Objective: We sought to identify risk factors for attrition among obstetrics and gynecology residents.

Study design: We analyzed 2001-2006 American Medical Association Graduate Medical Education (GME) Census data for all residents who entered obstetrics and gynecology in 2001 to characterize residents who did not complete a 4-year training period in their initial programs ("attrition"). Multivariable logistic regression models identified predictors of attrition from among age, gender, race, Hispanic ethnicity, medical school type, and medical school graduation year.

Results: Of 1055 residents entering obstetrics and gynecology in 2001, 228 (21.6%) were in the "attrition" group (133 changed obstetrics and gynecology programs and/or completed training on atypical cycles; 75 changed specialty; 20 discontinued GME). Residents who were older, underrepresented minority race, Asian race, osteopathic- or international medical school graduates were more likely to be in the "attrition" group (each P < .05).

Conclusion: Analysis of a national cohort of obstetrics and gynecology residents identified substantial attrition and demographic risk factors.