Predictors of Primary Care Career Choice: A Review of AMCAS Applications of Four Graduating Classes at a New Medical School

Abstract

The United States (U.S.) is currently facing a shortage of primary care physicians, an issue particularly salient in Arizona. The purpose of this project is to investigate predictors of students entering primary care specialties that may be apparent from their American Medical College Application Service (AMCAS) applications, in order to best serve the needs of the physician workforce. We hypothesized that factors such as female gender, older age at application (“non‐traditional” students), and being raised in a rural/underserved community background may be predictors of students who choose primary care fields. AMCAS applications are completed by aspiring medical students and contain demographic information including gender, age, race, languages spoken, and family/community characteristics. Data provided also include academic factors such as college major, grade point average (GPA), and Medical College Admissions Test (MCAT) scores. Other subjective data reported by applicants include descriptions of extracurricular activities and a personal statement. The National Resident Matching Program (NRMP) is the system used by graduating students during the last semester of medical school to match students with their choice of specialty and the residency program wherein they will spend an additional three years, minimum, in training. Based on the Association of American Medical Colleges (AAMC) designations, we are considering primary care to be Family Medicine, Pediatrics, Internal Medicine, and Medicine‐Pediatrics. We examined data from AMCAS applications of all 149 students who graduated from the University of Arizona College of Medicine‐Phoenix between 2011‐2014, and compared to their NRMP match outcomes. Comparisons were made between non‐primary care versus primary care‐overall, as well as Family Medicine alone versus all other matches given the increasing rate of specialization within Internal Medicine and Pediatrics. Multiple logistic regression revealed two predictors of primary care career choice compared to non‐primary care: having more siblings (P=.003) and non‐physician parents (P=.017). Specific to Family Medicine, several predictors were identified compared to the non‐Family Medicine cohort: a slightly greater percentage of earned community college credits (P=.03), lower MCAT physical science (P=.009), higher MCAT verbal scores (P=.02), and lower paternal education (P=.003). Our analyses suggest having a greater number of siblings and non‐physician parents may predict primary care career choice. Specific to Family Medicine, academic factors including community college enrollment and MCAT scores may be of predictive value. Though the exact implications behind these findings are unclear, it is important to continually examine such data as medical schools can shape admissions selection criteria targeted at increasing the number of graduates seeking careers in primary care.