If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Correspondence to: James Cawley, MPH, PA-C, School of Public Health and Health Services and School of Medicine and Health Sciences, The George Washington University, 2175 K Street, N.W., Washington, DC 20037. Phone: (202) 994-3573; fax: (202) 416-0433.
Salary discrepancies between male and female physicians are well documented; however, gender-based salary differences among clinically practicing physician assistants (PAs) have not been studied since 1992 (Willis, 1992). Therefore, the objectives of the current study are to evaluate the presence of salary discrepancies between clinically practicing male and female PAs and to analyze the effect of gender on income and practice characteristics.
Methods
Using data from the 2009 American Academy of Physician Assistants’ (AAPA) Annual Census Survey, we evaluated the salaries of PAs across multiple specialties. Differences between men and women were compared for practice characteristics (specialty, experience, etc) and salary (total pay, base pay, on-call pay, etc) in orthopedic surgery, emergency medicine, and family practice.
Findings
Men reported working more years as a PA in their current specialty, working more hours per month on-call, providing more direct care to patients, and more funding available from their employers for professional development (p < .001, all comparisons). In addition, men reported a higher total income, base pay, overtime pay, administrative pay, on-call pay, and incentive pay based on productivity and performance (p < .001, all comparisons). Multivariate analysis of covariance and analysis of variance revealed that men reported higher total income (p < .0001) and base pay (p = .001) in orthopedic surgery, higher total income (p = .011) and base pay (p = .005) in emergency medicine, and higher base pay in family practice (p < .001), independent of clinical experience or workload.
Conclusion
These results suggest that certain salary discrepancies remain between employed male and female PAs regardless of specialty, experience, or other practice characteristics.
Introduction
Despite years of progress, women in the United States workforce continue to earn less money than their male counterparts (
American Medical Association (AMA). (2008). Gender disparities in physician income and advancement (Board of Trustees Report 19). Available: http://www.ama-assn.org.
Association of American Medical Colleges. (2010). Total graduates by U.S. medical school and sex, 2006–2010 (Table 27). Available: https://www.aamc.org.
). Carter and Oliver first reported salary inequities between clinically practicing male and female PAs in 1983. In 1985, Oliver et al. again found that male PAs earned more than female Pas, but determined that the discrepancy largely resulted from the relatively greater economic contribution men made to practice revenue. In 1992, Willis reported that salaries differences between male and female PAs still existed despite comparable levels of experience and similar practice characteristics. Finally, in 2009, Zorn et al. determined that, among new graduate PAs, women earn less than men, even after controlling for numerous variables that included years of experience in health care, specialty, hours worked per week, and hours on-call per month. There have been no recent studies, however, evaluating gender-based salary differences among experienced PAs.
There is substantial research on salary differences between male and female physicians within the medical literature (
American Medical Association (AMA). (2008). Gender disparities in physician income and advancement (Board of Trustees Report 19). Available: http://www.ama-assn.org.
). Female physicians earn less across all specialty fields, including emergency medicine, internal medicine, family practice, pediatrics, and hospital medicine, even after accounting for practice variables such as number of hours worked and number of patient visits. In fact, recent literature identifies a widening salary gap between new male and female physicians (
). There are several hypotheses for why the disparities exist: Women fail to aggressively negotiate salary, sexism and/or discrimination, and/or demands of the clinician’s family. Various factors, including experience, practice characteristics, and specialty, contribute to the gender gap in physician income; however, they do not entirely explain the differences in compensation (
American Medical Association (AMA). (2008). Gender disparities in physician income and advancement (Board of Trustees Report 19). Available: http://www.ama-assn.org.
