Health Care Satisfaction in Relation to Gender Identity: Behavioral Risk Factor Surveillance Survey, 20 States (2014–2018)

Abstract

Background:

Healthcare satisfaction is a key component of patient-centered care. Prior research on transgender populations has been based on convenience samples, and/or grouped all gender minorities into a single category.

Objective:

To quantify differences in healthcare satisfaction among transgender men, transgender women, gender non-conforming, and cisgender adults in a diverse multistate sample.

Research Design:

Cross-sectional analysis of 2014-2018 Behavioral Risk Factor Surveillance System data from 20 states, using multilevel logistic models.

Subjects:

We identified 167,468 transgender men, transgender women, gender non-conforming people, cisgender women, and cisgender men and compared past year healthcare satisfaction across these groups.

Results:

Transgender men and women had the highest prevalence of being “not at all satisfied” with the healthcare they received (14.6% and 8.6%, respectively), and gender non-conforming people had the lowest prevalence of being “very satisfied” with their healthcare (55.7%). After adjustment for sociodemographic characteristics, transgender men were more likely to report being “not at all satisfied” with healthcare than cisgender men (odds ratio (OR): 4.45, 95% confidence interval (CI): 1.72–11.5) and cisgender women (OR: 3.40, 95% CI: 1.31–8.80).

Conclusions:

Findings indicate that transgender and gender non-conforming adults report considerably less healthcare satisfaction relative to their cisgender peers. Interventions to address factors driving these differences are needed.

Introduction

Considerable health disparities by gender identity exist, with gender minority (i.e., transgender and gender non-conforming) populations experiencing disproportionate rates of physical, mental, and behavioral health outcomes relative to their cisgender peers.¹ Accumulating evidence indicates that healthcare satisfaction, a multifaceted construct that includes the quality of healthcare delivery,² may be an important determinant of these disparities. While there is no universal measure of healthcare satisfaction, the Consumer Assessment of Healthcare Providers and Systems has identified major components, including the patient-provider relationship, specificity of care, ability to navigate coverage, and facility environment (Agency for Healthcare Research and Quality).³ These factors influence both the technical aspects of care and dimensions of healthcare access (e.g., availability, affordability, and acceptability), linking healthcare satisfaction directly and indirectly with health status.⁴

Qualitative research has shown that healthcare experiences often differ substantially between gender minority and cisgender populations as a direct result of cissexism and transphobia in medical settings.^5–8 For example, transgender and gender non-conforming people have described receiving care from providers who lack specific gender minority health knowledge;^9–11 being subject to microaggressions, obstruction of care, and explicit gender-based discrimination from providers;^12–16 and having to conceal their true gender identity in order to feel safe and/or listened to.¹⁷ As posited by the Gender Minority Stress Model,^18,19 these stigmatizing and discriminatory experiences likely have a considerable negative impact on healthcare satisfaction for gender minority populations, as they constitute social-structural barriers to receiving quality care.

Despite this qualitative evidence, few studies have collected data of substantive size to provide quantitative, population-based estimates of healthcare satisfaction for transgender and gender non-conforming people relative to cisgender people. Furthermore, no studies to our knowledge have explored potential within-group differences (e.g., between transgender men and women) in healthcare satisfaction. Understanding the relationships between gender identity and healthcare satisfaction is an important first step to improving the delivery of healthcare to gender minority populations, which is central to eliminating gender identity-related health disparities. This study thus aimed to fill these gaps by utilizing nationally-representative data from the Behavioral Risk Factor Surveillance System to quantify differences in healthcare satisfaction among U.S. adults within and between gender identity groups.

Methods

Data Source and Study Population

The Behavioral Risk Factor Surveillance System (BRFSS) is a concatenation of annual cross-sectional, population-based health surveys overseen by the Centers for Disease Control and Prevention (CDC) and conducted by local state, territory, and commonwealth health departments.²⁰ Surveys are conducted in English or Spanish translation, and in some locations, other languages. Eligible individuals include non-institutionalized adults aged 18 years or older residing in the 50 States, the District of Columbia, and the Atlantic and Pacific territories.

