Predictive Validity of Pre-Clinical Academic Achievements in Comprehensive Basic Science Examination: A Nationwide...

Medical education directly impacts patient care, yet the predictive validity of pre-clinical academic performance for licensure exam outcomes remains debated. This national, multi-institutional study (2019-2021) assessed the relationship between university course grades, cumulative grade point average (GPA), and Comprehensive Basic Science Examination (CBSE) scores in Iranian medical students.

Course grades and GPAs of 23 medical schools were linked to CBSE outcomes of 51 medical schools across five consecutive exam periods via student national ID. Pearson's correlation, paired t-tests, ANOVA, and chi-square assessed trends. Hierarchical cluster analysis (dendrogram) examined course grade correlations. Independent CBSE total score predictors were found using multiple linear regression.

Of the 25,757 individual records, 9,359 (45.2% female) had complete academic and CBSE data, making them eligible for primary analyses (84.5% passed CBSE on the first attempt). The GPA was 15.11±1.74, and the CBSE score was 101.68±24.61. All course grades correlated significantly with CBSE subtests (r=0.055-0.544, P<0.001). A significant moderate association (r=0.492, P<0.001) exists between overall GPA and CBSE. Repeat examinees had considerably lower GPAs and CBSE scores (P<0.001). GPA (β=0.318), Anatomy (β=0.158), Physiology (β=0.135), Epidemiology (β=0.043), and Virology (β=0.043) were the most significant predictors in regression modeling (R²=0.426). Cluster analysis showed that academic grades in anatomy, physiology, and biochemistry were strongly correlated with CBSE subtests.

This study represents the first large-scale national dataset in Iran pertaining to medical education. Pre-clinical GPA and course grades exhibit overall and subject-specific, notable predictive validity for CBSE performance. To enhance medical education and licensure results, it is advisable to implement standardized, cross-institutional comparisons alongside dynamic curriculum reviews. The regression model and clustering insights provide a framework for targeted educational interventions.