Journal of Osteopathic Medicine, 2025/10, 126(online 2025/10/01):213–219. doi: 10.1515/jom-2024-0157, type of study: retrospective study

Free full text   (https://www.degruyterbrill.com/document/doi/10.1515/jom-2024-0157/html)

Keywords:

exams [36]
medical students [693]
osteopathic medicine [2110]
predictive model [1]
retrospective study [339]
USA [1752]

Abstract:

Context As Comprehensive Osteopathic Medical Licensing Examination-USA (COMLEX-USA) Level 1 has changed to Pass/Fail scoring, residency programs that required minimum Level 1 scores for applicant consideration may choose to focus on COMLEX-USA Level 2-Cognitive Evaluation (Level 2-CE) target scores for applicant selection. Therefore, finding ways to predict passing and high Level 2-CE performance based on students’ past performance and to guide their study accordingly is essential for helping students succeed in and beyond medical school. Objectives The purpose of this retrospective study is to evaluate the predictive value of major performance measures from pre-admission to clerkship years on Level 2-CE. Then, based on the predictive value of those measures, the objective is to establish a predictive model and optimal cutoff scores with strong predictors to advise students on their preparation of Level 2-CE. Methods School-based performance measures for 948 first-time takers of the Level 2-CE Testing Cycles of 2019/20 to 2023/24 were analyzed. Correlational and multiple regression analyses were utilized to establish a predictive model utilizing: (1) preadmission and preclerkship performance (Medical College Admission Test [MCAT], undergraduate science grade point average [GPA], and preclerkship examination average); (2) national examination performance including the new COMLEX-USA Level 1 pass/fail-only status, individual and average clinical subject Comprehensive Osteopathic Medical Achievement Test (COMAT) scores, and the less studied Comprehensive Osteopathic Medical Self-Assessment Examination (COMSAE) Phase 2; and (3) clinical evaluation scores by preceptors. Then, receiver operating characteristic (ROC) curves were utilized to identify the optimal cutoff scores on the average clinical subject COMATs and COMSAE Phase 2 for student advising. Results A predictive model of COMLEX Level 2-CE was established with average clinical subject COMAT scores, first-time COMSAE Phase 2, preclerkship examination mean, and COMLEX Level 1 Pass/Fail status as the significant predictors. This model explained 73.9 % of the variance in Level 2-CE performance. Optimal cutoffs of the average clinical subject COMAT scores and first-time COMSAE Phase 2 performance were identified for passing Level 2-CE (COMSAE=447, average COMAT score=94.4) as well as having a high performance of Level 2-CE (650 & 700, respectively). Conclusions This study not only added evidence in support of previous studies on the bivariate associations between Level 2-CE and individual major performance measures from the preadmission to clerkship years, but also explored the use of the less-studied COMSAE Phase 2 in predicting Level 2-CE outcomes and provided a better predictive model utilizing the combination of individual performance measures. Most importantly, the current study also demonstrated a way to set optimal cutoff scores on strong predictors during the clerkship years to help the school guide students to better success on COMLEX Level 2-CE and long-term residency goals.