When the PA profession originated in the 1960s, the first recruits were ex-military corpsman and, until the mid-1980s, male graduates outnumbered female graduates (
). Additionally, the PA workforce, originally primary care based, has slowly redistributed into specialty practice, mirroring a similar trend in physician practice (
). As of 2009, 35% of PAs practice in primary care and 65% practice in specialty settings, including orthopedics, dermatology, and cardiovascular surgery (
). The AMA has recognized the contribution that women make to the medical workforce and, in a 2008 Report of the Board of Trustees, recommended that medical associations and other relevant organizations study gender-based inequities. Considering the overlap of physician and PA scope of practice, the increasing proportion of female PAs in the workforce, and the paucity of literature regarding gender influence on PA practice, further study of salary disparities in the PA profession is warranted. The results could have significant implications for recruiting and retention practices and the success of female PAs in clinical practice. Therefore, the primary objective of the current study is to evaluate the presence of gender-based salary discrepancies among clinically practicing PAs. The secondary research objective is to analyze the effect of gender on practice ownership, sources of income, and practice characteristics.
Methods
Sample and Instrument
The source of data was the 2009 American Academy of Physician Assistants’ Annual Census Survey; the survey is the largest of its kind and includes both AAPA members and nonmembers. Paper and web-based surveys were sent between February and August. The survey consisted of 30 multi-level, multiple choice questions related to PA workforce issues, including questions about participant demographics, salary, and practice characteristics (available: http://www.aapa.org/about-pas/data-and-statistics). Of the 72,433 AAPA members and nonmembers, 19,608 PAs participated in the 2009 census, for a response rate of 27% (
). The study sample was limited to civilian PAs (not employed by the federal, state, or local government). The rationale was two-fold: To model the study design completed in 1992 (Willis) and to account for the relative absence of gender-based salary discrepancies in government sectors.
Analysis
Standard descriptive statistics were used to calculate demographics (age, ethnicity, degree, practice setting, type and source of income, hours worked, type of job related activities, etc.). A series of chi-square, Mann–Whitney U, multivariate analysis of variance (MANOVA), analysis of variance (ANOVA), and multivariate analysis of covariance (MANCOVA) analyses were used to determine the effect of gender on total income, base pay, type and sources of income, and other practice characteristics. Data transformations were necessary to assume normality and maintain the integrity of multivariate analyses while avoiding the relatively limited statistical power offered by nonparametric analyses. Therefore, the study’s authors consulted two separate experts in statistics to assist with the analyses.
Type of Income, Ownership, and Specialty Fields
A series of “goodness-of-fit” chi-square analyses were utilized to determine significantly unequal distributions among men and women in variables that included source of income (such as a result of certain services or bonuses), specialties and worksites chosen, and practice ownership. To avoid issues with regard to unequal sample sizes in each group, data from 6,587 women were randomly selected for comparison with the 6,587 men in the present sample. Because this series of analyses represents 22 separate analyses, concerns regarding inflated type I error resulted in the application of a Bonferroni correction, resulting in a more conservative criterion for statistical significance (alpha = 0.002).
Practice Characteristics
A MANOVA was utilized to assess the impact of gender on a variety of practice characteristics. To satisfy the assumption of normality required by this statistic, square root transformation was applied to number of years worked, number of years in current area of specialty, number of outpatients per week, total on-call hours per month, funding available from the primary clinical employer for professional development, hours per week providing direct inpatient care, and number of patient visits per week. A Likert-scale item stating “Would you become a PA if you were choosing a career today?” with “Definitely yes” anchored at 1 and “Definitely no” anchored at 4, was also included in these analyses. Similarly, logarithmic transformation was applied to hours per week performing “other” patients-related activities and number of inpatient encounters per week. Inverse transformation was utilized to rectify significant skew for hours per week performed in non-patient duties, hours per week providing direct care to “other” patients, and number of “other” patient visits per week.
Sources of Income
Owing to violations of assumptions of normality that were not rectifiable by data transformation, Mann–Whitney U analyses were used to compare the medians of a variety of sources of income between male and female respondents. Because this represented nine individual analyses, a Bonferroni correction was applied to test at a more conservative level of significance (alpha = 0.005).
