Lesbian, Gay, And Bisexual Adults Report Continued Problems Affording Care Despite Coverage Gains

Abstract

The uninsurance rate among lesbian, gay, and bisexual (LGB) adults has dropped since the Affordable Care Act (ACA) and legalization of same-sex marriage. Less is known about whether disparities in access to care and health outcomes have narrowed in LGB adults compared to their straight peers in the post-ACA era.We used data from three waves of the Behavioral Risk Factor Surveillance System to examine access to a personal doctor, affordability of care, type of health insurance coverage, and self-reported health in LGB adults in the period January 2014–February 2017 in thirty-one states that implemented the system’s sexual orientation module. Compared to straight adults, more LGB adults reported avoiding necessary care because of cost and worse self-reported health outcomes, even if they had health insurance. More LGB adults reported having individually purchased insurance, which suggests that the repeal of the ACA’s individual mandate may create challenges in the affordability of necessary care.

About 3.5 percent of Americans self-identify as lesbian, gay, or bisexual (LGB).¹ Relative to their straight peers, LGB adults report having significantly more difficulty accessing health insurance coverage² and finding specialty care and forgoing necessary prescription drugs.^3,4 Low-income LGB adults and LGB people of color report additional challenges, including less adequate health insurance coverage⁵ and a higher likelihood of using the emergency department as a usual source of care.⁶ LGB adults are more likely to report worse health, more behavioral risk factors,⁷ and worse mental health outcomes, the latter largely attributed to minority stress and discrimination promulgated by societal stigma.⁸

Recent policies have attempted to reduce disparities in access to care, particularly insurance coverage. Among these are provisions of the Affordable Care Act (ACA), including the expansion of Medicaid coverage, the availability of individually purchased coverage, and a prohibition of denial of insurance because of preexisting conditions.⁹ For LGB adults, the ACA offered the potential to reduce important financial and medical barriers. Notably, preexisting depression and HIV would no longer hinder their enrollment in private insurance.

Furthermore, in 2015 the US Supreme Court ruling in Obergefell v. Hodges legalized same-sex marriage nationwide. Before the ruling, same-sex couples could only obtain spouse coverage as domestic partners if their employers offered such benefits.⁴ As of 2016 approximately 40 percent of employers overall—and 85 percent of large employers (firms with 200 or more employees) —that offered coverage to opposite-sex couples also offered it to same-sex spouses.⁴ Studies suggest that legalization of same-sex marriage may reduce societal stigma against LGB adults and discrimination by providers¹⁰ and improve self-reported mental health. Conversely, states’ institution of constitutional bans on same-sex marriage in 2004 and 2005 was associated with significant increases in the prevalence of psychiatric morbidity in LGB adults.¹¹

Recent studies indicate that disparities in uninsurance between LGB and straight Americans have narrowed since the implementation of the ACA and legalization of same-sex marriage.^9,12 Less is known about whether these trends have been accompanied by narrowed disparities in access to care or health outcomes several years into the post-ACA era. We examined access to care (having health insurance, having access to a personal doctor, having had a routine checkup, and affordability of care) and self-reported health outcomes for LGB adults relative to their straight peers in the period January 2014–February 2017. We also studied the role of health insurance coverage and primary type of health insurance in explaining differences in outcomes between LGB and straight adults.

Study Data And Methods

DATA AND STUDY SAMPLE

The Behavioral Risk Factor Surveillance System (BRFSS) is a nationally representative, random-digit-dialed telephone survey administered annually by the Centers for Disease Control and Prevention (CDC). All responses are weighted to account for nationwide differences in the distribution of sociodemographic characteristics, including age, race/ethnicity, education level, and marital status.¹³ The data we studied were collected in three survey waves during the period January 2014-February 2017.

Our outcomes were having insurance coverage, having a personal doctor, having had a routine checkup in the past year, having avoided necessary medical care because of cost, the number of bad mental health days (for example, stress, depression, and problems with emotions) in the past thirty days, and the number of days in the same period when poor physical or mental health kept the respondent from doing usual activities (such as self-care, work, or recreation).

