This is one school’s experience with NBME assessments. Results or outcomes may vary and depend on factors beyond NBME’s control, such as student participation, individual study habits, faculty support and engagement, and/or variations in curriculum, among other things.
Indiana University School of Medicine (IUSM) wanted to give students earlier, data-informed insight into their readiness for the USMLE® Step 2 Clinical Knowledge (CK) exam. To address this challenge, leaders from IUSM's Business Intelligence and Medical Student Education teams collaborated to identify the factors most strongly associated with Step 2 CK performance for their students. After evaluating hundreds of student data points, they developed a predictive model that incorporates MCAT scores, performance in select Phase 1 courses, and results from certain NBME® Subject Exams to generate a predicted score. To support meaningful use of that score, IUSM Student Affairs developed a Canvas-based resource for their students that explains the model, delivers individualized results directly to students and provides advising recommendations.
An important challenge was determining how to incorporate Subject Exam scores into the model when students complete clerkships and subject examinations in different sequences. IUSM needed an approach that could generate meaningful predictions at different points in the curriculum. The team’s model addressed this concern, allowing for predicted score as soon as a single Subject Exam was complete. Methods of their model were published in the Journal of Medical Education and Curricular Development (link provided below).
The IUSM Predictor also highlighted a broader opportunity: making assessment data easier to access and use across advising and analytics workflows. IUSM participated in an early pilot of the NBME Assessment Data Application Programming Interface (API), which will allow schools to securely integrate NBME assessment data directly into their own reporting and analytics systems and reduce the need for manual downloads. Through the pilot, Indiana University explored opportunities to automate manual data workflows and deliver more timely information to advisors and educators. The initiative supports the school's broader efforts to streamline assessment workflows, improve access to performance data, and enhance advising programs such as the IUSM Predictor.
I felt like I had no grasp on where I stood in terms of being competitive as a residency applicant, and the estimator was my first guess that I might be on the right track… it allowed me to feel more confident in my decision to apply to a more competitive specialty.
IUSM Student
A Step 2 predictor combined with individualized guidance and advising can play an important role in supporting student success.
Anthony Shanks, MD, MEd, Senior Associate Dean of Medical Education
Our school was excited to partner with the NBME in developing its API. This new tool will simplify access to the data that is critical for advising and mentoring our students.
Ben Steckler, Director of Business Intelligence, Indiana University School of Medicine
NBME Portfolio Parts Used
IUSM incorporates NBME Clinical Science Subject Exams during Phase 2 (year 3) of their curriculum. Results from Emergency Medicine, Family Medicine, Internal Medicine, Neurology, Obstetrics and Gynecology, Pediatrics, Psychiatry, and Surgery were used in their Step 2 CK predictive model.
Supporting Data
- Shanks A, Steckler B, Smith S, et al.: Impact of a USMLE Step 2 prediction model on medical student motivations. J Med Educ Curric Dev. 2025, 12:1-7. 10.1177/23821205251321812.
- Bird JB, Olvet DM, Willey JM, Brenner JM. A Generalizable Approach to Predicting Performance on USMLE Step 2 CK. Adv Med Educ Pract. 2022;13:939-944 https://doi.org/10.2147/AMEP.S373300
- Kracaw R A, Dizon W, Antonio S, et al. (September 20, 2021) Predicting United States Medical Licensing Examination Step 2 Clinical Knowledge Scores from Previous Academic Performance Measures within a Longitudinal Interleaved Curriculum. Cureus 13(9): e18143. doi:10.7759/cureus.18143
More to Discover
Join our educator community, explore assessment research fellowship opportunities, and see how your peers are innovating assessment best practices.
