Abstract
Purpose: This study explored the utility of the California Health Interview Survey (CHIS) to compare health-related outcomes among gay men, lesbians, and heterosexuals who reported being in a legally recognized partnership.
Methods: We regressed sexual identity and marriage/legally recognized partnership status on seven different outcomes related to health insurance coverage, medical services access and use, and general health and well-being using CHIS data collected between 2009 and 2013.
Results: There were 1432 respondents who identified as gay, lesbian, or homosexual, and 67,746 who identified as heterosexual. The percentage of participants who reported being married/legally partnered was 54.06% for heterosexual women, 52.93% for heterosexual men, 38.83% for lesbians, and 23.56% for gay men. Legally partnered/married gay and lesbian respondents were more likely to have health insurance and use healthcare than their counterparts not in such partnerships; few trends were statistically significant. Gay men in legally recognized partnerships were more likely than their heterosexual counterparts to report continuous health insurance coverage, a usual medical care source, and at least one provider visit within the past 12 months. We found statistically significant poorer health status outcomes among lesbians in legally recognized partnerships compared to married heterosexual women.
Conclusions: Lesbians in legally recognized partnerships did not fare as well as married heterosexual women. Gay men in legally recognized partnerships fared better than married heterosexual men on some measures. CHIS questionnaire structures limited our sample and analyses. We recommend that CHIS and other researchers ask partnered status-, marriage-, and sexual identity-related questions en bloc to ensure more robust representation, analyses, recommendations, and policy resolutions.
Keywords: : healthcare use, health insurance coverage, health status, legal partnerships and marriage
Introduction
As the issue of marriage equality progressed across the United States, citizens cited legal precedents and research findings that the lack of legal recognition for same-sex couples had negative health consequences1–4 in comparison to opposite-sex couples for whom the health benefits associated with marriage are well-documented.5–7 Policy and health literature reviews acknowledged the need for data to support the idea that legal recognition could facilitate health benefits for lesbian and gay male couples,2–4,8 yet rigorous studies had been limited due to the piecemeal, state-based legalization process in the United States and the lack of large representative surveys that include same-sex couples in survey language.9
Researchers previously explored the relationship between cohabitation and health status between same- and opposite-sex cohabiters through sexual identity questions in large U.S. population-based studies. For example, Liu et al.10 pooled 1997–2009 National Health Interview Survey (NHIS) data and found that same-sex cohabiting couples reported better health than unpartnered adults and opposite-sex cohabiting couples, although socioeconomic status explained that difference. Similarly, Denney et al.11 examined 1997–2008 NHIS data yet found that same- and opposite-sex cohabiters reported poorer health than heterosexual married couples. At the time, legal recognition for same-sex couples was so rare that the studies focused only on cohabitation. In 2013, NHIS12 began regular implementation of sexual identity questions on its entire sample, and has published on lesbian, gay, and bisexual health indicators, health-related behaviors, access, and use, but not yet marriage.
The U.S. Supreme Court's 2015 decision that the Constitution's “Fourteenth Amendment require[s] a state to recognize a marriage between two people of the same sex when their marriage was lawfully licensed and performed out-of-state”13(p1) ceased piecemeal legalization of same-sex marriage in the United States. We believe it is now incumbent upon LGBT health researchers to assess the influence of marriage on the health and well-being of same-sex couples and their immediate families. To achieve these aims, standardized wording for same-sex marriage needs to be implemented and tested in population-level studies.
The purpose of this study is to compare health-related outcomes among gay men, lesbians, and heterosexuals who reported being part of a legally recognized partnership versus other gay men, lesbians, and heterosexuals.
Hypothesis 1: Adults who are in legally recognized partnerships, compared to those who are not, are more likely to report general good health, health insurance coverage, and use of medical services regardless of sexual identity.
Hypothesis 2: Legally recognized, partnered lesbian and gay (LG) adults are as likely as married heterosexuals to report general good health, health insurance, and use of medical services.