In 2014 the CDC added an optional module to the BRFSS to enable states to collect data on sexual orientation, making it one of the few federally administered surveys to collect data on sexual orientation and health insurance coverage among nonmarried adults with multiple years of data. In the period 2014–17, thirty-one states implemented the module for at least one survey year. It asks respondents to report their sexual orientation as “straight,” “lesbian or gay,” “bisexual,” “other,” or “don’t know/not sure,” and respondents may refuse to answer. We categorized respondents as LGB (“lesbian or gay” or “bisexual”) or “straight” based on self-reported orientation, excluding people with any other responses.We conducted a sensitivity analysis that included respondents who answered “other” or “don’t know/not sure” or who refused to answer as an additional comparison group.

Though the umbrella term LGBT, which includes transgender people, is often used to describe the broader population, transgender people face different challenges in accessing health care, relative to their cisgender LGB peers.¹⁴ For example, even after ACA implementation, some gender-affirming health services are not covered by public or private health plans.⁴ We therefore focused our analysis on sexual orientation and adjusted our models for respondents who self-identified as transgender.

Online appendix exhibit A1 shows the construction of our study sample.¹⁵ We focused on nonelderly US adults (ages 18–64), given our questions surrounding access to insurance coverage and care. Adults ages 65 and older typically qualify for Medicare and likely have different experiences.

ANALYSIS

We used Pearson’s chi-square tests to describe differences between LGB and straight nonelderly adults on demographic characteristics and health status.We then estimated logistic regression models across four access to care outcomes and estimated linear regression models across two self-reported health outcomes. As a sensitivity analysis, we also estimated Poisson regression models for the self-reported health outcomes. Our adjusted analyses aimed to isolate the association of identifying as LGB with access to care and health outcomes. We therefore controlled for several sociodemographic, clinical, and policy covariates that either are associated with our outcomes or are key differences across sexual orientation samples.^9,16 All models adjusted for age, sex, identifying as transgender, race/ethnicity, primary language, highest level of education attained, annual household income, insurance coverage (excluded in “uninsured” outcome), marital status, having children, number of chronic conditions, self-reported health status, living in a Medicaid expansion state, and state and year fixed effects.

To examine whether having health insurance coverage mediated the relationship between sexual orientation and our outcomes, we stratified our models by insurance status. We accounted for nationwide differences in the distribution of sociodemographic characteristics by applying survey weights to all analyses. All analyses were conducted in Stata, version 15.

LIMITATIONS

Our study had several limitations. First, it relied on self-reported survey data in terms of sexual orientation, insurance coverage, and health outcomes. Data on sexual orientation were available only for the thirty-one states that implemented the optional module for at least one survey year in 2014–17. However, these states contain about 70 percent of the US population.¹⁷ Additionally, sexual orientation data were not collected in the BRFSS before 2014, which therefore limited our ability to examine trends before the implementation of the ACA and the legalization of same-sex marriage. Some correlates of access to care either were unavailable in the BRFSS (for example, the availability of transportation) or were heavily missing (such as an urban/rural indicator).¹⁸ Data on primary type of health insurance coverage are not collected annually and constitute an optional module in the BRFSS. However, our estimates contribute new data, as few studies have examined type of insurance coverage in the LGB population in the post-ACA era.

Second, the sample size for nonelderly adults identifying as LGB is relatively small, which limited our ability to robustly assess intra-LGB heterogeneity for certain subgroups or through a more intersectional lens. We attempted to address sample-size issues by pooling data from three waves of the BRFSS and by grouping lesbian, gay, and bisexual respondents together, consistent with previous literature.^7,19 Prior studies have suggested that there are important differences in socioeconomic status,²⁰ health care access, and health outcomes^19,21 when data on lesbian women, gay men, and bisexual people are examined separately. However, our sample size did not permit us to conduct these stratified analyses.

Study Results

Our unweighted sample consisted of 329,658 nonelderly adults who self-identified as LGB or straight (exhibit 1). In our weighted sample, 4.3 percent of respondents self-identified as LGB (data not shown).

EXHIBIT 1.