Base Pay and Total Income
To more precisely determine the impact of gender on base pay and total salary in the absence of the influence of practice characteristics as well as type of practice, MANCOVAs were attempted within the three most commonly reported general specialty areas, as listed in the 2009 AAPA PA Census: Family practice, emergency medicine, and orthopedic surgery. To limit issues related to intercorrelation, covariates were limited to years worked as a PA, number of patient visits per week, hours worked per week for primary clinical employer, and total on-call hours served per month. The assumption of linearity was satisfied in each analysis.
When analyzing data for orthopedic surgery, however, the covariate “hours worked per week for primary clinical employer” was removed owing to inter-correlation with total on-call hours per month (r = 0.113; p ≤ .001) and number of patient visits (r = 0.210; p ≤ .001). The remaining covariates, years worked as a PA, number of patient visits per week, and total on-call hours served per month, were subjected to square root transformation to establish normality. Similarly, MANCOVA assumptions for analyses within the emergency medicine practice data required square root transformation of years worked as a PA, total on-call hours per month, and the removal of nine outliers based on hours worked per week. Total patient visits per week were removed as a covariate owing to violation of assumptions of normality that were not rectifiable by data transformation or outlier elimination.
Assumptions for normality within the family practice data were met through square root transformation of total on-call hours per month, years worked as a PA, and logarithmic transformation of patient visits per week. MANCOVA results conducted with the family practice data could not be interpreted owing to heterogeneity of regression hyperplanes. However, all assumptions were met when utilizing ANCOVA assessing only base pay as a dependent variable.
Results
Subjects
Two thirds of all survey respondents were women (65%), averaged approximately 42 years of age, were predominantly Caucasian (82%); three quarters (73%) possessed bachelor’s degrees (Table 1).
Table 12009 AAPA Census Participant Demographics by Gender
Effect of Gender on Type of Income Received, Ownership, and Field of Specialty
Women more frequently reported income based on surgical assisting, overtime, shift differential, on-call availability, on-call services, bonuses based on practice productivity, and bonuses based on individual productivity (p < .001 for all comparisons; Table 2).
Table 2Frequency of Reported Sources of Income by Gender
Pay Based on
Men, n (%)
Women, n (%)
χ2(1)
p
Surgical assisting
1,395 (21.18)
1,787 (27.13)
48.29
<.001
Overtime
1,943 (29.50)
4,405 (66.87)
954.86
<.001
Shift differential
1,491 (22.63)
1,906 (28.94)
50.70
<.001
On-call availability
1,577 (23.94)
1,926 (29.24)
34.77
<.001
On-call services
1,486 (22.56)
1,861 (28.25)
42.01
<.001
Bonuses paid based on the productivity of the practice
More women reported working in internal medicine, pediatrics, dermatology, and obstetrics and gynecology than men. Men were more likely to report working in emergency medicine, cardiothoracic surgery, and in orthopedic surgery. Men and women were equally likely to report working in general surgery and cardiology (Table 3).
Table 3Frequency of Reported Primary Specialty by Gender
Additionally, men (2,700; 40.99%) were more likely than women (1,841; 27.95%) to report having some form of supervisory duty (p < .001). Men (831; 12.62%) and women (800; 12.16%) were equally likely to report some form of ownership in the practice (p = .443).
Effect of Gender on Practice Characteristics
MANOVA results indicated that gender significantly affected the chosen practice variables (p < .001). ANOVA results utilized as post hoc analyses indicated that men were slightly less likely to choose PA as a career in hindsight (p = .043), worked more years as a PA (p < .001), and spent more years in their current primary specialty (p < .001). Men reported working more hours per week for their primary clinical employer (p < .001), providing more direct care to inpatients (p < .001), providing more direct care to “other patients” on a weekly basis (p = .005), and greater involvement in non-patient duties (p < .008). Similarly, men reported serving more hours on-call per month compared with women (120.36 vs. 101.68 hours/month, respectively; p < .001). However, women reported performing more “other” patient-related activities relative to men (p < .001; Figure 1).