The exploratory aims of this article are to describe the survey instrument and skip pattern used to assess same-sex marriage/legal partnership and document any benefits or concerns for adaptation of the measure to other populations. California Health Interview Survey (CHIS) is a unique dataset, as its questions on legal partnership and marriage for same-sex couples stayed abreast of legal California events.
Methods
CHIS is an ongoing, independent, cross-sectional telephone survey that monitors the health status of California residents. Multiple U.S. Department of Health and Human Services agencies have provided financial support to CHIS. First fielded biennially, CHIS became a continuous survey in 2011.14–17 We combined and analyzed CHIS data collected between 2009 and 2013; CHIS labels these two waves of data as “CHIS 2009” and “CHIS 2011–2012,” respectively.16,17 The University of California, Los Angeles (UCLA) Center for Health Policy Research made the data available upon our request. The survey's multistage probability design collects a comprehensive sample representative of sex/gender and race/ethnicity for all California counties. The California Information Practices Act (section 1798.24) permits the release of CHIS data only for statistical research and reporting, and precludes the release of geography-related information to ensure participant confidentiality.14–17 UCLA obtained original CHIS human subjects approval. The National Institutes of Health Office of Human Subjects Research Protection determined that our project was exempt from review.
Dependent variables
We explored the relationship of marriage and legally recognized partnership with seven binary, dependent variables. We dichotomized (1) self-reported general health status as excellent/very good/good and poor/fair15; (2) health insurance as insured over all past 12 months versus lack of continuous coverage during the same period; (3) usual source of care as having a regular place to visit when sick or need health advice versus no usual source; (4) seen a medical doctor as ever in the past 12 months; (5) mammogram as having an examination within the past 2 years for women aged 50 and older; (6) delay in medical care in past 12 months; and (7) delay in prescription in past 12 months. We recoded responses to variables 2–7 as yes or no.
Independent variables
Sexual identity
CHIS asked participants, “Do you think of yourself as straight or heterosexual, as gay, lesbian, or homosexual, as bisexual, or other?” We included only those participants who identified as gay, heterosexual, homosexual or lesbian.
Marital or partnered status
All respondents were asked the standard martial question “Are you now married, living with a partner in a marriage-like relationship, widowed, divorced, separated, or never married?”
Legally recognized couple status
In 2009, CHIS added the following question for adults who reported having sex with someone of the same sex over the preceding 12 months: “Are you legally registered as a domestic partner or legally married in California with someone of the same sex?”15 The CHIS interview schedule placed its question on (heterosexual) marital status within sociodemographics and its question on same-sex marriage and/or legal partnership within the context of recent same-sex sexual activity.
We recoded legally recognized couple status based on responses to both the standard marital status question and the legally registered and married question. If participants responded as legally partnered or married, that response superseded their response to the standard marital status question. The ensuing relatively small number of legally partnered or married responses led us to combine responses into two categories: (1) legally recognized and (2) other, a category that includes unmarried people who may be divorced, widowed, never married, or living with another person without legal recognition. We created this binary variable to correspond to our other binary variables and focus on cohabiting couples with legally recognized partnerships, because at the time that CHIS collected these data, statewide, same-sex legal partnerships and marriage were permitted inconsistently. To simplify the reading process, we use legally recognized, partner(s), and partnerships to refer to people who are legally married, legally partnered, or in another civilly recognized union.
Covariates
Sociodemographics included age, education, race/ethnicity, employment status of respondent and partner, household income, presence of children at home, and gender to explore variations within and across lesbian, gay, and heterosexual couples. We recoded educational achievement into three categories: less than high school or equivalency, high school diploma, including some college coursework, or a 4-year college degree or higher. Participant race/ethnicity was categorized as Latina/Latino, African American, White, Asian, Native (American Indian/Native American, Alaska Native, and Pacific Islander), multiple race, and other. We recoded employment as employed, unemployed looking for work, or unemployed not looking for work and applied the same trichotomy to the employment status of the respondent's legal partner. Participants reported household income ranges and the number and age of children in the home; we created binary variables to reflect whether household incomes were above or below 350% of the U.S. Federal poverty level (a standard marker for public assistance) and whether a household included at least one child under age 18.