Selected characteristics of nonelderly adults, by lesbian, gay, and bisexual (LGB) status, 2014–17

	Straight (n = 317,680)	LGB (n = 11,978)
Age range (years)
18–34	34.2%	55.9%
35–44	20.4	15.2
45–54	22.9	16.7
55–64	22.6	12.1
Female	50.4	54.2
Transgender	0.4	3.0
Race/ethnicity
White	63.9	62.3
Black	11.9	12.8
Hispanic/Latino	14.9	14.1
Asian	5.4	5.1
American Indian/Alaska Native	0.6	1.1
Other	1.8	3.4
Don’t know/not sure/refused	1.5	1.2
Primary language not English	6.1	2.6
Employment status
Employed	68.1	61.5
Not in workforce	24.8	28.5
Unemployed	7.1	10.0
Education
Some high school	12.3	11.7
High school graduate	28.6	26.7
Some college	31.4	34.8
College graduate or more	27.4	26.5
Don’t know/not sure/missing	0.2	0.2
Annual household income
Less than $15,000	8.7	12.4
$15,000–$24,999	12.8	16.8
$25,000–$34,999	8.1	8.7
$35,000–$49,999	11.6	12.0
$50,000–$74,999	13.7	10.9
$75,000 or more	32.6	24.7
Don’t know/not sure/refused	12.6	14.4
Has insurance	87.2	86.8
Primary source of coveragea
Employer-sponsored	66.6	55.1
Individually purchased	11.4	13.3
Medicaid or other state program	9.4	14.9
Otherb	12.6	16.7
Marital status
Married	52.0	21.6
Never married	27.3	54.1
Divorced, widowed, or separated	15.4	11.3
Unmarried couple	4.9	12.3
Refused/missing	0.5	0.7
Self-reported fair or poor health	15.2	18.4

SOURCE Authors’ analysis of pooled data for 2014–15, 2015–16, and 2016–17 from the Behavioral Risk Factor Surveillance System. NOTES All differences between straight and LGB respondents (based on Pearson’s chi-square tests) in each category were significant (p < 0.01) except for “has health insurance” (p = 0.51).

^aRespondents were not required to give this information, which was collected in 2014–15 and 2016–17. The unweighted sample was 89,026 straight and 2,932 LGB adults.

^bMedicare, TRICARE, Department of Veterans Affairs, military, or Indian Health Service.

Comparable rates of LGB and straight adults reported having health insurance (86.8 percent and 87.2 percent, respectively; exhibit 1). We compared the distribution of self-reported primary source of health insurance coverage for LGB and straight adults in 2014 and 2016. Given that states did not have to ask the question, data were available for approximately 27.9 percent of our unweighted sample. More LGB adults (13.3 percent) than straight adults (11.4 percent) reported having an individually purchased plan.

In unadjusted analyses, we found that similar rates of LGB and straight adults reported being uninsured (13.2 percent and 12.8 percent, respectively) and having had a routine checkup within the past year (64.9 percent and 66.5 percent, respectively) (exhibit 2). However, LGB adults were significantly less likely than straight adults to report having a personal doctor (74.7 percent versus 77.3 percent) and significantly more likely to report having avoided necessary medical care because of cost (20.0 percent versus 14.1 percent). LGB adults reported nearly twice as many bad mental health days (7.1 versus 3.9 days) and more days when poor physical or mental health limited their ability to do usual activities (5.1 versus 4.4 days) in the past thirty days.

EXHIBIT 2.

Unadjusted and adjusted associations of access to care and self-reported health status, by lesbian, gay, and bisexual (LGB) status, 2014–17

	Unadjusted, predicted value			Adjusted, predicted value
	Straight (ref)	LGB	p value	Straight (ref)	LGB	p value
ACCESS TO CARE
Uninsured	12.8%	13.2%	0.51	12.9%	12.4%	0.35
Has personal doctor	77.3	74.7	<0.01	77.2	77.3	0.94
Had checkup within 1 year	66.5	64.9	0.05	66.5	66.1	0.61
Avoided receiving care because of cost in past year	14.1	20.0	<0.01	14.2	16.4	<0.01
SELF-REPORTED HEALTH STATUS (DAYS WITHIN PAST 30 DAYS)
Bad mental healtha	3.9	7.1	<0.01	3.9	5.6	<0.01
Limited usual activitiesb	4.4	5.1	<0.01	4.9	5.3	<0.01

SOURCE Authors’ analysis of pooled data for 2014–15, 2015–16, and 2016–17 from the Behavioral Risk Factor Surveillance System. NOTES We estimated logistic regression models for access to care outcomes and linear regression models for self-reported health status outcomes. The adjusted models were adjusted for age, sex, identifying as transgender, race/ethnicity, primary language, education, annual household income, insurance coverage (except for “uninsured” outcome), marital status, having children, number of chronic conditions (heart attack or myocardial infarction; angina or coronary heart disease; stroke; asthma; cancer; chronic obstructive pulmonary disease, emphysema, or chronic bronchitis; arthritis, rheumatoid arthritis, gout, lupus, or fibromyalgia; depressive disorder; kidney disease; and diabetes), living in a state that expanded eligibility for Medicaid, and state and year fixed effects. Predicted values were obtained using the MARGINS command in Stata. Observations (no more than 0.3 percent of our sample) when strata had a single sampling unit were dropped.