Figure 1Hours worked per week by job-related activities by gender. ∗p ≤ .01; ∗∗p ≤ .001.
Men reported more patient visits per week (p = .042), more inpatient encounters per week (p < .001), and more “other” patient visits per week (p = .040). However, no significant difference was noted with regard to total outpatient visits per week (p < .071). Finally, men reported having more funding available to them from their primary clinical employers for professional development compared with women ($1759.70 vs. $1456.57/year, respectively; p < .001; Figure 2).
Figure 2Number of visits or encounters per week by gender. ∗p ≤ .05; ∗∗p ≤ .001.
Results indicated that men reported receiving greater income for surgical assisting fees, overtime, shift differential, administrative pay, on-call availability, on-call services, income based on practice productivity or performance, and incentive based on individual productivity or performance than women (p < .001 for all comparisons; Figure 3).
Figure 3Median annual income based on source by gender. ∗p ≤ .001.
MANCOVA analyses indicated that gender predicted income in orthopedic surgery independent of the practice variables years worked as a PA, number of patient visits per week, and total on-call hours served per month. Post hoc ANOVA results indicated that men in orthopedic surgery report higher total income and base pay relative to women in the same field (p ≤ .001 for all comparisons; Figure 4).
Figure 4Median annual base pay and total income by practice setting (dollars). ∗p ≤ .01.
Consistent with these findings, MANCOVA results in emergency medicine predicted that base pay and total income seem to differ across genders when controlling for years worked as a PA, number of on-call hours per month, and hours worked per week (p = .016). Post hoc ANOVAs indicated that men in emergency medicine report higher total income (p = .011) as well as base pay (p = .005) than women (Figure 4).
Similar to the results yielded in orthopedic surgery and emergency medicine, men reported significantly higher base pay in family practice than did women when controlling for years worked as a PA, number of patient visits per week, hours worked per week for primary clinical employer, and total on-call hours served per month (p < .001; Figure 4). Simple side-by-side comparisons show that men in fact earn more than women across all specialties (Table 4); however, inferential analyses controlling for practice characteristics were limited to orthopedic surgery, emergency medicine, and family practice.
Table 4Annual Base Pay in Other Specialties by Gender (Dollars)
The goal of this study was to evaluate the presence of gender-based salary discrepancies among practicing PAs. The results indicate that male PAs continue to earn more than female PAs, even when controlling for practice variables including specialty, experience, work effort (number of hours worked or number of patient visits completed), and number of on-call hours worked. Willis found gender-based salary differences among experienced PAs in 1992 and since then an increasingly higher percentage of women have entered the PA profession (
American Medical Association (AMA). (2008). Gender disparities in physician income and advancement (Board of Trustees Report 19). Available: http://www.ama-assn.org.
). Comparably, the 2009 AAPA census data reveal that female PAs are more likely than male PAs to practice in general internal medicine, pediatrics, and obstetrics and gynecology. Although a higher percentage of women also reported working in family practice, the difference was not significant. Similar to female physicians, female PAs also work fewer hours and see fewer patients than their male colleagues.
Female health care providers may be choosing to practice in primary care specialties and work fewer hours due, in large part, to compatibility with family life (
American Medical Association (AMA). (2008). Gender disparities in physician income and advancement (Board of Trustees Report 19). Available: http://www.ama-assn.org.
Salsberg, E. (2007, November). Balancing tomorrow’s physician expectations and workforce realties: Results of the AAMC/AMA survey of physicians under 50. Presented at the Association of American Medical Colleges 2007 Annual Meeting, Washington, DC. Cited by American Medical Association. (2008). Gender disparities in physician income and advancement (Board of Trustees Report 19). Available: http://www.ama-assn.org.
American Medical Association (AMA). (2008). Gender disparities in physician income and advancement (Board of Trustees Report 19). Available: http://www.ama-assn.org.