Statistical analysis
Descriptive statistics for sociodemographic, gender, and sexual identity variables appear in Table 1. We used chi-square tests to assess the significance of difference in proportions and ANOVA to test for difference in means by sexual identity for women and men separately. We constructed separate logistic regression models for each of the seven dependent variables, with separate models for women and men. Each model included sexual identity, partnership status, and the interaction term of sexual identity and partnership status, adjusted for the above-described sociodemographic covariates. We used the interaction term of sexual identity by legal partnership status to compare if regression coefficients for legally recognized status differed by sexual identity (e.g., does legal recognition relate to health insurance differently for lesbians in legally recognized partnerships compared to their heterosexual counterparts?). We could have conducted separate regressions for each sexual identity, but we would have been unable to compare the relationship between legal partnership and the dependent variable for persons of different sexual identities. The interaction term allows us to formally test if legal recognition relates to a health outcome differently for various statistical conditions18—in this case, sexual identity.
Table 1.
Sociodemographic Characteristics of Gay, Heterosexual, and Lesbian Participants, Aged 18–70: California Health Interview Survey 2009–2012
| Women | Men | |||||
|---|---|---|---|---|---|---|
| Lesbian | Heterosexual | P value | Gay | Heterosexual | P value | |
| Sample size (%) | 583 (1.28%) | 39,323 (96.06%) | 849 (2.45%) | 28,423 (95.91%) | ||
| Legally recognized partner status—four levels | ||||||
| Legally recognized partnerships | 224 (38.83%) | 21,092 (54.06%) | <0.001 | 203 (23.56%) | 16,598 (52.93%) | <0.001 |
| Living with partner without legal recognition | 89 (19.72%) | 2105 (7.81%) | 126 (20.88%) | 1557 (7.59%) | ||
| Widowed/divorced/separated | 79 (10.86%) | 9976 (15.43%) | 68 (3.77%) | 4035 (9.23%) | ||
| Never legally partnered | 191 (30.59%) | 6150 (22.70%) | 452 (51.79%) | 6233 (30.25%) | ||
| Legally recognized partner status—dichotomized | ||||||
| Legally recognized partnerships | 224 (38.83%) | 21,092 (54.06%) | <0.001 | 203 (23.56%) | 16,598 (52.93%) | <0.001 |
| Other | 359 (61.17%) | 18,231 (45.94%) | 646 (76.44%) | 11,825 (47.07%) | ||
| Age | ||||||
| Mean (standard error) | 41.81 (0.90) | 42.08 (0.05) | 0.773 | 40.57 (0.71) | 41.25 (0.05) | 0.358 |
| Race/ethnicity | ||||||
| Latino | 104 (29.32%) | 9142 (34.57%) | <0.001 | 141 (26.95%) | 6416 (35.93%) | 0.060 |
| Asian/PI/AI/AN | 31 (7.20%) | 4656 (15.61%) | 54 (14.19%) | 3513 (13.68%) | ||
| African American | 29 (4.90%) | 1793 (5.84%) | 31 (4.71%) | 1156 (5.35%) | ||
| White | 398 (53.61%) | 22,795 (42.13%) | 602 (51.46%) | 16,651 (43.11%) | ||
| Other | 21 (4.97%) | 937 (1.86%) | 21 (2.69%) | 687 (1.93%) | ||
| Have child aged 0–17 in household | ||||||
| Yes | 117 (28.94%) | 15,095 (49.34%) | <0.001 | 66 (15.51%) | 10,257 (44.45%) | <0.001 |
| No | 466 (71.06%) | 24,228 (50.66%) | 783 (84.49%) | 18,166 (55.55%) | ||
| Educational attainment | ||||||
| < High school education | 29 (8.50%) | 4413 (15.66%) | <0.001 | 26 (4.28%) | 2984 (16.14%) | <0.001 |
| High school diploma or some college | 231 (46.48%) | 19,191 (48.82%) | 328 (41.17%) | 13,665 (49.93%) | ||
| College degree or above | 323 (45.03%) | 15,719 (35.52%) | 495 (54.55%) | 11,774 (33.93%) | ||
