Association Between Admission Criteria and Written Summative Examination Performance Among Medical Students: A Retrospective Cohort Study

Abstract

Medical schools combine prior academic achievement, aptitude tests, interviews, and other measures to select applicants, but the predictive contribution of each component is context-dependent, and the study institution's own weighted admission composite had not been evaluated against its written examinations since its adoption. This retrospective cohort study evaluated the association between admission criteria and subsequent performance on the written summative examination in a Philippine medical program. The analytic cohort comprised 318 students admitted across eight cohorts from 2007 to 2014 who completed the academic curriculum through Year Level 3 by 2017, enrolled through total enumeration of all students meeting eligibility criteria. Data came from institutional admission records (undergraduate school and course, Latin honors, weighted point average [WPA], mental ability test, personality assessment, National Medical Admission Test component and percentile scores, interview score, and essay score) and registrar-recorded scores from 19 end-of-course written summative examinations. These measures were drawn directly from official institutional registrar and admissions records rather than newly collected instruments, and records were cross-checked for completeness before analysis. Frequencies, percentages, means, and ranges described the sample; Spearman correlation assessed categorical and ordinal admission variables, and simple and multiple linear regression assessed continuous admission variables against the Total Summative Score (TSS). The existing composite Total Admission Score (TAS) significantly predicted TSS: each one-point increase in the composite was associated with a 0.352-point increase in examination performance (95% CI 0.256–0.448; p<.001). In the multivariable model, WPA (B=0.505, 95% CI 0.314–0.696; p<.001) and Latin honors (B=1.324, 95% CI 0.299–2.349; p=.011) remained significant, while the aptitude, admission-test, and interview variables did not. Students with no remedial examinations had a higher mean WPA at admission than students with increasing numbers of remedials (88.7 versus 86.8, 84.9, and 83.9). Prior academic achievement was the most consistent predictor of written examination performance. The current composite showed statistically significant but limited predictive utility, and exact reweighting of admission components should not be implemented without reanalysis of the original data, internal validation, and evaluation against broader outcomes including clinical, professional, and community-based performance.