^aNumber of bad mental health days (defined in the text).

^bDays when poor physical or mental health limited the respondent’s ability to do usual activities.

After adjusting for sociodemographic factors and health status, we found that comparable rates of LGB and straight adults reported being uninsured (12.4 percent and 12.9 percent, respectively), having a personal doctor (77.3 percent and 77.2 percent, respectively), and having had a routine checkup in the past year (66.1 percent and 66.5 percent, respectively). However, significantly more LGB adults still reported having avoided necessary medical care because of cost (16.4 percent versus 14.2 percent), more bad mental health days (5.6 versus 3.9), and more days when poor physical or mental health limited their ability to do usual activities (5.3 versus 4.9) in the past month.

In terms of magnitude and significance, the association between our outcomes and identifying as LGB was similar when we included an additional comparison group of respondents who answered “other” or “don’t know/not sure” or who refused to answer the question about sexual orientation in our adjusted models. However, the comparison group reported having significantly fewer days with limited usual activities, relative to the straight respondents. Unadjusted and adjusted estimates of self-reported health outcomes using Poisson regression models were comparable.

In our adjusted analyses, uninsured respondents (both LGB and straight) were significantly less likely to have a personal doctor and to have had a checkup in the past year. However, they were more likely to have avoided getting necessary medical care because of cost and had more days with limited usual activities.When we stratified our models by health insurance coverage status, our unweighted sample consisted of 298,377 adults with health insurance coverage (exhibit 3), of whom 10,656 identified as LGB. The 1,262 people (0.38 percent of the sample) who responded “don’t know/not sure” or refused to respond to the insurance coverage item were excluded from this analysis. Among the insured, LGB and straight adults had comparable rates of having a personal doctor (82.7 percent and 82.5 percent, respectively) and having had a checkup within the past year (69.9 percent and 70.3 percent, respectively). And LGB adults were significantly more likely than straight adults to avoid getting necessary medical care because of cost (12.5 percent versus 10.4 percent). Insured LGB adults also reported significantly more bad mental health days (5.3 versus 3.9) and more days with limited usual activities (5.2 versus 4.9). Among the uninsured, most results were consistent, though the estimated proportions were larger, and there was no significant difference between LGB and straight adults in avoiding necessary medical care because of cost (42.0 percent and 39.9 percent, respectively).

EXHIBIT 3.

Adjusted associations of access to care and self-reported health status with lesbian, gay, and bisexual (LGB) status, stratified by insurance status, 2014–17

	Insured (n = 298,377), predicted value			Uninsured (n = 30,019), predicted value
	Straight (ref)	LGB	p value	Straight (ref)	LGB	p value
ACCESS TO CARE
Has personal doctor	82.5%	82.7%	0.84	41.0%	40.1%	0.71
Had checkup within 1 year	70.3	69.9	0.63	40.7	40.1	0.80
Avoided receiving care because of cost in past year	10.4	12.5	<0.01	39.9	42.0	0.34
SELF-REPORTED HEALTH STATUS (DAYS WITHIN PAST 30 DAYS)
Bad mental healtha	3.9	5.3	<0.01	4.4	7.3	<0.01
Limited usual activitiesb	4.9	5.2	<0.01	5.0	6.0	<0.01

SOURCE Authors’ analysis of pooled data for 2014–15, 2015–16, and 2016–17 from the Behavioral Risk Factor Surveillance System. NOTES We estimated logistic regression models for access-to-care outcomes and linear regression models for self-reported health status outcomes. Adjusted models are explained in the notes to exhibit 2. Predicted values were obtained using the MARGINS command in Stata. Observations (no more than 0.3 percent of our sample) when strata had a single sampling unit were dropped.

^aNumber of bad mental health days (defined in the text).

^bDays when poor physical or mental health limited the respondent’s ability to do usual activities.

