Correlation of American Board of Surgery in-Training Exam Metrics in a Postpercentile Reporting Era

The annual ABSITE exam is eliminating percentile reporting beginning in 2025. This study aims to develop and validate predictive targets for standard scores and percent correct to effectively identify residents at risk of scoring below the 30th percentile, as well as provide score context for residents.

Retrospective data (program year, percentile, percent correct, and standard score) were collected. Linear regression analysis was used to determine standard score and percent correct targets for each program year. Target scores were compared to predict performance below the 30th percentile for each program year.

Community-based general surgery residency program (Henry Ford Jackson, MI) PARTICIPANTS: All ABSITE exams taken by residents between 2012 and 2024. Data includes 148 exams taken by 42 residents collected anonymously.

Standard score targets were set at 350, 445, 210, 535, 545 for program year 1 through 5, respectively. These standard score targets had an overall sensitivity of 0.87 and specificity of 0.93. Percent correct targets were set at 55%, 64%, 70%, 72%, 73% for program year 1 through 5, respectively. These percent correct targets had an overall sensitivity of 0.87 and specificity of 0.99.

The predicted targets for standard score and percent correct were effective in identifying resident performance under the 30th percentile, an evidence-based cutoff for identifying struggling learners. Further investigation is needed to validate these results.