The BRFSS consists of a core set of questions included in each jurisdiction’s survey, a set of standard optional modules, and jurisdiction-specific questions. In 2013, the CDC developed a standard optional module on sexual orientation and gender identity (SOGI) for the BRFSS; states had the option to administer this module beginning in 2014,²¹ with the latest data available from 2018. Healthcare satisfaction, our outcome of interest, was also measured within a standard optional module fielded in 2014, 2016, 2017, and 2018. Because of the optional nature of these modules, only 20 states fielded both concurrently in one or more of the years 2014, 2016, 2017, and/or 2018 (228,470 of the 1,839,019 respondents across these four years). We excluded respondents who were not administered the standard optional modules by design (e.g., out-of-state cell phone respondents), declined to answer or answered “don’t know/not sure” on questions about gender identity and/or healthcare satisfaction, and/or reported not having received healthcare in the past 12 months, resulting in a final analytic sample of 167,468 respondents (Figure 1).

Figure 1.

Description of Study Sample Construction

Measures

Gender Identity.

The SOGI module includes two questions, one assessing gender identity and one assessing sexual orientation. The gender identity question has remained constant since its introduction in 2014 and asks: “Do you consider yourself to be transgender?” (If yes, ask: “Do you consider yourself to be male-to-female, female-to-male, or gender non-conforming?”).²¹ The item wording in BRFSS follows the Gender Identity in U.S. Surveillance group’s recommendation for a single item assessment of gender identity,²² and this variable has been used successfully to assess the health status of transgender adults on a multi-state basis.^23–27 We categorized respondents into one of five mutually exclusive categories based on their response to this question and self-reported sex assigned at birth: transgender women (transgender, male-to-female), transgender men (transgender, female-to-male), gender non-conforming people (transgender, gender non-conforming), cisgender women, and cisgender men.

Healthcare Satisfaction.

The healthcare satisfaction item is a single global assessment of satisfaction with healthcare that asks: “In general, how satisfied are you with the health care you received?” with three response options: “Very satisfied”, “Somewhat satisfied”, and “Not at all satisfied”. This item appears sequentially after a series of items asking about healthcare coverage and visits in the past 12 months.²¹ Thus, although not specified in the text of this question, we believe most respondents interpreted this item in light of healthcare experiences in the past 12 months, and we restricted the analysis to respondents who reported a routine checkup in the past 12 months and/or a visit with a healthcare provider in the past 12 months.

Analytic Approach

We first examined the distribution of sociodemographic factors (age group, race/ethnicity, sexual orientation, marital status, and educational attainment), healthcare access and utilization factors (has health insurance, able to afford needed care, has a primary care provider, and had a routine checkup in the past 12 months), general health status, and number of diagnosed chronic conditions (depression, myocardial infarction, chronic heart disease, stroke, cancer, chronic obstructive pulmonary disorder, asthma, arthritis, kidney disease, and diabetes) across the five gender identity groups. We then examined the prevalence of being “very”, “somewhat”, or “not at all” satisfied with their healthcare across these groups, and considered prevalence differences ≥5% to be notable.

We developed multivariable multinomial logistic models to compare the odds of being “somewhat satisfied” and “not at all satisfied” relative to “very satisfied” with healthcare. The three gender minority groups were compared to both cisgender females and cisgender males. We developed a directed acyclic graph (see Supplemental Digital Content Figure 1) to guide covariate selection for these models, and thus included only age group, race/ethnicity, sexual orientation, and educational attainment as adjustment variables. For all analyses, estimates based on 10 or fewer observations are suppressed.

The BRFSS is a hierarchically clustered probability sample with weights that are intended to account for both design factors and survey non-response probabilities. Accordingly, all analyses were weighted using the appropriate clustering variables in proc surveyfreq and proc surveylogistic in SAS 9.4. We multiplied the standard weights for each state with multiple years of data by the proportion of respondents from that state in that year, as recommended by BRFSS.²⁸ Additional details about the weighting scheme are contained in the Supplemental Digital Content.

Sensitivity Analyses

Similar multivariable models were run among a population restricted to those who reported having health insurance, were able to afford needed care, had a primary care provider, and had received a routine checkup within the last 12 months (N = 114,481). This analysis was performed to verify that healthcare access was not the primary driver of any potentially observed relationship between gender identity and healthcare satisfaction. The approach was chosen in lieu of model adjustment or matching because healthcare access is a central aspect of our conceptual model (see Supplemental Digital Content Figure 1), and thus we sought to explore the relationship among those with adequate access to healthcare rather than adjust out the effects of this variable.

