Objective Data Reveals Gender Preferences for Patients’ Primary Care Physician

Abstract

Background and Objectives:

Most studies based on self-reported data indicate that female patients more often than males have a same-gender preference for their primary care physician (PCP). Because self-reported preferences may not reflect true preferences, we analyzed objective data to investigate patients’ preferences for PCP gender.

Methods:

Analyses were performed on 2192 new patients seen within a university-based healthcare system by 13 PCPs (2 male, 11 female) during 2017. New patients were asked about their PCP gender preference when assigned a PCP. We compared the expected prevalence (proportion of males/females in overall patient population) and observed prevalence (gender distribution of patients for each PCP) by PCP gender. A mixed model with PCP as a random effect examined the odds of male and female patients being assigned a same-gender physician.

Results:

The expected prevalence of new patients was 65% female and 35% male. The observed prevalence (95% confidence interval [CI]) of male patients among male and female PCPs was, respectively, 59.7% (49.0%-69.5%) and 28.0% (24.0%-32.4%), with neither CI containing the expected prevalence of male patients (35%). Similarly, the observed prevalence of female patients among male and female PCPs was, respectively, 40.3% (95% CI 30.5%-51.0%) and 72.0% (95% CI 67.6%-76.0%), with neither CI containing the expected prevalence of female patients (65%).

Conclusions:

Both male and female patients often preferred to see a same-gender PCP with this preference more pronounced in males. Future research should seek to clarify the relationships between patients’ gender preferences, patient-physician gender concordance/discordance, patient satisfaction, and health outcomes.

Introduction

In the United States (US), patients have considerable autonomy when choosing their primary care physician (PCP). The US population of PCPs has become more diverse as the number of female providers and providers of different races has increased.¹ When queried about the relative importance of physician characteristics in selecting a PCP, patients consistently rank physician demographics, including gender, much lower than variables related to physician skill and experience.^2-5 Although gender may not be the most important factor when selecting a physician, patients who express a PCP gender preference usually prefer to see a physician of the same gender,^6-12 with women more often expressing a same-gender preference than men.^6-10

Considering patients’ preferences regarding PCP gender when matching patients and providers may impact patient satisfaction, an increasingly important outcome that affects provider reimbursment and public rankings.¹³ During medical encounters, male and female physicians have different communication styles with women providing more patient-centered care than men.^14-17 Patient-centered care is associated with increased patient satisfaction^17-19 as well as better adherence and self-management behaviors,^20,21 which are mediators of health outcomes. Patient-centered care is maximized when female providers treat female patients and lowest in female patient/male physician and male patient/male physician dyads.^16,22,23 Although not conclusive, these studies suggest that patient satisfaction may be mediated by both the physician’s communication style and the patient-physician gender dyad.

Important limitations of the current literature regarding patient preferences for physican gender is its reliance on patient self report, which may not reflect patients’ true preferences, and the age of the literature. To investigate whether patients have same-gender PCP preferences based on objective data, we examined the gender distribution of new adult (age ≥18 years) primary care patients within the Lousiana State University (LSU) Health System during 2017. This study was considered exempt from review by the LSU institutional review board.

Methods

Eligible patients had their first visit with 1 of 13 PCPs with medical degrees (MD or DO) between January 1, 2017 and December 31, 2017. When new patients are assigned to a PCP within the LSU Health System, they are asked about their preferred location and PCP gender by the scheduler. The scheduler offers the patient the first available apppointment with a PCP who meets their preferences for gender and location.

Analyses compared the expected gender prevalence (overall proportion of male/female patients) with the observed gender prevalence (proportion of male/female patients for each PCP). If PCP selection is not influenced by gender preference, the observed gender prevalence should be similar to the expected gender prevalence. A mixed model with PCP as a random effect examined the odds of new male and female patients being assigned a same-gender physician.

Results

Among the 13 PCPs, 11 (84.6%) were female and 2 (15.4%) were male. Each PCP saw an average of 169 new patients (range, 22-408 patients); the total number of new patients was 2192 (1427 females and 765 males). Patients’ average age was 50 years (range, 18-92 years).

Among all new patients, 65% were female and 35% were male; this is the expected prevalence of each gender. The observed prevalence (95% confidence interval [CI]) of new male patients among male and female PCPs was, respectively, 59.7% (49.0%-69.5%) and 28.0% (24.0%-32.4%); neither of these 95% CIs contains the expected prevalence of male patients (35%). Similarly, the expected prevalence of new female patients (65%) was outside the 95% CIs of the observed prevalance of new female patients among both male PCPs (40.3%; 95% CI 30.5%-51.0%) and female PCPs (72.0%; 95% CI 67.6%-76.0%).