| Working status | ||||||
| Employed | 368 (63.73%) | 22,603 (61.09%) | 0.009 | 526 (71.42%) | 19,566 (74.59%) | 0.448 |
| Unemployed, looking for work | 53 (13.75%) | 2556 (8.92%) | 64 (10.96%) | 2258 (10.21%) | ||
| Unemployed, not looking for work | 162 (22.51%) | 14,164 (29.99%) | 259 (17.62%) | 6599 (15.20%) | ||
| Household income | ||||||
| Household income under 350% FPL | 231 (40.62%) | 19,960 (55.55%) | <0.001 | 299 (33.27%) | 13,077 (52.45%) | <0.001 |
| Household income greater than or equal to 350% FPL | 352 (59.38%) | 19,363 (44.45%) | 550 (66.73%) | 15,346 (47.55%) | ||
| Partner's working status (among participants with a legally recognized partner) | ||||||
| Employed | 89 (85.23%) | 15,324 (78.90%) | <0.001 | 58 (80.50%) | 9722 (59.14%) | 0.010 |
| Unemployed, looking for work | 0 (0.00%) | 829 (4.74%) | 5 (3.23%) | 582 (4.05%) | ||
| Unemployed, not looking for work | 19 (14.77%) | 4933 (16.36%) | 20 (16.27%) | 6290 (36.81%) | ||
Except for age, numbers in cells are weighted frequencies and percentages for variables. Response rates are estimates of the percentage of sampled persons who participated in the survey, where the sample may be across the entire state, restricted to a county, or some other subgroup. To estimate response rates, CHIS takes the probability of sampling persons into account. Thus, response rates are weighted percentages of the entire number of CHIS participants rather than percentages of a subsample—in this case, gay, heterosexual, and lesbian California adults younger than 70 years. For age, we report means and standard deviations. Chi-square and t-tests for each variable by sexual identity were conducted separately for men and women. P values less than 0.05 are discussed as significant within the text.
CHIS, California Health Interview Survey; FPL, Federal Poverty Level (U.S.).
Model-based estimates and unadjusted prevalences appear in Table 2. We analyzed each variable in discrete models to test our two hypotheses: the “Association of Legal Recognition” column presents the association of this status within each sexual identity group. “Legal Recognition” is a binary variable, comparing legally partnered participants to “other,” which includes unmarried and nonlegally partnered individuals. For example, we compared legally partnered lesbians to other (i.e., not legally partnered) lesbians. The “Association of Sexual Identity” column presents the comparison between legally partnered gay men and lesbians, respectively, to the reference group, heterosexual respondents in legally recognized partnerships. For example, we compared legally partnered lesbians to legally partnered heterosexual women.
Table 2.
Health and Access Variables Regressed on Sexual Identity and Legally Recognized Status: California Health Interview Survey 2009–2012
| Women | Men | |||||
|---|---|---|---|---|---|---|
| Outcome | Sexual identity | Partnership status | Association of legal recognition, AOR [95% CI] | Association of sexual identity, AOR [95% CI] | Association of legal recognition, AOR [95% CI] | Association of sexual identity, AOR [95% CI] |
| Good/very good/excellent health, self-reported | Lesbian/gay | Legally partnered | 0.77 [0.35, 1.68] | 0.45 [0.25, 0.81]** | 0.91 [0.42, 1.97] | 0.70 [0.36, 1.40] |
| Other | 1.00 (REF) | — | 1.00 (REF) | — | ||
| Heterosexual | Legally partnered | 1.26 [1.12, 1.43]*** | 1.00 (REF) | 1.25 [1.06, 1.47]** | 1.00 (REF) | |
| Other | 1.00 (REF) | — | 1.00 (REF) | — | ||
