Primary care providers’ willingness to continue gender-affirming hormone therapy for transgender patients

Abstract

Background

Most transgender individuals either use or are interested in using gender-affirming hormone therapy (HT). Making gender-affirming HT available in primary care is critical for quality care to this vulnerable population. The barriers that transgender patients experience to accessing this treatment may be exacerbated if primary care providers (PCPs) will not provide it. Little is known about PCPs’ willingness to administer HT to transgender patients.

Objective

To examine whether PCPs are willing to continue prescribing HT for transgender patients and the factors that predict such willingness.

Methods

An online survey of internal and family medicine physicians and residents practising in a large integrated Midwest health system (n = 308); 158 responded to the relevant questions (51.3%).

Results

Approximately 50% of respondents were willing to continue HT for transgender patients. Most participants had previously met a transgender person (77%), and approximately half of them had cared for a transgender patient in the past 5 years. Multivariate logistic regression results indicate that attending physicians had lower odds of willingness to continue HT compared with medical residents, and those who reported perceived capability of providing routine care to transgender patients had higher odds of willingness.

Conclusions

Only about half of PCPs surveyed were willing to continue HT for transgender patients. Our study indicates that both personal and clinical factors play a role. Future research should address ways to increase PCPs’ willingness and comfort related to continuing HT for transgender patients.

Introduction

Recent estimates suggest that there are nearly one million transgender individuals in the US population (1). Many transgender individuals—those who experience conflict between their gender identity and assigned sex at birth—either seek or would like to seek medical transition care to align their physical characteristics and presentation with their gender identity. The medical aspect of transition may include several interventions, including the use of gender-affirming hormone therapy (HT) (2). This usually involves prescribing testosterone in various forms for male-identifying transgender individuals and oestrogen and androgen blockers for transgender women, to induce virilization and feminization, respectively (3).

Gender dysphoria can cause severe stress and anxiety, and HT has a number of positive mental health effects for transgender individuals. These include improvements in depression, anxiety and functional impairment (4) and lower risk of binge drinking, drug use and suicidal ideation among transgender women (5). Transgender men receiving HT have reported better quality of life and mental health than a control group (6). Given these documented positive effects, HT should be widely available for transgender individuals.

However, transgender men and women encounter numerous barriers to accessing both general health care services and gender-affirming (medical transition) care. Many transgender people report being ridiculed or harassed in health care settings (7,8) and have also reported that providers refuse to use their correct gender pronouns (corresponding to their gender identity) (7). Of respondents to the 2015 US Transgender Survey (USTS), 33% reported at least one negative experience with a health care provider in the past year (8). Emerging evidence also suggests that not at all transgender individuals who want HT can access it. Among transgender men and women who participated in the USTS, 24% of respondents wanted to receive HT but had never received it (8). A study published in 2013 suggests that about 8% of transgender people have recently been refused HT by a provider (9).

Whether transgender patients receive HT from an endocrinologist or a primary care provider (PCP) may depend on the type of provider available to them who is accessible, comfortable and knowledgeable about such care. Endocrinologists have greater expertise on the subject and may often initiate HT. However, long-term endocrinological management of HT should be reserved for complex cases, as with other common endocrinological needs, such as thyroid disease. HT for transgender patients is considered straightforward, and PCPs or other non-specialists can prescribe it in the long term (10). Considering barriers to health care access and the importance of HT to this population, it is imperative that PCPs be able and willing to provide HT for transgender patients.

Prior studies have found that primary care physicians do not routinely discuss gender identity with patients (11) and report a general lack of knowledge as well as ethical concerns about transition care for transgender patients (12). Comfort with sexual and gender minority patients in general appears to be correlated with clinical experience with this patient population (13). However, very little is known about the extent to which PCPs are willing to provide HT to transgender patients. This study seeks to determine whether primary care physicians would be willing to prescribe HT in cases where another provider, such as an endocrinologist, initiated the HT regimen. We aim to (i) determine the extent to which PCPs are willing to refill, or continue, HT for transgender patients and (ii) the factors that predict willingness to continue HT. Specifically, we explore the role of factors identified by prior studies mentioned above as well as variables informed by the Theory of Planned Behavior (14) and Intergroup Contact Theory (15), including provider characteristics, personal and clinical exposure to transgender people, empathy, transphobia and barriers and facilitators related to caring for this patient population.