Ethical Review

The BRFSS data used in this analysis are de-identified and publicly available (CDC 2020). As such, they are not considered human subjects data.

Results

Relative to cisgender people, gender minority people were more likely to report being bisexual and less likely to report being heterosexual, more likely to have a high school education, and less likely to have completed a 4-year degree (Table 1). Further, transgender men were somewhat more likely to be Hispanic and somewhat less likely to be White, and gender non-conforming people were more likely to report “other”, “don’t know”, or “not sure” to the sexual orientation item. With respect to health status, transgender men and gender non-conforming people were more likely to report being in fair or poor health and less likely to report having no chronic conditions relative to cisgender groups.

Table 1.

Characteristics of Eligible Transgender and Cisgender Behavioral Risk Factor Surveillance Survey Respondents: 2014, 2016, 2017, and 2018.^*

	Transgender			Cisgender
	Men	Women	Gender non-conforming	Men	Women
	(n=244†)	(n=354)	(n=116)	(n=67,102)	(n=99,652)
	%‡	%	%	%	%
Age group
18 to 24	13.3	22.4	28.1	19.7	17.9
25 to 44	16.5	20.5	29.3	22.9	23.5
45 to 64	48.1	39.0	22.2	35.8	34.6
65 to 74	8.1	10.4	8.7	13.4	13.4
75 and older	14.0	7.6	11.7	8.3	10.6
Race/ethnicity§
Hispanic of any race(s)	19.6	6.0	—ǁ	9.2	9.0
Non-Hispanic White	59.0	66.4	73.5	71.7	70.4
Non-Hispanic Black	16.6	19.6	15.2	13.8	15.4
Sexual orientation
Heterosexual	86.4	83.0	68.7	96.0	95.7
Lesbian or Gay	3.3	6.4	7.9	2.3	1.2
Bisexual	6.4	8.8	10.6	1.3	2.6
Something else/don’t know/not sure	—ǁ	—ǁ	—ǁ	0.3	0.5
Marital status
Married/widowed	57.9	55.7	45.4	58.5	60.6
Separated/divorced	15.8	16.4	19.6	12.3	14.2
Single, never married	23.7	24.4	28.0	24.8	21.2
Unmarried partnership	—ǁ	—ǁ	—ǁ	4.1	3.7
Educational attainment
Less than high school	14.8	22.4	11.2	12.1	11.0
High school	49.1	37.0	37.8	31.6	28.9
1–3 years college	28.4	28.7	29.9	28.9	32.9
4+ years college	7.6	11.9	21.1	27.5	27.5
Healthcare access
Has health insurance	79.0	85.4	92.4	90.3	91.6
Able to afford needed care	83.6	80.6	80.3	88.8	86.0
Has a primary care provider	70.8	81.2	83.7	79.8	87.3
Routine checkup in past 12 months	85.6	84.5	77.1	79.0	82.4
All 4 indicators of healthcare access	59.5	53.6	58.2	60.6	65.2
General health
Excellent	11.8	13.9	8.0	17.7	17.5
Very good/good	47.6	63.6	59.4	63.5	63.6
Fair/poor	40.1	22.5	33.6	18.4	18.5
Chronic conditions**
None	34.6	47.3	34.9	49.2	42.5
One	29.9	26.6	35.1	26.6	28.4
Two	17.1	12.2	21.2	13.0	15.6
Three or more	18.5	13.8	8.8	11.2	13.4

^*States with data from both the Sexual Orientation/Gender Identity and Health Care Access modules: Delaware (2014, 2016, 2017), Florida (2017), Georgia (2016), Idaho (2014), Indiana (2014), Iowa (2014), Kentucky (2014, 2016), Louisiana (2014, 2016, 2018), Maryland (2014), Minnesota (2014, 2016), Mississippi (2018), Montana (2014), Nevada (2014), New York (2014), Ohio (2014), Pennsylvania (2014, 2016), Tennessee (2018), Vermont (2014), Virginia (2014), Wisconsin (2014, 2017).

^†Unweighted sample size.

^‡Weighted to reflect sampling probability, non-response, and the summarization across multiple years of data, according to standard Behavioral Risk Factor Surveillance Survey methodology. Column percentages may not sum to 100.0% due to rounding or omitted categories.

