Can a resident's publication record predict fellowship publications?

Area of Science:

• Medical Education
• Internal Medicine Fellowships
• Academic Publishing

Background:

• Internal medicine fellowship programs aim to recruit residents with strong publication potential.
• The predictive value of an applicant's publication record for long-term scholarly output is not well-established.

Purpose of the Study:

• To determine if an internal medicine resident's publication record at fellowship application predicts future publications.
• To assess the validity of using publication counts as a selection criterion for fellowship programs.

Main Methods:

• Analysis of publication data for internal medicine residents applying for fellowships.
• Evaluation of publications occurring 3-7 years post-fellowship match.
• Calculation of diagnostic test characteristics (sensitivity, specificity, predictive values, likelihood ratios) and ROC curve analysis.

Main Results:

• Of 307 residents, 126 (41%) published post-fellowship, while 181 (59%) did not.
• The area under the receiver operator characteristic curve was 0.59, indicating poor predictive ability.
• No specific number of pre-application publications demonstrated adequate test characteristics for predicting long-term publishing.

Conclusions:

• The number of publications at fellowship application is an unreliable predictor of sustained scholarly productivity.
• Current practices of using publication records for fellow selection lack empirical validation based on this study's findings.