Discussion

Previous studies indicate that the uninsurance rate in the LGB population has declined since the implementation of the ACA and the legalization of same-sex marriage.^7,12 Our findings suggest not only that LGB and straight adults’ uninsurance rates are comparable in the post-ACA era, but also that LGB adults report similar rates of having a personal doctor and having had a regular checkup within the past year, compared to their straight peers. Nevertheless, even after adjusting for sociodemographic factors and health status, we found that significantly more insured LGB adults reported avoiding necessary medical care because of cost and worse health outcomes, compared to their straight peers.

It is likely that the availability of health insurance through the ACA coverage expansions and LGB-specific outreach efforts during open enrollment reduced uninsurance rates in LGB adults to levels comparable to those of their straight peers. In 2013 a coalition of organizations formed Out2Enroll for LGB-specific targeted outreach and enrollment efforts following implementation of the ACA Marketplaces, though there was wide variation across states in the level of effort.²² Given the lack of a question about sexual orientation in the BRFSS at that time, we could not calculate a pre-ACA uninsurance rate for the LGB population. However, uninsurance rates in 2014–17 were 12.8 percent in the straight population and 13.2 percent in the LGB population, while uninsurance rates (not differentiated) in the BRFSS for 2012–13 were 18.3 percent.²³

Studies using data that predate the ACA have reported that fewer LGB adults had a usual place of care before the ACA’s implementation.^4,12 Our results suggest that in the post-ACA era, comparable rates of LGB and straight adults report having a personal doctor, which is important as many LGB adults report both a scarcity of providers who are competent to address LGB-specific health care needs and discrimination based on their sexual orientation.^3,14

Our finding that, compared to straight adults, LGB adults report higher rates of avoiding necessary medical care because of cost is consistent with the results of a recent study using data for 2013–15 from the National Health Interview Survey.⁶ Even when we limited results to people with insurance, we observed significant differences between LGB and straight adults, which suggests that insurance coverage might not be the only driver of differences in avoiding medicallynecessary care because of cost. One potential explanation is that forgoing care may reflect the limited generosity or adequacy of health insurance coverage taken up by LGB adults,⁵ given the proportion of LGB adults who reported having individually purchased coverage. A previous study indicated that, among those who reported having individually purchased coverage in 2014, LGB adults had greater dissatisfaction health care than their straight peers,²⁴ and the author suggested that LGB adults’ greater health care needs led to higher out-of-pocket spending, particularly given the trends toward higher premiums and deductibles of plans in the nongroup insurance market.

Though many of the previously documented gaps in access to care between LGB and straight adults have narrowed in the post-ACA era, LGB adults still report having significantly more bad mental health days and days when their physical or mental health limited their ability to participate in usual activities relative to their straight peers. Both insurance coverage and legalization of same-sex marriage have been associated with improved psychological well-being and rapidly improved self-reported mental health.^25,26 Nonetheless, we still observed significantly more bad mental health days among LGB adults. Factors specific to LGB adults that could explain this disparity include internalized stigma, discrimination, the political environment, and the limited availability of LGB behavioral health specialists.^8,18

Policy Implications

Despite previously documented gains in health insurance coverage,¹² LGB adults continue to face challenges in affording care. Given that compared to straight adults, LGB adults are less likely to have employer-sponsored insurance and more LGB adults report having individually purchased insurance, the repeal of the individual mandate—and anticipated increases in monthly premiums—may have negative consequences for LGB Americans.

Social policy has been associated with changes in mental health among LGB adults.^10,11 Rhetoric during the 2016 campaign season was associated with a substantial increase in the volume of calls to LGBT suicide hotlines, particularly immediately after the November 2016 election.²⁷ Furthermore, religious protections were recently expanded to medical professionals, which is concerning to LGB adults who fear discrimination or denial of care.²⁸ The administration of President Donald Trump has also proposed removing sexual orientation questions from federal surveys, including the National Survey of Older Americans Act Participants,²⁹ and it has not yet been confirmed whether the optional sexual orientation module will be included in the finalized 2019 BRFSS.³⁰ Such actions may limit the understanding of health disparities among LGB adults in the future.

Conclusion

In the post-ACA era, LGB adults reported similar rates of insurance coverage and comparable access to a personal doctor and routine checkups, relative to their straight peers. However, LGB adults were still significantly more likely to report avoiding necessary medical care because of cost, even if they had insurance. Irrespective of insurance coverage, LGB adults consistently reported worse mental health outcomes and more days when physical or mental health limited their participation in usual activities.