Yeo Heather L, Abelson Jonathan S, Mao Jialin, Lewis Frank, Michelassi Fabrizio, Bell Richard, Sedrakyan Art, Sosa Julie A
*NewYork-Presbyterian Hospital/Weill Cornell Medicine, Department of Surgery, New York, NY †Weill Cornell Medicine, Department of Healthcare Policy and Research, New York, NY ‡American Board of Surgery, Inc., Philadelphia, PA §Department of Surgery, Duke Cancer Institute and Duke Clinical Research Institute, Duke University Medical Center, Durham, NC.
Ann Surg. 2017 Sep;266(3):499-507. doi: 10.1097/SLA.0000000000002384.
We present 8-year follow-up data from the intern class of 2007 to 2008 using a novel, nonparametric predictive model to identify those residents who are at greatest risk of not completing their training.
Nearly 1 in every 4 categorical general surgery residents does not complete training. There has been no study at a national level to identify individual resident and programmatic factors that can be used to accurately anticipate which residents are most at risk of attrition out.
A cross-sectional survey of categorical general surgery interns was conducted between June and August 2007. Intern data including demographics, attendance at US or Canadian medical school, proximity of family members, and presence of family members in medicine were de-identified and linked with American Board of Surgery data to determine residency completion and program characteristics. A Classification and Regression Tree analysis was performed to identify groups at greatest risk for non-completion.
Of 1048 interns, 870 completed the initial survey (response rate 83%), 836 of which had linkage data (96%). Also, 672 residents had evidence of completion of residency (noncompletion rate 20%). On Classification and Regression Tree analysis, sex was the independent factor most strongly associated with attrition. The lowest noncompletion rate for men was among interns at small community programs who were White, non-Hispanic, and married (6%). The lowest noncompletion rate for women was among interns training at smaller academic programs (11%).
This is the first longitudinal cohort study to identify factors at the start of training that put residents at risk for not completing training. Data from this study offer a method to identify interns at higher risk for attrition at the start of training, and next steps would be to create and test interventions in a directed fashion.
我们展示了2007年至2008年实习医生班级的8年随访数据，使用一种新颖的非参数预测模型来识别那些完成培训风险最高的住院医师。
每4名普通外科住院医师中就有近1人未能完成培训。尚未有全国性研究来确定可用于准确预测哪些住院医师流失风险最高的个体住院医师因素和项目因素。
2007年6月至8月对普通外科实习医生进行了横断面调查。实习医生的数据包括人口统计学信息、在美国或加拿大医学院就读情况、家庭成员的距离以及家庭成员从事医学工作的情况，这些数据经过去识别处理后与美国外科委员会的数据相链接，以确定住院医师培训的完成情况和项目特征。进行了分类回归树分析，以识别未完成培训风险最高的群体。
在1048名实习医生中，870人完成了初始调查（回复率83%），其中836人有链接数据（96%）。此外，672名住院医师有完成住院医师培训的证据（未完成率20%）。在分类回归树分析中，性别是与流失最密切相关的独立因素。男性未完成率最低的是在小型社区项目中实习的白人、非西班牙裔且已婚的实习医生（6%）。女性未完成率最低的是在较小学术项目中培训的实习医生（11%）。
这是第一项纵向队列研究，旨在识别培训开始时使住院医师面临无法完成培训风险的因素。本研究的数据提供了一种在培训开始时识别流失风险较高的实习医生的方法，下一步将是有针对性地创建和测试干预措施。
