Who Makes It to the End?: A Novel Predictive Model for Identifying Surgical Residents at Risk for Attrition

Area of Science:

• Medical Education
• Surgical Training
• Predictive Analytics

Background:

• High attrition rates (nearly 25%) exist for categorical general surgery residents.
• Limited national-level data identifies individual or program factors predicting resident attrition.

Purpose of the Study:

• To identify factors predicting general surgery resident attrition using a novel predictive model.
• To provide data for early identification of at-risk residents for targeted interventions.

Main Methods:

• A cross-sectional survey of general surgery interns (2007-2008) was conducted.
• Demographic and program data were linked with American Board of Surgery data.
• Classification and Regression Tree analysis identified risk factors for non-completion.

Main Results:

• The study analyzed 1048 interns, with 672 completing residency (20% noncompletion rate).
• Sex was the strongest predictor of attrition.
• Lowest attrition rates were observed in White, non-Hispanic, married men in small community programs (6%) and women in smaller academic programs (11%).

Conclusions:

• This is the first longitudinal study identifying early training risk factors for resident attrition.
• The findings offer a method to identify high-risk interns at the start of training.
• Future steps include developing and testing targeted interventions for at-risk residents.