Entice With Procedures, Inspire With Primary Care: A Preclerkship Pipeline Course

Primary care in the United States faces a growing supply-demand mismatch. By 2033, the Association of American Medical Colleges projects a shortage of 21,100 to 55,200 primary care physicians and a fall in the ratio of physicians to physician assistants (PA) from 2:1 to 1:1.¹ It remains to be seen whether the rising supply of PAs will offset demand for primary care.¹ However, increasing proportions of US MD and PA graduates choose to subspecialize each year.^2-4 US medical institutions are not increasing the number of students committed to primary care careers, with less than 10% of graduates over the past decade entering a family medicine residency.^5-6

Stanford University School of Medicine is a private academic medical center with 90 MD and 25 PA students annually. From 2011 to 2020, fewer than 5% of each graduating class on average matched into family medicine residencies. Although more students enter internal medicine (21%) or pediatric (7%) residencies, most subspecialize.⁷ Current efforts to foster primary care interest include a patient-centered elective course, student interest groups, and faculty mentorship for research. Clerkship students spend a month in a family medicine rotation and may opt for primary care electives. Despite these opportunities, the pool of students interested in primary care remains modest. 

To cultivate early primary care interest among preclerkship students, we developed a novel seminar titled “Primary Care Defined: Perspectives and Procedures.” The course was split into procedural workshops and interactive sessions highlighting clinician perspectives to promote experiential learning and professional mentorship.^8,9 While similar educational strategies have been successfully applied to emergency medicine, anesthesiology, and surgical fields, none have been reported for primary care specialties.¹⁰ We hypothesized that early preclinical exposure to hands-on learning and meaningful clinical interactions would improve enrollees’ interest and attitudes toward a career in primary care, as measured by a modified validated Family Medicine Attitudes Questionnaire.

This elective was a weekly, 1-hour, lunchtime seminar for Stanford preclerkship students. Following an introductory session, there were four case-based procedural workshops and four interactive sessions. Procedures included abscess incision and drainage, joint injection, intrauterine device insertion, and shave biopsy. Workshops included a patient case, a live demonstration, and small-group practice with an MD or PA facilitator. The interactive sessions addressed key topics such as work-life fit, salary, and posttraining trajectories through panel discussions or breakout groups with clinicians. Speakers from family medicine, internal medicine, and pediatrics across different practice settings were invited. Full details on our curriculum and workshop supplies can be found in the STFM Resource Library.¹¹

To evaluate changes in enrollees’ attitudes toward primary care, we modified the previously-validated 14-item Family Medicine Attitudes Questionnaire, which covers six domains (patient relationships, competence and expertise, lifestyle, research, specialty importance, and shortage).¹² Our final unvalidated survey contained 11 5-point Likert scale items from the original instrument,¹³ with an additional question assessing interest in primary care, scaled from 1 to 7 for greater statistical sensitivity.¹⁴ We administered the survey during the first and last weeks of the academic quarter. A nonenrolled control group was recruited through school mailing lists. No compensation was offered for completing the questionnaire. We programmed the tool using Qualtrics XM software (Provo, UT).

We used one-way analyses of covariance (ANCOVA) to test the null hypothesis that enrollee attitudes toward primary care were equal to those of control participants. We assigned enrollment status as the fixed factor, with gender, program (MD vs PA), and precourse responses as covariates, based on variables with major differences between enrollees and controls. We also performed an ANOVA of change from baseline to confirm our findings due to concern for unconsidered selection bias.¹⁵ We used SPSS software, version 24.0 (IBM Corp, Armonk, NY) for statistical analysis, and set statistical significance at P<.05.

The Stanford University Institutional Review Board determined that this study did not meet the definition of human subjects’ research.

Twenty-four enrolled and 10 nonenrolled learners completed the pre- and postcourse surveys. Nine enrollees (38%) were MDs and 15 (62%) were PAs, compared to seven (70%) and three (30%) in the nonenrollee group, respectively. Gender distribution was comparable between enrollees (71% female) and nonenrollees (70% female). Table 1 shows other demographic information. Precourse responses were not significantly different between enrollees and nonenrollees (Table 2).

The one-way ANCOVA assessing postcourse level of interest in a career in primary care (correcting for gender, program, and pre-course ratings as covariates) demonstrated a statistically significant difference (F_1,32=9.22; P=.005) between enrollees (5.50 out of 7; 95% CI, 5.16-5.84) and nonenrollees (4.49; 95% CI, 3.94-5.05). The one-way ANOVA of change from baseline was similarly significant (F_1,19=5.10; P=.036), differing primarily in effect size. Significant differences were found in four of the 11 statements, specifically about income, job complexity, career enjoyment, and procedural involvement (Table 3). Statements regarding patient relationships and the importance of primary care (6, 9, 10, 11) received high levels of agreement from both groups. Differences were noted for statements about research and knowledge depth (5, 7, 8), but were not significant.

This novel curriculum received highly positive feedback. Learners favored the small class size, high faculty-to-student ratio, and diversity of skills learned. Constructive feedback included allotting more time for the procedural walk-through and reorganizing stations for smoother workflow. Learners also stressed the value of hearing honest perspectives from clinicians on important topics like income and job fulfillment. We observed high levels of engagement across all sessions, lending credibility to the course design and content.

 After the course, enrollees rated their interest in primary care significantly higher than did nonenrollees. They also scored more positively on statements regarding compensation, scope of practice, and job fulfillment. These differences suggest that this elective effectively defined key features of primary care and addressed common misconceptions.¹⁶ Enrollees and nonenrollees had high levels of agreement with statements concerning patient-physician interactions and the value of primary care, suggesting baseline appreciation of this information. Topics such as research opportunities and interactions between generalists and specialists may deserve more emphasis in future iterations.

Students likely enrolled in this course based on preexisting interest in primary care. Other selection factors include interest in procedural exposure or elective credit for graduation. Enrollees may have known more about the domains highlighted during the elective or felt obliged to submit higher ratings despite guarantee of anonymity. Additionally, many enrollees were PA students who might differ from medical students in their educational goals. Effects of precourse scores and program type, however, were controlled for in the ANCOVA. We did not control for home town size, which differed between enrollees and nonenrollees. Notably, a higher proportion of controls came from rural backgrounds, which may correlate with a higher interest in primary care specialties.¹⁷

While the original survey was validated as an instrument to assess fourth-year medical students’ attitudes toward family medicine and likelihood of matching into the specialty, we administered a modified version to preclinical medical students without revalidation, thus limiting the ability to make definitive conclusions based on the survey results. Long-term outcomes, such as match rates into primary care specialties, could not be evaluated given the recency of this course and its small sample size. We also did not stratify student responses by individual primary care specialty. The predominance of PA enrollees will additionally limit generalizability to other institutions. Future instructors should longitudinally follow a larger sample of past and future students for the analysis of residency match outcomes and possible validation of the modified survey instrument.

Our novel elective offers a framework for other medical schools to generate early positive messaging around the activities and impact of primary care specialties. Coupling experiential learning with meaningful student-faculty engagement may address concerns that deter preclerkship learners from pursuing careers in primary care.