of passing college exams on first attempt, but the effect size was smaller than for those assessments that occur nearer to postgraduate training - EPM, SJT, and PSA scores. The findings suggest that the MMI is a valid predictor of educational success in both medical school and postgraduate training. Although the proportion of variance explained by MMI and all other predictors is small, MMI remains a valuable tool for medical school admissions. In the absence of innovations that can improve prediction, medical schools should continue using MMI in combination with other factors, such as UCAS and UCAT scores, to make admissions decisions.