Health Disparities, Inpatient Stays, and Emergency Room Visits Among Lesbian, Gay, and Bisexual People: Evidence From a Mental Health System

There is a growing emphasis on addressing specific mental health needs among gender and sexual minority communities, both nationally and internationally (1, 2). A recent literature review on the service experiences of lesbian, gay, bisexual, transsexual, and transgender (LGBT) people with severe mental illness suggested that there are distinct health care needs in this growing population (3). Despite the importance of this issue, research has been limited, mainly because of difficulties delineating LGBT populations, issues measuring stigmatization and discriminatory practices as they relate to this population, and challenges related to generating representative samples in population-based studies.

A meta-analysis by Meyer (4) confirmed that the prevalence of mental disorders is higher among lesbian, gay, and bisexual (LGB) people than among heterosexuals. More specifically, compared with heterosexuals, gay men and lesbians experience more mental health problems—including substance use disorders, affective disorders, posttraumatic stress disorder, and suicide (5–9). Some researchers have speculated that the higher prevalence of mental disorders among LGB people stems from stigma, prejudice, and discrimination. These factors create a stressful social environment that leads to mental health problems in stigmatized minority groups, contributing to a condition that has been termed “minority stress” (10, 11). For example, compared with heterosexual adults, gay men and lesbians are about 2.5 times more likely to have had a mental disorder at any point over their lifetimes (4).

In 2007, Burgess et al. (12) investigated the association between discrimination incidence in the past year and LGB utilization of mental health services. Results indicated that compared with heterosexuals, LGB individuals reported poorer mental health and higher perceived need for mental health services. In addition, research indicates that LGB individuals experience increased prejudice in the form of higher rates of harassment and employment discrimination (13).

In terms of overall mental health estimates, only 43% of adults in the United States with a mental health condition received mental health services in the past year (14). Among adults with a serious mental illness, 65% received mental health services in 2016.

In summary, the literature indicates that although the LGB population often presents with more mental and general medical issues, it remains unclear whether there are differences in health care or service use between the LGB population and heterosexual adults. This study examined the potential differences in use of acute services between LGB individuals and individuals in a matched heterosexual control group within the New York State mental health system. The study utilized the Gelberg-Andersen behavioral model for vulnerable populations (15) to evaluate acute care use among LGB individuals. The model suggests that service use is a function of predisposing characteristics that exist before the onset of illness, factors that enable service use, and an individual’s need for health care. The major research questions of this study included the following: Compared with heterosexuals, what was the health profile of LGB individuals? Relative to heterosexuals, what was the likelihood among LGB individuals of ever using an inpatient or emergency service, and on average, how frequent were emergency room (ER) visits, and how long were inpatient stays? On the basis of our research questions, we predicted that compared with heterosexuals, LGB individuals would have lower rates of having ever used an inpatient or emergency service, fewer ER visits, shorter inpatient stays, and in general more severe health profiles.

Methods

Study Setting

This was a retrospective, matched-control study using data from the 2015 New York State Office of Mental Health (NYSOMH) Patient Characteristics Survey (PCS) and Medicaid. The PCS is a 1-week survey of all patients served by the public mental health system in New York State. Conducted every 2 years, the PCS collects demographic, clinical, and socioeconomic data from over 4,000 NYSOMH-licensed or -funded programs providing direct services to approximately 180,000 patients. The PCS is a Web-entry application that allows reporting providers to electronically load, import, or enter information from existing electronic health record (EHR) systems. Data are not self-reported by patients. Providers collect and submit information about patients served during the survey week. Patients can have any type of insurance coverage, including Medicaid, Medicare, private insurance, Child Health Plus, self-pay, and other health insurance.

The study population was limited to individuals between the ages of 18 and 85 at the time of the survey week (October 19–25, 2015) with Medicaid eligibility between October 2014 and October 2015. Individuals whose EHR indicated a sexual orientation as gay, lesbian, or bisexual were matched with heterosexual adults. Individuals with an unknown sexual orientation were excluded.

