Prospective Prediction of Match Outcomes in Integrated Plastic Surgery: A Novel, Mixed-Methods Approach Incorporating...

Area of Science:

• Medical Education
• Graduate Medical Education
• Surgical Residency Admissions

Background:

• Integrated plastic surgery residency applications are highly competitive.
• Previous research focused on objective metrics, neglecting qualitative aspects of recommendation letters.
• The role of letter of recommendation (LOR) content and authorship in predicting match outcomes remains understudied.

Purpose of the Study:

• To develop and validate a predictive model for integrated plastic surgery residency match outcomes.
• The model integrates quantitative and qualitative application data, including LOR content and authorship characteristics.
• To enhance the understanding of factors influencing successful residency applications.

Main Methods:

• Retrospective analysis of US MD applicants to integrated plastic surgery programs (2023-2025).
• Data included demographics, academic metrics, research, LOR content (linguistic analysis via LIWC-22), and LOR author characteristics (H-index).
• A machine-learning model (LASSO logistic regression) was developed and prospectively validated, with performance measured by AUC.

Main Results:

• Higher Step 2 scores, increased podium presentations, higher letter writer H-index, and "communal/likable" language in LORs predicted successful matches.
• A standardized LOR ranking between 5 and 10 was a negative predictor.
• Prospective validation achieved an Area Under the Curve (AUC) of 0.7791.

Conclusions:

• A machine-learning model effectively predicts integrated plastic surgery residency match success.
• Key predictors include academic/research achievements, LOR author's standing, and specific LOR language.
• Improving transparency and standardization in residency selection is crucial for equity and optimal outcomes.