^§Percentages for Non-Hispanic American Indian/Alaska Native, Asian, Native Hawaiian/other Pacific Islander, some other single race, and multiracial individuals suppressed due to small cell sizes.

^ǁCell sizes of 10 or fewer suppressed.

^**Ever told by a doctor that the respondent had: depression; myocardial infarction or heart attack; coronary heart disease or angina; stroke; cancer (other than skin cancer); chronic obstructive pulmonary disorder, emphysema or chronic bronchitis; asthma (current); arthritis; kidney disease; or diabetes.

As shown in Table 2, gender non-conforming people were the least likely to report being “very” satisfied with the healthcare they had received (56%, 95% confidence interval (CI): 40%−71%), followed by transgender men (59%, 95% CI: 47%−72%), cisgender men (64%, 95% CI: 63%−65%), cisgender women (67%, 95% CI: 66%, 68%), and transgender women (67%, 95% CI: 57%−77%). Transgender men were the most likely to report being “not at all” satisfied with the healthcare they had received (15%, 95% CI: 3%−26%), followed by transgender women (9%, 95% CI: 1%−16%), cisgender men (4%, 95% CI: 4%−4%), and cisgender women (3%, 95% CI: 3%−3%). The number of gender non-conforming people reporting being “not at all” satisfied was fewer than 10 and thus not reported.

Table 2.

General healthcare by gender identity, overall and among respondents with excellent healthcare access^*: Behavioral Risk Factor Surveillance System, 2014, 2016, 2017, and 2018.

	Very satisfied	Somewhat satisfied	Not at all satisfied
	% (95% CI†)	% (95% CI)	% (95% CI)
Total eligible population	65.5 (65.0–66.1)	30.9 (30.4–31.5)	3.5 ( 3.3– 3.7)
Transgender men	59.4 (46.9–72.0)	25.9 (16.9–35.0)	14.6 ( 2.7–26.5)
Transgender women	66.7 (56.7–76.7)	24.8 (16.2–33.3)	8.6 ( 1.4–15.7)
Transgender, gender non-conforming	55.7 (40.1–71.4)	36.0 (20.8–51.2)	—‡
Cisgender men	64.0 (63.2–64.8)	32.0 (31.2–32.8)	4.0 ( 3.7– 4.3)
Cisgender women	66.8 (66.1–67.5)	30.1 (29.4–30.8)	3.1 ( 2.8– 3.4)
Excellent access to care*	73.4 (72.8–74.0)	25.1 (24.5–25.7)	1.5 ( 1.3– 1.6)
Transgender men	66.1 (49.1–83.1)	20.7 (11.0–30.3)	—‡
Transgender women	73.4 (61.2–85.6)	22.0 (11.8–32.3)	—‡
Transgender, gender non-conforming	59.8 (37.5–82.2)	39.3 (17.0–61.7)	—‡
Cisgender men	71.6 (70.6–72.6)	26.8 (25.8–27.8)	1.7 ( 1.4– 1.9)
Cisgender women	74.9 (74.2–75.7)	23.8 (23.0–24.5)	1.3 ( 1.1– 1.5)

^*Had health insurance, was able to afford needed care, had a primary care provider, and had a routine checkup in the past 12 months.

^†CI=Confidence Interval.

^‡Estimates based on 10 or fewer respondents suppressed.

Among individuals with all four markers of healthcare access (have insurance, able to pay for needed care, have a primary care provider, and had a routine checkup in the last 12 months), gender non-conforming people again were the least likely to report being “very” satisfied with the healthcare they had received (60%, 95% CI: 38%−82%), followed by transgender men (66%, 95% CI: 49%−83%), cisgender men (72%, 95% CI: 71%−73%), transgender women (73%, 95% CI: 61%−86%), with cisgender women being the most likely to report being “very” satisfied with their healthcare (75%, 95% CI: 74%−76%). Too few gender minority respondents reported being “not at all” satisfied with their healthcare to report results on this outcome.

