Physician Assistants in Primary Care: Trends and Characteristics

DISCUSSION

Overall, we found that PAs working in primary care have demographic characteristics similar to those of medical students who choose general primary care specialties. A 2010 systematic review determined that 6 factors are associated with a medical student’s commitment to primary care practice: female sex, older age, Latino ethnicity, lower socioeconomic status, receipt of a National Health Service Corps scholarship, and intention to practice in primary care at the time of medical school matriculation.⁷ We found that PA practice in primary care is also more likely among those who are female, older, and Hispanic. The only other recent study of the relationship between demographics and PA practice setting also found that minority and older PAs are more likely to work in primary care.⁸ Furthermore, similar to the findings for medical students, the data on PAs identified a positive association between primary care practice and lower economic status.⁶

A US study of specialty choices among PAs between 1997 and 2006 found that the growth rate of PA practice in primary care was lower than the overall growth of the profession.⁵ Similarly, we found an overall decline in the percentage of PAs in primary care each year. In the cohort of 2004-2008 graduates, however, the percentage of primary care PAs increased modestly, by an average of 0.9% each year. It should be noted that nearly one-half of all PAs change specialties at some time during their careers.⁹ Interestingly, primary care specialties in general have the most staying power, with 53% of those in family practice making no change.⁹ We cannot predict the practice patterns of new graduates, but historically, the net decrease in the number of PAs in family medicine correlates with smaller percentages of new graduates choosing family practice.⁹ The slight upward trend of new graduate selection of primary care may therefore reflect some stabilization in the percentage of primary care PAs.

In general, medical schools and PA programs have had little success recruiting matriculants into primary care.^7,10 Educational programs may therefore want to consider factors associated with an increased likelihood of primary care practice when recruiting and selecting applicants. In addition, further research could be aimed at identifying the particular factors that cause practicing PAs to transition into and out of primary care.

The AAPA Census Survey targets all practicing PAs, and the proportion of respondents by sex was nearly identical to the current sex distribution of certified PAs.¹¹ The survey’s 27% response rate, however, raises the risk of self-selection, self-report, and nonresponse bias. For example, those in the profession longer are less likely to respond.¹² This group represents a smaller proportion of the profession but also includes PAs who graduated from PA school at a time when more graduates entered primary care. We were therefore limited in our ability to generalize the results to the entire population of PAs. Finally, the survey was cross-sectional and not prospective, and was not designed for analysis of demographic characteristics associated with specialties. Despite these limitations, our findings support prior research that indicates that women, older clinicians, and Hispanics are more likely to practice in primary care.^6,7,13

PA practice trends mirror the practice trends of physicians. Consequently, actions that increase the number of primary care physicians are also likely to increase the percentage of primary care PAs. The PA profession should therefore continue to support workforce policy measures that successfully increase the number of primary care clinicians including loan repayment, improved levels of reimbursement for primary care physicians, and expansion of Title VII, Section 747 of the Public Health Service Act, which aims to increase the quality, size, and diversity of the primary care workforce.¹⁴

Acknowledgments: The authors wish to thank the American Academy of Physician Assistants for providing the data for analysis and to acknowledge Kent Armeson for statistical support.