Methods

Setting and sample

The cross-sectional survey study drew on the salaried clinicians in 26 primary care clinics throughout metropolitan Detroit employed by a large integrated health system. At the time of the study, there was no formal specialty care programme for transgender patients in the health system. The nearest comprehensive gender services programme was located ~45 miles away from the main campus of the health system. Eligible participants included internal medicine and family medicine PCPs, including attending physicians, physician assistants and residents. Health system departmental records were used to identify all PCPs working in the 26 clinics (n = 308).

Procedures

In November 2015, we invited PCPs to complete an online survey. Research Electronic Data Capture (REDCap) tools hosted at the health system were used (16). Each eligible participant received a unique link to the survey. The survey took ~15 minutes to complete. Those who completed the survey were each offered a $30-gift card and automatically entered into a random draw to receive one of three $100-gift cards.

Survey instrument and variables

The survey was developed in part based on adaptations of previously published survey questions or previous studies whose authors were willing to share survey instruments; some questions were developed specifically for this study. The domains addressed in the survey include provider characteristics, personal and clinical exposure to transgender individuals, transphobia, provider empathy, treatment barriers and facilitators, and willingness to care for transgender patients.

Variables

Provider characteristics

Provider characteristics included age, gender, race/ethnicity, sexual orientation, religious identity, religiosity, political views, medical specialty and provider type. Religiosity was measured using the question, ‘To what extent do you consider yourself a religious person?’ and response options included ‘Very religious’ (4), ‘Moderately religious’ (3), ‘Somewhat religious’ (2) and ‘Not at all religious’ (1) (17). The item score was used for the analysis (theoretical range, 1–4). Religious identity was categorized as Atheist/Agnostic, Christian, Muslim, Jewish, Hindu and Other. Political views were categorized as liberal, moderate or conservative.

Personal exposure

Personal exposure was measured using three binary variables: ever met a transgender person, transgender acquaintances or colleagues, and transgender friends or family. Response categories included ‘yes’ and ‘no.’

Clinical exposure

Two binary variables constructed for this study characterized clinical exposure to transgender patients: (i) cared for transgender patient in past 5 years (yes/no); and (ii) ever continued HT for transgender patient (yes/no).

Transphobia

Eight items from a previously validated transphobia scale were used to measure transphobia, with items measured on a 7-point Likert scale (1 = ‘strongly disagree’ and 7 = ‘strongly agree’) (18). To calculate scale scores, item scores were summed and averaged. Higher scores indicated a greater degree of transphobia. Exploratory factor analysis indicated a one-factor solution; factor loadings ranged from 0.510 to 0.814. Cronbach’s alpha for the resulting scale was 0.846, similar to the alpha for the original scale.

Empathy

Three items from a previously validated empathy scale for health care providers (19) were used to characterize empathy towards transgender patients. The questions used a 7-point Likert scale (1 = ‘strongly disagree’ and 7 = ‘strongly agree’). Scale scores were determined by averaging the sum of item scores; higher scores indicate greater empathy. An exploratory factor analysis produced a one-factor solution (factor loading scores ranged from 0.431 to 0.849). Cronbach’s alpha was low (0.320); however, the scaled variable was used due to the consistent relationships between each of the three individual items and the outcome variable.

Barriers and facilitators

Three items that were adapted from a previous study (20) assessed perceived capacity to provide transgender care. These included a 4-item knowledge barriers scale, familiarity with any hormone regimen for transgender patient care, and a question addressing one’s perceived capability to provide routine care to transgender patients. Knowledge barriers included barriers related to training on transgender health issues and familiarity with gender transition guidelines. Each barrier where there was agreement (somewhat agree, agree, or strongly agree) constituted 1 point on the barriers scale (theoretical range = 0–4). This scale demonstrated good internal consistency (Cronbach’s alpha = 0.849), with item means ranging from 0.393 to 0.578. All four items loaded onto one factor, with factor loadings ranging from 0.790 to 0.886. Similarly, any agreement with a statement about familiarity about HT regimens for transgender patients (somewhat agree, agree, or strongly agree) was coded as ‘yes’, and any agreement regarding perceived capability to provide routine care to transgender patients (somewhat agree, agree, or strongly agree) was also coded ‘yes’ to create two binary variables. Those who did not express agreement were coded ‘no’.