In models adjusted for age group, race/ethnicity, and educational attainment, transgender men were more likely to report being “not at all” satisfied (adjusted odds ratio (aOR) 4.45, 95% CI: 1.72–11.5 relative to cisgender women; aOR 3.40, 95% CI: 1.31–8.80 relative to cisgender men) (Table 3). Transgender women were nearly twice as likely to report being “not at all” satisfied but with wide confidence intervals; aOR 2.17, 95% CI: 0.85–5.57 relative to cisgender women; aOR 1.66. 95% CI: 0.65–4.25 relative to cisgender men. Estimates for gender non-conforming people were imprecise. Transgender men and women appeared to be somewhat less likely to be “somewhat” satisfied than cisgender women and men, although these differences had wide confidence intervals.

Table 3.

Odds ratios of being “somewhat” or “not at all” satisfied with healthcare received in past year, by gender identity: Behavioral Risk Factor Surveillance System, 2014, 2016, 2017, and 2018.

	All eligible respondents		Respondents with excellent healthcare access*
	Somewhat v. very satisfied	Not at all v. very satisfied	Somewhat v. very satisfied	Not at all v. very satisfied
	aOR† (95% CI‡)	aOR (95% CI)	aOR (95% CI)	aOR (95% CI)
Transgender men
v. Cisgender women	0.94 (0.58–1.52)	4.45 (1.72–11.5)	0.96 (0.54–1.73)	—§
v. Cisgender men	0.86 (0.53–1.38)	3.40 (1.31–8.80)	0.81 (0.45–1.45)	—§
Transgender women
v. Cisgender women	0.76 (0.47–1.23)	2.17 (0.85–5.57)	0.91 (0.49–1.69)	—§
v. Cisgender men	0.69 (0.42–1.12)	1.66 (0.65–4.25)	0.76 (0.41–1.42)	—§
Transgender, gender non-conforming
v. Cisgender women	1.36 (0.69–2.70)	—§	1.90 (0.74–4.90)	—§
v. Cisgender men	1.23 (0.62–2.44)	—§	1.59 (0.62–4.11)	—§

^*Had health insurance, was able to afford needed care, had a primary care provider, and had a routine checkup in the past 12 months.

^†aOR=adjusted Odds Ratio, adjusted for age group, race/ethnicity, sexual orientation, and educational attainment.

^‡CI=Confidence Interval.

^§Estimates based on 10 or fewer respondents suppressed.

In analyses restricted to individuals with all four markers of healthcare access, the odds for being “somewhat” satisfied were similar between transgender men and women and their cisgender peers. It should be noted that the proportion of transgender women (54%), and gender non-conforming people (58%) with such excellent healthcare access is lower than the proportion of cisgender women (65%) and men (61%) (Table 1).

Discussion

Main Findings

We found that healthcare satisfaction varied considerably by gender identity among a nationally-representative sample of U.S. adults. Specifically, transgender men and gender non-conforming people were least likely to be satisfied with their healthcare, while cisgender women were most likely to be satisfied, and there were more polarized differences in healthcare satisfaction by sex assigned at birth between transgender men and women relative to cisgender men and women. These observed differences were not explained by sociodemographic factors (age group, race/ethnicity, and educational attainment). Notably however, levels of healthcare satisfaction were similar between all gender identity groups within the sub-group of respondents who had reported markers of good healthcare access (i.e., having insurance, a primary care provider, the ability to pay, and a recent routine physical).

These findings are consistent with several possible interpretations consistent with the gender minority stress model, which posits that transgender and gender minority populations face substantial social stressors as a result of cissexism and transphobia and that such stressors shape health and healthcare-related experiences.¹⁸ Many gender minority individuals seek gender-affirming pharmaceutical and surgical procedures, which often necessitate interaction with medical gatekeepers.^27,29 These populations also experience frequent and substantive discrimination, rejection, and lack of clinical competence in healthcare delivery.^30,31 As such, gender minority people who chose to medically transition may put substantial effort into identifying affirming and supportive providers, potentially increasing levels of healthcare satisfaction. However, our analyses also revealed that while many transgender people are able to identify care providers with whom they are “very” satisfied, a disturbingly high proportion (15% of transgender men and 8% of transgender women) were “not at all” satisfied with their healthcare. Furthermore, our sensitivity analysis revealed that levels of healthcare satisfaction were similar between gender identity groups among respondents with all four markers of healthcare access. Thus, findings from our main analysis may also indicate that healthcare satisfaction is strongly linked to healthcare access among gender minority populations, perhaps reflecting the role of gender-affirming care in reducing gender dysphoria and/or risk of experiencing physical violence for gender minority populations.