Measures

Except for acute service use, all measures used in this study were based on data gathered from provider EHRs through the PCS. The PCS was used to identify the study groups and individual characteristics and factors that are known to influence service use (15, 16). These characteristics and factors include predisposing characteristics, which are presumed to be present before the onset of illness; enabling characteristics, the factors known to facilitate access to health services; and need variables, such as the type and severity of health condition and indicators of risky behavior (17).

Predisposing characteristics.

The predisposing variables used in this study were sex at birth, sexual orientation, age, race, and Hispanic ethnicity. Five age groups were created: 18–24, 25–34, 35–44, 45–54, and 55–85. In the matched-control analysis, individuals’ age was used. Race and Hispanic ethnicity were categorized as white non-Hispanic, black non-Hispanic, Hispanic, and other–multiracial.

Enabling characteristics.

Individual enabling characteristics were categorized as follows: education level (less than high school, high school, some college/business/vocation, and college), employment status (paid employment, nonpaid work, unemployed, and not in the labor force), cohabitation (lives with others and lives alone), homeless (yes and no), parental status (no children; children under 18 in one’s custody; children under 18, but not in one’s custody; children over 18; and multiparental status [more than one choice]), preferred language (English and not English), criminal justice involvement (yes and no), and region of residence (Western, Central, Hudson River, New York City, and Long Island). In the matching analysis, individuals’ county of residence was used. Two categories for cash assistance benefits were established: public cash assistance (yes and no) and Supplemental Security Income/Social Security Disability Insurance (yes and no).

Need variables.

The need variables were based on diagnoses or a significant disability as indicated in the EHR system. Chronic medical condition was defined as having at least one of the following conditions: hyperlipidemia, high blood pressure, diabetes, obesity, heart attack, stroke, other cardiac condition, pulmonary condition, Alzheimer’s disease or dementia, kidney disease, liver disease, endocrine condition, progressive neurologic disease, traumatic brain injury, joint and connective tissue disorder, cancer, or other chronic physical condition (see online supplement). The following were characterized as present or not present: physical disability (hearing, vision, or mobility), primary psychiatric disorders (schizophrenia spectrum and other psychotic, depressive, bipolar, anxiety, and other disorders), developmental disability, intellectual disability, and autism spectrum disability. Serious mental illness refers to having one or more diagnoses of mental disorders combined with significant functional impairment. In addition, three indicators of risky behavior were established (tobacco use, alcohol-related disorder, and substance use–related disorder, all categorized as yes and no).

Acute service use.

Acute services were assessed by using Medicaid claims and encounters for ER visits or inpatient stays in the 12 months preceding the PCS week. Mental health–related inpatient stays comprised psychiatry-related stays at licensed general hospitals, mental health clinics, and state-operated psychiatric centers. Similarly, inpatient stays for substance use disorders comprised detox, rehabilitation, and other substance use–related stays. Nonbehavioral inpatient stays comprised inpatient stays that did not meet the criteria described above for stays related to treatment of mental and substance use disorders.

Mental health–related ER use comprised psychiatry-related visits to general hospitals and comprehensive psychiatric emergency programs, whereas substance use–related ER use comprised substance use–related visits. Nonbehavioral ER use comprised visits that were related neither to mental health nor substance use. Inpatient and ER services were categorized by using a combination of rate codes, DRG codes, specialty codes, procedure codes, revenue codes, diagnosis of mental or substance use disorders, and hospital provider type (see online supplement). The primary outcome was having used an inpatient or ER service in the past 12 months. Secondary outcomes included length of inpatient stay and frequency of ER visits.

Statistical Analysis

A matched case-control approach was used to control for potential confounding and to form groups sufficiently balanced to provide efficient statistical analysis. The matching of LGB individuals was performed by using the PROC SURVEYSELECT procedure from SAS, as outlined elsewhere (18). LGB individuals were matched 1:1 to heterosexual adults on exact values of assigned sex at birth, age, race, Hispanic ethnicity, and county of residence. In this procedure, matched LGB individuals and heterosexual adults were selected by using a simple random sampling without replacement from independent samples that were stratified on the matched criteria.