Salsberg, E. (2007, November). Balancing tomorrow’s physician expectations and workforce realties: Results of the AAMC/AMA survey of physicians under 50. Presented at the Association of American Medical Colleges 2007 Annual Meeting, Washington, DC. Cited by American Medical Association. (2008). Gender disparities in physician income and advancement (Board of Trustees Report 19). Available: http://www.ama-assn.org.
American Medical Association (AMA). (2008). Gender disparities in physician income and advancement (Board of Trustees Report 19). Available: http://www.ama-assn.org.
American Medical Association (AMA). (2008). Gender disparities in physician income and advancement (Board of Trustees Report 19). Available: http://www.ama-assn.org.
also identified differences between male and female practice characteristics by analyzing the 2002 AAPA Census Survey. Similar to the current study, responses to the 2002 survey indicated that female PAs were more likely to work in lower paying primary care specialties, whereas men were more likely to practice in emergency medicine and surgical specialties (
). However, the current study demonstrates that, although a greater percentage of men practice in cardiothoracic and orthopedic surgery, comparable percentages of men and women practice in neurosurgery. In addition, a greater percentage of female PAs practice in dermatology, which is associated with one of the highest incomes overall and provides the highest total income for women. Interestingly, Lindsay also noted the historic sociologic trend that occupations become “feminized” as a result of deteriorating work conditions, a phenomenon not observed within the PA profession. Unlike other occupations, the PA profession was not experiencing a decrease in status or decline in autonomy when the gender shift occurred (
Owing to changes in the AAPA’s Annual Census Survey and differences in statistical analyses, one cannot directly compare the current survey results with the results of
study; however, in general the current study suggests that the income gap based on gender is not improving. Willis concluded that male PAs practicing in primary care made $5,000 (approximately 12%) more annually than female PAs working in family practice, despite nearly identical practice characteristics. Although the salary differences between male and female family practice PAs in the current study cannot be entirely attributed to gender, results demonstrate a $14,685 (approximately 17%) difference in base pay.
found that male and female PAs working in orthopedics made comparable incomes, whereas the current results indicate a significant income difference between male and female orthopedic surgery PAs, even after controlling for experience, number of patient visits per week, and total on-call hours. This may be explained by the relatively small number of PAs in orthopedics in 1991. Willis also found no significant difference in the salaries of male and female PAs practicing less than 1 year, whereas the study of new graduate income by
found that, between 1998 and 2006, female new graduate PAs earned an average of $3,490 less than male new graduates. If these comparisons are reflective of all practicing PAs, including experienced PAs and those working in other specialties, the apparent trend toward increasing gender-based salary differences is concerning.
Based on her finding that male PAs were more likely than female PAs to negotiate compensation based on some form of productivity,
suggested that type of compensation negotiated may have influenced gender-based income differences. In the current study, men and women were equally likely to report some form of ownership in a practice. Interestingly, female respondents to the 2009 AAPA survey were more likely to report receiving pay based on surgical assisting, shift differential, on-call availability, on-call services, practice productivity, and individual productivity, whereas male respondents reported receiving greater income from all of these activities. Men also reported receiving more funding for professional development than women. These results suggest that, although women may be negotiating pay based on services performed, they receive lower rates of compensation and perhaps fewer benefits than men.
Other Influencing Factors
The current study attempted to control for practice characteristics such as number of patients seen and on-call hours; however, the influence of administrative duties could not be assessed. Male respondents to the 2002 AAPA Census Survey were more likely to report supervising others (
). Similarly, a greater percentage of male respondents to the 2009 survey reported some form of supervisory duty. They also reported more involvement in non-patient duties and greater income from administrative pay than women. These findings suggest that greater responsibility for practice management duties, as opposed to patient care, may be a factor contributing to gender-based salary differences.
American Medical Association (AMA). (2008). Gender disparities in physician income and advancement (Board of Trustees Report 19). Available: http://www.ama-assn.org.
has determined that, although experience, specialty, and work effort contribute to the gender-based gap in physician income, part of the disparity results from gender bias and discrimination within the medical profession and in society as a whole. Income disparity between male and female PAs also exists after controlling for influencing variables. In addition, physicians and PAs share common workplace settings. Therefore, gender bias and discrimination likely exist to some degree within the PA profession as well.