A novel finding from our analyses is that transgender men and women reported lower levels of healthcare satisfaction relative to both cisgender men and women. These differences likely stem from considerable differences in healthcare experiences between transgender men and women, including dramatic differences in transition procedures for transgender men versus transgender women, as well as stigma around and cultural erasure of the specific medical needs of some transgender men (e.g., reproductive care).³² Importantly, the majority of extant research groups all gender minority respondents into a single category (often due to inadequate sample sizes),^33–38 and the few studies that do distinguish transmasculine and transfeminine healthcare experiences have been underpowered;³⁹ further research is needed to elucidate this important relationship.

Alternate Explanations

The fact that gender minority respondents with excellent healthcare access reported comparable levels of being “very” satisfied with their healthcare, despite evidence of widespread implicit and explicit bias against gender minority populations,^30,31 requires explanation. Consistent with the gender minority stress model,^18,19 gender minority individuals may internalize some degree of unjust or deprecating treatment as an expected norm, and thus report comparable levels of satisfaction despite objectively less satisfactory care. It is also possible that some gender minority respondents did not come out as transgender or gender non-conforming to their providers in expectation of disparate care; future work exploring within-group differences by gender expression and degree of disclosure is warranted.

Limitations

The BRFSS includes only one item about healthcare satisfaction, intended to broadly capture generic satisfaction with care. Thus, it is possible that gender minority respondents may not have interpreted the item to address aspects of healthcare relevant to their gender care, and these aspects of healthcare delivery might well show more dramatic differences than those found in this study. The survey also contains only one item assessing gender identity, with limited response options, and no items regarding gender expression. Thus, it was not possible to assess whether gender minority respondents who are open about their identity experience greater or lesser degrees of healthcare satisfaction than those who elect not to discuss these facets with their healthcare providers. This is the first study of our knowledge to quantify differences in healthcare satisfaction between gender identity groups using a large, nationally-representative sample.

Conclusions

Gender minority populations experience a range of stigma and discrimination-based encounters in their interactions with the healthcare system, including with their providers, which likely negatively impacts healthcare satisfaction and have downstream effects on health status. As a result of structural-level cissexism and transphobia, gender minority patients often face financial burdens that limit their ability to afford gender affirming procedures, further exacerbating the stress of navigating the healthcare system.²⁹ This is especially true for patients with socially and medically stigmatized illnesses such as HIV, as well as for those with additional historically marginalized social identities.^1,29,31 Additionally, there is substantial evidence that healthcare providers fail to maintain an adequate knowledge of gender minority health literature and practice (likely due to the systemic exclusion of this topic from standard medical training) and/or refuse to provide care to gender minority populations.^40–43

Ultimately, our findings are indicative of a need to improve the way healthcare is delivered to gender minority populations. Qualitative research with gender minority groups provides important insights into elements of healthcare delivery that improve perceived quality and thus satisfaction, including more inclusive/comprehensive care,³³ provider knowledge of gender minority health,^10,34 respect for patients,¹⁶ willingness to make referrals,³⁴ and receiving care from a center that specializes in gender minority health.^33,35 Encouragingly, demonstrated improvements have been achieved in clinics and health systems that ensure their providers apply up to date knowledge of transgender health and maintain a culturally sensitive and welcoming environment for all patients,⁴⁰ suggesting that efforts should be directed towards addressing gaps in medical education (in addition to advocacy around broader social-structural stigma). Specifically, increasing course time spent on sexual and gender minority health beyond the average of five hours has been shown to improve provider knowledge of the subject area. Other educational approaches have been used including elective and mandatory lunchtime lectures, clinical observations, and online educational modules.⁴⁴ Research is needed to assess the long-term effects of such interventions on healthcare satisfaction among gender minority groups, as well as to explore the links between improving healthcare satisfaction and reducing health disparities. Similar experiences and disparities in care and satisfaction have been observed among racial and ethnic minorities within the U.S. healthcare system.⁴⁵ It is likely that these experiences intersect to shape healthcare satisfaction for groups at the nexus of more than one marginalized identity (e.g., gender and racial/ethnic minorities),⁴⁶ and thus it will also be essential for future research to investigate how to optimally design and deliver such medical education interventions that address the multiple sources of stigma and discrimination groups may face.