Willingness to continue hormone therapy for transgender patients

Willingness to continue HT (the dependent variable) was measured on a 7-point Likert scale with answers ranging from ‘strongly agree’ to ‘strongly disagree’. Participants indicating that they ‘strongly agreed’, ‘agreed’ or ‘somewhat agreed’ were coded ‘yes’ and all others were coded ‘no’.

Statistical analysis

Descriptive analyses were conducted for all study variables, including frequencies for nominal measures, and means and standard deviations for ordinal and continuous measures. Chi-square tests and t-tests for nominal and continuous variables, respectively, were conducted to assess the relationship between each predictor variable and willingness to continue hormone treatment. Independent variables that were associated with willingness to continue HT in bivariate analyses at the P < 0.10 level were included in multivariate analysis. Binary logistic regression was used to predict willingness to continue HT for transgender patients. Variables were entered simultaneously.

Results

A total of 163 PCPs completed the survey. Respondents were included in the analysis if they answered the survey question about willingness to continue HT (N = 158; response rate = 51.3%). Although response rates did not vary considerably by specialty (50.2% of internal medicine providers responded versus 53.9% of family medicine providers), medical residents were more likely to respond to the survey (56.7%) compared to attending physicians (46.0%).

Provider characteristics

Providers were on average 40.2 years old, 57% women and mostly heterosexual (92.4%). Half (50.6%) defined their race/ethnicity as white and 25% were Asian or Pacific Islanders. Nearly half identified as Christian (45.6%) and politically liberal (45.6%). In terms of specialty, most were internal medicine providers (69.6%). Half (50.6%) were medical residents, 47% were attending physicians and the remainder were physician assistants (Table 1).

Willingness to continue hormone therapy and other variables

Overall, 50.6% of respondents were willing to continue HT that another provider had initiated. Most participants had met a transgender person in the past (77.8%), though few respondents identified having transgender family, friends, acquaintances or colleagues (Table 2). Over half had cared for a transgender patient in the past 5 years (53.2%), but few (12.7%) had ever continued an HT prescription for a transgender patient. The mean empathy and transphobia scores were 5.7 (SD = 0.9) and 3.2 (SD = 1.1), respectively (out of 7-point scores). Respondents reported a mean of 1.7 knowledge-related barriers (out of 4-point score). About two-thirds of respondents (67.7%) felt capable of providing routine care to transgender patients.

Bivariate results

Provider characteristics significantly associated with willingness to continue HT included race/ethnicity, religious identity and provider type (Table 1). In addition, transphobia, perceived capability of providing routine care, familiarity with any HT regimen, willingness to provide routine care and willingness to initiate HT were all significantly associated with willingness to continue HT.

As shown in Table 2, compared to an overall transphobia score of 3.2 on the 7-point scale, participants who were willing to continue HT had a mean transphobia score of 1.0 (SD = 0.1), indicating lower transphobia (t = –2.239, P = .027). Of those who felt capable of providing routine care, 57% were willing to continue HT compared to 36% of those who did not feel capable of providing routine care (χ² = 6.017, P =0.014). Most participants (73.7%) who were familiar with HT for this population were willing to continue HT, compared to 48% among those who were not familiar with HT regimens (χ² = 4.591, P = 0.032).

Multivariate results

The logistic regression model predicting willingness to continue HT was significant, χ² = 45.637, P < 0.001, R² = 0.367 (Table 3). Variables that significantly contributed to the model included religious identity, provider type and perceived capability of providing routine care. Compared to Christian respondents, the adjusted odds of Jewish providers being willing to continue an HT regimen was over 12 times higher (AOR = 12.12, P = 0.043), while the odds were lower for respondents who reported their religious identity as ‘Other’ (AOR = 0.137, P = 0.036). The adjusted odds of attending physicians being willing to continue HT were significantly lower than those of medical residents (AOR = 0.220, P = 0.001). Finally, the odds of being willing to continue HT were over twice as high for participants who felt capable of providing routine care to transgender patients compared to those who did not feel capable (AOR = 2.447, P = 0.049).

Discussion

We found that only half of PCPs surveyed in this study were willing to continue HT for patients who had been started on HT by another physician. Furthermore, logistic regression results suggest that medical residents were more willing to continue HT compared to attending physicians, and those who felt capable of providing routine care to this patient population were nearly 2.5 times more willing compared to those who did not feel capable. Notably, our results suggest that factors such as lack of familiarity with HT regimens were not significant once other factors were controlled.