Among male PCPs, the observed prevalence of new male patients was 24.8 percentage points higher than expected (P < .05) while the observed prevalence of new male patients among female PCPs was 6.9 percentage points lower than expected (P < .05). Male PCPs were 3.8 times more likely to see new male patients compared with female PCPs (odds ratio 3.8; 95% CI 2.4-6.1, P = .004). There was no statistically significant difference in patient age between male and female PCPs (P = .283). When stratifying patient age by 15-year increments, the higher than expected prevalence of new male patients seen by male PCPs was unrelated to patient age.

Discussion

Our data indicate that both female and male patients tend to prefer a same-gender PCP, with this preference more pronounced among male patients. These findings are in contrast with the results of most previous studies reporting a stronger same-gender PCP preference among female patients^6-10 but consistent with 2 previous studies showing that men have a stronger same-gender PCP preference than women.^11,12 The basis for patients’ PCP gender preferences has not been well studied. In a survey of Dutch patients, male and female patients endorsed similar reasons for preferring a same-gender PCP: they felt they could talk more easily to the provider and would be more at ease during the physical exam.¹¹ Gender-based assumptions about expertise and professionalism did not appear play a role in determining patients’ preferences.¹¹ Another study found that preference for a same-gender physician was strongest for specific issues, such as discussing sexual dysfunction and undergoing a physical exam.⁹ In a qualitative study, women who expressed a preference for a same-gender PCP felt that a female physician would listen better and have a greater understanding of their problems.⁶ Future research should investigate whether patient-physician interactions differ when patients who prefer a same-gender PCP are matched with a same- versus opposite-gender physician and whether matching patients with a PCP of their preferred gender impacts patient satisfaction, health behaviors, and health outcomes.

The relationship between patient gender, physician gender, and and perceived quality of care is not well understood due a lack of research on this topic and conflicting findings in existing literature. A study of cancer patients who were randomly assigned to a male or female physician for aftercare found that female-female dyads were associated with the highest satisfaction ratings and male-male dyads with the lowest ratings.²⁴ Among new patients randomly assigned to a PCP, those seeing female PCPs, regardless of gender concordance, were more satisfied than those seeing male PCPs.²⁵ A large survey of health maintenance organization members in Northern California conducted during 1994 to 1995 found that patients who chose female or male physicians were equally satisfied with their care, but that female patients who chose a female physician were less satisfied than patients in other dyads.²⁶ The authors posited that female patients who choose female physicians may be disappointed when these relationships do not meet patients’ ideals for gender-based care. Finally, a study that assigned college students to a computer-generated virtual physician found that patients were generally more satisfied with physicians who adopted a more caring communication style and that satisfaction was unrelated to physician gender.²⁷ However, there were interactions between patient gender, physician gender, and physician communication style: in female-female dyads, female patients were more satisfied with encounters with physicians who adopted a more caring communication style whereas the physician’s communication style had no impact on satisfaction in male-male dyads. Given these disparate findings and that only 1 study examined patients’ gender preferences, more research is needed to determine the ways in which patient gender, physician gender, and patient preferences for physician gender contribute to patient satisfaction and other outcomes.

This study had several limitations. Findings are specific to a single clinical institution and may not be generalizable to other settings. Although we were able to establish that, when given a choice and explicitly asked about their PCP gender preference, male and female patients tended to choose a same-gender PCP, we were not able to ascertain the reasons for these preferences and no data were available to examine potential relationships between patient gender preferences, patient-physician gender concordance/discordance, and outcomes. Additionally, our PCP panel was not sufficiently racially diverse to allow examination of patient preferences regarding PCP race and potential interactions between preferences for PCP gender and race.

In sum, our objective examination of patient preferences for PCP gender suggests that male patients tend to prefer male PCPs and that, albeit less strongly, female patients tend to prefer female PCPs. As the gender distribution of practicing physicians becomes increasingly equal,²⁸ healthcare systems should be able to match primary care patients with a PCP of their preferred gender. Future research should seek to examine reasons for patient preferences as well as establish whether matching patients with PCPs of their preferred gender impacts processes and outcomes of care.