Matching into orthopaedic residency has become difficult, and the US Medical Licensing Examination Step 1 transition to pass/fail scoring has complicated the process. Advisors' ability to mentor students has decreased, and program directors may rely on Step 2 Clinical Knowledge (CK) scores in selecting which candidates to interview. This study aims to offer a method to predict Step 2 CK outcomes based on preadmission and preclinical performance. The study investigated 486 students from a US medical school who enrolled in 2017 and 2018. Data on demographics, preadmission, and preclinical performance were collected. Before model creation, it was found that sex, Medical College Admission Test scores, Comprehensive Basic Science Examination performance, and preclinical curriculum performance produced optimal models. Multivariate ordinal logistic regression models were built to predict probabilities of four outcome levels of Step 2 CK: <235, 235 to 249, 250 to 265, and >265. Finally, nomograms were created to visualize probability calculations. Each model's odds ratios revealed that female sex, higher MCAT scores, and better Comprehensive Basic Science Examination and preclinical performance were associated with an increased likelihood of being in higher Step 2 CK scoring groups. Preclinical performance had a profound effect, especially for those in the top 1/3. Models were successful in assigning higher probabilities to students in higher Step 2 CK scoring groups in more than 80% of instances. Nomograms presented provide examples of how to apply these models to an individual student. This study presents a novel method for predicting probabilities of Step 2 CK outcomes that can be used to mentor students at a time point when Step 1 previously filled this role. It may assist in identifying orthopaedic hopefuls at risk of performing poorly on Step 2 CK and can foster the development of individualized guidance and mitigation strategies.
Read full abstract