It is not clear why medical residents would be more likely to be willing to continue HT compared to attending physicians, but it is unlikely to be related to age or generational shifts in attitudes; there was no relationship between age and willingness to provide HT. It may be that residents are more eager to learn about new medication regimens because they are in training, or more accustomed to providing care with which they are not completely comfortable. Certainly, improving perceptions of capability among providers should begin with including transgender health content in provider education curricula. A recent study found that the median total time spent on sexual and gender minority health in medical school curricula in Canada and the United States was 5 h. Worse, only 30.3% of medical schools report spending any time at all on the topic of gender transitioning and 34.8% cover sex reassignment surgery (21). Yet, a recent study found that adding just a 1-h lecture on transgender health and HT can significantly increase medical students’ comfort and willingness to treat transgender patients (22).

Although our results indicate potential differences in willingness to provide HT by religious identity, small sample sizes in several categories limit our ability to interpret these results. However, past studies have clearly indicated that health care providers’ values and biases play a role in dictating their behaviour and that physicians are often not even aware of their biases (23). In addition, ‘religious freedom’ has been often invoked to excuse discrimination against transgender people (24).

If internal medicine and family medicine physicians are willing to continue HT, transgender patients will be more likely to receive comprehensive health care services, which could mitigate the physical and mental health disparities that this population experiences. Being able to access HT in primary care settings may lead to a higher likelihood of establishing a medical home, and the resulting use of preventive care services, risk reduction and cancer screening, and chronic disease management, as well as other important routine care services. Notably, respondents in the USTS were three times more likely to have to travel more than 50 miles for transition-related care than for routine care, and among those who reported having two separate providers for routine and transition care, over half were unsure about how much their PCP knew about transgender health (8).

Limitations

This study has several limitations. First, causality cannot be established due to the cross-sectional nature of the data, and a relatively small sample size may have limited the power of the study. However, this is the first study to examine associations between willingness to provide care to transgender individuals in terms of continuing HT and both personal and clinical factors that may influence PCPs. Second, this sample may be biased towards providers who are more sympathetic to or interested in the transgender population or transgender health care. We were unable to evaluate characteristics of respondents versus non-respondents, with the exception of provider type and specialty. Furthermore, health care providers may have over-reported their willingness to treat transgender patients due to social desirability. Data were collected in one integrated Midwestern health system; thus, results may not be generalizable to health care providers in less integrated health systems or other parts of the country. In addition, there may be other factors that influence providers’ willingness to continue HT that we did not examine in this study.

Conclusions

Results of this study indicate that many PCPs are not willing to continue HT for transgender patients. Perceived capability significantly predicted willingness to provide HT above and beyond self-reported knowledge, familiarity, exposure and other factors. This suggests that capability may be related less to medical and pharmaceutical knowledge and more to the interpersonal skills that the provision of excellent medical care requires. Providers may not feel capable of interacting with transgender patients if they do not know which pronoun to use or what name to call a patient, feel comfortable discussing gender-affirming care, and other areas related to being culturally sensitive to transgender patients. While adding training on gender-affirming care for transgender patients to medical school curricula may be helpful, adding cultural competence or humility modules is also critical (25). Future studies should explore factors associated with PCPs’ perceived capability to care for transgender patients and whether improving perceived capability will, in turn, affect providers’ willingness to participate in medical transition care for this patient population.

Acknowledgements

Many thanks to Dr. Jessica Shill for her insight and comments on this manuscript. REDCap is a secure, web-based application designed to support data capture for research studies, providing (i) an intuitive interface for validated data entry; (ii) audit trails for tracking data manipulation and export procedures; (iii) automated export procedures for seamless data downloads to common statistical packages; and (iv) procedures for importing data from external sources.

Declaration

Funding: Funding sources include CSWE/NASW Foundation Social Work HEALS Doctoral Fellowship, funded by the New York Community Trust’s Robert and Ellen Popper Scholarship Fund; the Blue Cross and Blue Shield of Michigan Student Award Program; and the Graduate Medical Education Fund at Henry Ford Hospital.

Ethical approval: This study received IRB approval at the participating health system.

Conflict of interest: The authors report no conflicts of interest.

References