To examine the differences in acute service use between LGB individuals and heterosexual adults, crude and adjusted odds and rate ratios were calculated by using a generalized estimating equation (GEE) model (accounting for matching pairs) with a binomial or negative binomial distribution and an independent working correlation structure (19, 20). For binary outcomes (having used or not having used a service), a GEE logistic regression model was used. For count outcomes (number of inpatient days or ER visits), a GEE negative binomial regression model was used. The models were fitted for each LGB group, with adjustment for employment status, education, cohabitation, homelessness, tobacco use, alcohol use–related disability, substance use–related disability, physical disability, chronic medical conditions, and parental status. These variables were maintained in the adjusted models regardless of statistical significance. Odds ratios (ORs) and rate ratios (RRs) with 95% confidence intervals are presented. The study protocol was approved by the Nathan S. Kline Institute Institutional Review Board with a full waiver of informed consent. All statistical analyses were performed with SAS software, version 9.4.

Results

Population Characteristics

The 2015 PCS collected information on 179,096 individuals. Overall, for 9,683 (5.4%) individuals, the EHR indicated sexual orientation as gay, lesbian, bisexual, or other. Of these, 2,579 (26.6%) were gay men, 1,062 (11.0%) were bisexual men, 2,469 (25.5%) were lesbians, 3,095 (32.0%) were bisexual women, and 478 (4.9%) were persons with other sexual orientation. The matched cohort included 5,775 LGB individuals and 5,775 heterosexual adults. Detailed descriptive characteristics of the LGB and matched-control samples are presented in Table 1.

General Medical and Mental Health Conditions

Gay men, lesbians, and bisexuals were more likely than heterosexuals to have one or more chronic general medical conditions. However, gay men were less likely than heterosexuals to have developmental and intellectual disabilities and alcohol or substance use–related disabilities. Bisexuals and lesbians were more likely than heterosexuals to have a physical disability. Lesbians and bisexual women were more likely than heterosexuals to be categorized as having serious mental illness, to use tobacco products, and to have alcohol and substance use–related disabilities. However, lesbians were less likely than heterosexuals to have intellectual disabilities (Table 2).

Acute Health Care

Gay men were significantly less likely than heterosexual men to have used inpatient mental health services (adjusted OR [AOR]=0.69) or visited an ER for mental health (AOR=0.63) or nonbehavioral health (AOR=0.87) reasons in the past 12 months. In rate of service use, gay men had fewer non behavioral health ER visits than heterosexuals (adjusted RR [ARR]=0.72) (Table 3).

Lesbians were significantly less likely than heterosexual women to have used inpatient mental health services AOR=0.73) or visited an ER in the past 12 months for mental health (AOR=0.79) or substance use disorder (AOR=0.53) reasons. However, they were more likely than heterosexuals to have visited an ER for nonbehavioral health reasons (AOR=1.21) (Table 4). In rate of service use, the average number of ER visits for nonbehavioral health and substance use disorders was lower among lesbians than heterosexuals (ARR=0.89 and ARR=0.63, respectively) (Table 4).

For bisexual men, the likelihood of having used any of these acute services in the past 12 months was no different from heterosexual men after covariate adjustment. However, on average, they had significantly shorter substance use–related inpatient stays (ARR=0.65) and substance use–related ER visits (ARR=0.69) compared with heterosexuals (Table 3). Interestingly, bisexual women were significantly more likely than heterosexual women to have used ER services for a nonbehavioral health matter (AOR=1.48). However, on average, they had significantly shorter substance use–related inpatient stays (ARR=0.66) and substance use–related ER visits (ARR=0.55) compared with heterosexuals (Table 4). In summary, gay men and lesbians showed less acute service use than heterosexual adults. Compared with heterosexuals, bisexual men showed no significant difference in the odds of having used acute services. However, bisexual women had higher odds of having used nonbehavioral health ER services. Both bisexual men and women had a lower rate of acute substance use services compared with heterosexuals.

Discussion

This study builds upon the broader literature on the variations in health conditions and service use among sexual minority populations in general and extends the findings within a mental health system. This study provides a unique perspective on LGB health care because PCS data are collected from EHR systems and are not self-reported. In addition, the study was guided by a well-established model on health care access and utilization (15). Therefore, its findings provide a comprehensive view on health conditions and acute service use among LGB individuals within a mental health system.