The persistence of gender-based income inequities among male and female PAs may have significant implications for the profession. PAs’ incomes have steadily increased over the years (
). The ongoing difference in income may indicate that, although women earn less, their salaries have little influence on male PA salaries. Alternatively, lower pay for women in a female-dominated profession may be exerting a negative effect on salaries for the entire profession as a whole (
). Other considerations include the distribution of PAs in primary care and the mobility of PA practice. Among physicians, an increase in the number of female primary care physicians has helped to compensate for a decrease in the number of men practicing in primary care (
). Female PAs are also more likely than their male colleagues to practice in primary care but, unlike physicians, PAs can readily change specialties. At least half of all practicing PAs change specialties at some time in their careers (
). Lower income for women practicing in primary care, even when compared with men in the same disciplines, may influence female PAs to seek employment in more lucrative specialties, thus exacerbating the redistribution of PAs out of primary care.
Limitations to the current study include potential self-selection and self-report biases that may have influenced survey results. More important, the American Academy of Physician Assistants’ annual census was not designed for gender-based income analyses or inferential statistics. Therefore, substantial data transformations and nonparametric analyses were used to compare salaries between genders. Finally, analyses of the effects of gender on income within all specialties were beyond the practical scope of this project; therefore, only family practice, emergency medicine, and orthopedic surgery were examined. Despite the study limitations, the results highlight the need for further research on gender-based income inequity within the PA profession.
Acknowledgments
The authors gratefully acknowledge the American Academy of Physician Assistants for providing data from their 2009 Annual Census Survey and thank Mitsy Audrain for editorial assistance.
References
American Academy of Physician Assistants. (2009). 2009 AAPA Physician Assistant Census National Report. Available: http://www.aapa.org.
American Medical Association (AMA). (2008). Gender disparities in physician income and advancement (Board of Trustees Report 19). Available: http://www.ama-assn.org.
Association of American Medical Colleges. (2010). Total graduates by U.S. medical school and sex, 2006–2010 (Table 27). Available: https://www.aamc.org.
Salsberg, E. (2007, November). Balancing tomorrow’s physician expectations and workforce realties: Results of the AAMC/AMA survey of physicians under 50. Presented at the Association of American Medical Colleges 2007 Annual Meeting, Washington, DC. Cited by American Medical Association. (2008). Gender disparities in physician income and advancement (Board of Trustees Report 19). Available: http://www.ama-assn.org.
Bettie Coplan, MPAS, PA-C, is an Assistant Professor for the Physician Assistant Program at Midwestern University, Glendale, Arizona. She serves as coordinator the program’s clinical master’s track.
Alison C. Essary, MHPE, PA-C, is Interim Director and Associate Professor at the Midwestern University Physician Assistant Program, Glendale, Arizona.
Thomas B. Virden, III, PhD, is an Associate Professor in the Clinical Psychology Program, Midwestern University, Glendale, Arizona, and served as the statistical consultant on the project.
James F. Cawley, MPH, PA-C, is Professor of Prevention and Community Health in the School of Public Health and Health Services at The George Washington University.
James D. Stoehr, PhD, is a Professor and Associate Director for Master’s Education at the Midwestern University Physician Assistant Program. He also serves as the coordinator of the research master’s track.
Article info
Publication history
Published online: August 08, 2011
Accepted:
June 15,
2011
Received in revised form:
June 15,
2011
Received:
January 10,
2011
Footnotes
Funding for statistical consultation was provided by Midwestern University’s College of Health Sciences.
00159-9&id=gr1.jpg){kind=link}
00159-9&id=gr2.jpg){kind=link}
00159-9&id=gr3.jpg){kind=link}
00159-9&id=gr4.jpg){kind=link}