In this retrospective study, students who volunteered for the ICC were 29% more likely to fit into a primary care specialty than students who did not volunteer (RR 1.29, 95% CI 1.05-1.59) . ICC senior officers were more than twice as likely to fit into the GP practice compared to volunteers who were non-senior officers (RR 2.41, 95% CI 1.10, 5.26) or non-volunteers (RR 2, 27, 95% CI 1.05-4.89). . However, a post-hoc analysis found that this difference is most likely due to higher levels of senior volunteering compared to non-managerial employees. Stratified by the number of shifts worked, the proportion of executives who fit into family practice was no different from the non-executive volunteers. In addition, in multiple logistic regression, the number of shifts worked correlated significantly with adjustment to family practice (p= 0.042), while executive status does not (p= 0.247). In fact, each additional volunteer shift was associated with a 3.9% greater likelihood of fitting into the GP practice when the senior executive’s status as a potential confounder is accounted for. There was no association between volunteering and competitiveness of specialty choice.
These results build on two previous studies that showed a relationship between SRFC volunteering and student interest in primary care subjects. In a survey of 914 medical students at the University of California San Diego before and after volunteering at their SRFC, Smith et al. found that involvement in the SRFC increased students’ interest in primary care and increased students’ interest in working with underserved people. [14]. Limitations included the lack of a control group and the subjective assessment of interest in primary care rather than an objective measure such as specialty choice, if any. Similarly, Campos-Outcalt et al. showed in 1985 that volunteers at the University of California Davis SRFC were more likely to enter primary care residential homes than non-volunteers. [15]. Their study was more objective because it used students’ choice of specialty using match lists as a measure of outcome, but it is outdated and limited in scope because only Hispanic medical and pre-med students were analyzed.
On the other hand, several recent studies have found no significant association between volunteering at an SRFC and the choice of primary care specialist. However, these studies had notable limitations [16,17,18,19]. Brown et al. stated that their survey had a low response rate (39.8%) and may not have been representative of the volunteers at their clinic. [18] Similarly, Tran et al. suggested that their conclusions were limited by a small sample size (136 students) and large standard deviations in survey responses. [16]. Vaikunth et al. performed a more robust analysis in their study. They used objective measures such as alumni match statistics for subject interests (rather than surveys) and considered students’ academic performance. However, their SRFC was specifically focused on catering to Hispanic populations, who may have chosen to recruit student volunteers with knowledge of Spanish. [17] Her resulting analysis was less representative of her entire med school class. In addition, unlike this study, they included obstetrics/gynecology as a primary care specialty.
The present study used the quantitative outcome measure of the alumni concordance statistics and did not rely on self-reported survey data. Our analysis involved a large sample size of 506 students, evenly distributed between the voluntary (50.2%) and non-voluntary (49.8%) cohorts. Additionally, our clinic does not target volunteers who speak any particular language and is highly representative of the entire medical student class at CMS. Finally, this analysis included several unique measures, including leadership involvement, number of shifts worked, and specialty competitiveness. The results of this research indicate that each SRFC involvement at our facility is associated with the pursuit of primary care specializations and the level of involvement with the pursuit of primary care practice. However, whether these associations represent a causal relationship requires further investigation.
There is some evidence in the literature that supports the notion that the associations observed in the present study are in fact causal. Long-term experience in primary care has been reported to increase the likelihood that medical students will pursue specialties in primary care [20, 21]. Volunteering at the ICC, which provides primary care services to underserved patient populations, can provide such lengthy primary care experience and thus inspire students to strive for primary care. When this is the case, SRFCs can be a valuable tool in stimulating interest in primary care specialties among medical students. However, it is worth noting that despite the increase in the prevalence of SRFCs in the US over the past ~15 years, the shortage of primary care physicians continues to increase [1, 3,4,5]. Not all SRFCs focus on primary care services and offer their students opportunities for long-term volunteer experience like our ICC does. Therefore, it remains possible that even if our ICC has a causal impact on student interest in primary care, other SRFCs may not. Future research should examine not only whether the associations found in this study are indeed causal, but also whether they extend to all SRFCs or only SRFCs with certain properties.
limitations
It is important to recognize the possibility that students with a pre-existing interest in primary care may have volunteered with the ICC more often than students with no interest in primary care, leading to selection bias. On the other hand, most students who volunteer with ICC are freshmen and sophomores, and studies have shown that most medical students change their desired specialty during medical school. [22, 23]. The AAMC reports that only 25.6% of medical school graduates studied the specialty they indicated as their intended field of practice prior to entering medical school [22]. Compton et al. report that only 30% of undergraduate medical students who were interested in primary care in their freshman year remained interested through their fourth year. [23]. Therefore, instability in medical student subject preferences, if present, could have a mitigating effect on this type of selection bias.
In addition, it is worth considering the quantification method used to approximate the competitiveness of medical specialties. Few studies have attempted to analyze competitiveness, and most have used compliance rate as a proxy for competitiveness [8,9,10,11]. However, the completion rate discourages student self-selection (i.e., medical students select the specialty to which they apply according to the strength of their application). A better measure of competitiveness should account for these variables, perhaps by including factors such as clerkship grades, board scores, and number of publications. Additionally, associations between volunteering and performance in later residencies were not assessed in this research. Another limitation is that the choice of specialty, if any, was defined using residency match data and no information on fellowships or further specialization was used. Therefore, some of the students identified as matching primary care subject areas may move on to subject areas outside of primary care. Finally, the number of shifts worked by ICC volunteers was determined using shift schedules, which may not have included last-minute changes such as volunteer cancellations or shift coverage.