Although several studies have examined health and behavioral conditions and access to and use of health care use among sexual minority groups (21), the findings have not always been consistent because of difficulties in obtaining a representative LGB sample and appropriate control groups. Nevertheless, much of the findings support our hypothesis that there are differences in health conditions and service use between LGB individuals and heterosexuals. Some studies reported that, in general, LGB individuals compared with heterosexuals appear to have poor health/behavioral conditions (4, 12, 21). Although many studies associate poor health and use of fewer services with not having health insurance (22–27), this cannot be the case in a Medicaid-insured mental health system.

A study by Tjepkema (27) on the relationship between health care use and sexual identity among Canadians ages 18 to 59 showed that relative to heterosexuals, gay men and bisexual women were more likely to report multiple chronic conditions and physical disabilities. Bisexuals were two or more times more likely than heterosexuals to perceive their mental health as poor. These findings appear to be similar to the findings of this study: gay men, lesbians, and bisexual individuals were more likely than heterosexuals to have at least one chronic physical condition. Also, compared with heterosexuals, bisexuals and lesbians were more likely to have a physical disability. These findings suggest that even among persons with insurance and a similar propensity to access and utilize services, significant differences remain in service use and health conditions among heterosexuals and LGB individuals, with greater morbidity among the LGB population.

Although research findings on health care use among LBG individuals have varied by study location, methods, and sample, the odds of accessing or using care were either equivocal or lower in the LGB population than in the comparable population (27, 28). This analysis showed that gay men and lesbians were the least likely to either visit an ER or have a mental health–related inpatient stay. Notably, both gay men and lesbians on average tended to visit general ERs less frequently than heterosexuals. Also evident was the less frequent use of acute substance use–related services among bisexuals compared with heterosexuals. On average, bisexuals had shorter substance use–related inpatient stays and fewer substance use–related ER visits. Overall, LGB individuals appeared to have more chronic physical conditions and physical disabilities and tended to use fewer inpatient and emergency services.

The tendency for Medicaid-insured LGB individuals to use fewer acute services may be related to several factors. The stigmatization of mental illness cannot be overemphasized (29). Also, the interpersonal and structural discrimination (12) toward homosexuals, coupled with the stigmatization of homosexuality and provider attitudes toward homosexuality, indirectly serves as a barrier to care for LGB individuals (16, 30–32). This may explain the underutilization of needed health care services (12). Further, prejudice toward homosexuals may have resulted in high rates of concealment of gender and sexual identity when using health care services. In fact, about 13% (N=23,270) of the records submitted to PCS listed gender and sexual identity as unknown.The study had some limitations. First, the PCS is conducted every 2 years and uses a cross-sectional survey methodology. Although all programs licensed or funded by NYS OMH are required to complete the survey, the PCS does not capture the entire service population. Not every patient receives services during the survey week. Also, programs may not report data for all patients served during the survey week. Second, mental health providers outside the public mental health system, such as private practitioners, were not surveyed. However, the PCS is a provider-reported, not a client-reported, survey. Reporting providers collect and submit the required information from each client’s medical records. Therefore, the information collected is free of self-reporting errors. Third, services not covered under Medicaid, such as those paid out of pocket, were not accounted for in this study. Last, the study used data from Medicaid-insured individuals in the New York State mental health system and may not be generalizable to other populations.

Conclusions

The study provides evidence of disparities in health conditions and use of acute services among sexual-minority populations within a public mental health system. Overall, compared with matched heterosexuals, LGB individuals appeared to have more chronic physical conditions and physical disabilities and tended to use fewer inpatient and emergency services. Gay men and lesbians were the least likely to have used an inpatient or emergency room service in the past 12 months. Future work will examine use of outpatient services by LGB individuals for both general medical and behavioral conditions.

Increasing access to screening, diagnostic, and preventive services among members of sexual minority groups should be a key component of federal, state, and local health reforms. These efforts will require evidence-based strategies to effectively engage LGB individuals in treatment for both general medical and mental illness.