ORIGINAL RESEARCH
William E. Freeman, MD; Carl E. Sanders, PhD; Andrés Hincapié, PhD; Donna B. Jeffe, PhD
Corresponding Author: Donna B. Jeffe, PhD, Department of Medicine, Washington University in St. Louis School of Medicine.
Email: jeffedonnab@wustl.edu
DOI: 10.3122/jabfm.2025.250266R2
Keywords: Certification, Clinical Clerkships, Data Set, Family Medicine, Health Workforce, Internal Medicine, Logistic Regression, Medical Education, Pediatrics, Racial Groups, Retrospective Studies, Secondary Data Analysis, Underserved Populations
Dates: Submitted: 07-11-2025; Revised: 10-16-2025; 12-01-2025; Accepted: 12-08-2025
Status: In Press.
PURPOSE: The shortage of primary-care physicians in the U.S. is exacerbated by declining numbers of U.S. medical-school matriculants planning careers in primary-care specialties (general internal medicine, family medicine, or pediatrics). We examined whether primare-care-specialty plans at matriculation, especially in family medicine, would be associated with long-term primary-care practice.
METHOD: Longitudinal data from multiple sources were obtained from the Association of American Medical Colleges for the national cohort of U.S. MD-granting medical-school matriculants in academic years 1998-1999 through 2000-2001. Graduates from 2002-2013 in office-based practice by May 2020 were included. Using two multivariable logistic regression models, we identified matriculating students’ specialty plans and medical-school experiences associated with office-based primary-care (vs. non-primary-care) practice.
RESULTS: Of 48,533 matriculants, 30,079 graduates (45.1% female) had complete data for analysis. Of 10,331 matriculants reporting primary-care-specialty plans, 4,177 (40.4%) reported primary-care practice at follow-up. Matriculants reporting primary-care-specialty plans, especially in family medicine (versus non-primary-care specialty plans: Model 1) were more likely to report primary-care practice at follow-up (adjusted odds ratios [aORs] range: 2.11-4.83). Matriculants reporting internal medicine (aOR: 0.44) and pediatrics (aOR: 0.72) plans (versus family medicine plans: Model 2) were less likely to report primary-care practice at follow-up. Graduating students planning to practice in underserved areas and rating some or all primary-care clerkships “Excellent” were more likely to be in primary-care practice.
CONCLUSIONS: As early as medical-school matriculation, primary-care-specialty plans, particularly in family medicine, were associated with long-term primary-care practice. Evaluations of interventions that promote interest in primary-care practice before and during medical school are warranted.