| Have usual place to go when sick or needing health advice | Lesbian/gay | Legally partnered | 1.11 [0.37, 3.33] | 0.68 [0.26, 1.78] | 1.73 [0.70, 4.25] | 2.78 [1.27, 6.05]* |
| Other | 1.00 (REF) | — | 1.00 (REF) | — | ||
| Heterosexual | Legally partnered | 1.20 [1.03, 1.41]* | 1.00 (REF) | 1.38 [1.15, 1.65]*** | 1.00 (REF) | |
| Other | 1.00 (REF) | — | 1.00 (REF) | — | ||
| Visited doctor during past 12 months | Lesbian/gay | Legally partnered | 0.88 [0.38, 2.03] | 0.68 [0.33, 1.39] | 3.21 [1.35, 7.64]** | 5.27 [2.34, 11.85]**** |
| Other | 1.00 (REF) | — | 1.00 (REF) | — | ||
| Heterosexual | Legally partnered | 1.07 [0.95, 1.21] | 1.00 (REF) | 1.17 [1.01, 1.35]* | 1.00 (REF) | |
| Other | 1.00 (REF) | — | 1.00 (REF) | — | ||
| Insured all past 12 months | Lesbian/gay | Legally partnered | 1.58 [0.80, 3.11]**** | 0.72 [0.42, 1.24]**** | 2.67 [1.26, 5.68]**** | 1.62 [0.82, 3.18]**** |
| Other | 1.00 (REF) | — | 1.00 (REF) | — | ||
| Heterosexual | Legally partnered | 1.32 [1.17, 1.49] | 1.00 (REF) | 2.30 [1.96, 2.69] | 1.00 (REF) | |
| Other | 1.00 (REF) | — | 1.00 (REF) | — | ||
| Delayed medical care in the past year | Lesbian/gay | Legally partnered | 0.59 [0.31, 1.12] | 1.31 [0.79, 2.16] | 0.55 [0.29, 1.04]# | 1.15 [0.74, 1.80] |
| Other | 1.00 (REF) | — | 1.00 (REF) | — | ||
| Heterosexual | Legally partnered | 0.86 [0.77, 0.98]* | 1.00 (REF) | 0.79 [0.69, 0.90]*** | 1.00 (REF) | |
| Other | 1.00 (REF) | — | 1.00 (REF) | — | ||
| Delayed getting prescription in past year | Lesbian/gay | Legally partnered | 1.47 [0.72, 3.00] | 2.19 [1.26, 3.83]** | 1.10 [0.56, 2.16] | 1.42 [0.83, 2.45] |
| Other | 1.00 (REF) | — | 1.00 (REF) | — | ||
| Heterosexual | Legally partnered | 0.79 [0.69, 0.90]*** | 1.00 (REF) | 0.81 [0.66, 0.99]* | 1.00 (REF) | |
| Other | 1.00 (REF) | — | 1.00 (REF) | — | ||
| Had mammogram in the past 2 years (universe: women 50+) | Lesbian/gay | Legally partnered | 2.34 [0.94, 5.85]# | 0.87 [0.44, 1.75] | ||
| Other | 1.00 (REF) | — | ||||
| Heterosexual | Legally partnered | 1.63 [1.41, 1.89]**** | 1.00 (REF) | |||
| Other | 1.00 (REF) | — | ||||
Each row displays results for a separate multivariate regression for one of the seven dependent variables. For all AORs, the logistic regression models included age, race, education, children in household, employment status, household income, insured past 12 months, sexual identity, legally partnered status (or not), as well as the interaction of sexual identity by partnered status. When insured in the past 12 months was the dependent variable, we did not include insurance as an independent variable.
Numbers in cells represent (1) AORs for people in legally recognized partnerships and (2) AORs for sexual identity among respondents in legally recognized partnerships. For the column, “Association of Legal Recognition, AOR,” legally partnered respondents are compared to other respondents (reference group) within each sexual identity and gender combination (e.g., lesbians in legally recognized partnerships compared to other lesbians). For the column, “Association of Sexual Identity, AOR,” lesbians and gay men in legally recognized partnerships are compared to heterosexuals in legally recognized partnerships within each gender (e.g., legally partnered lesbians compared to heterosexual women in legally recognized partnerships).
P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001; #P < 0.10.
AOR, adjusted odds ratios; CI, confidence interval.
Adjusted odds ratios (AOR) are reported. We rejected the null hypothesis if a P value is less than or equal to 0.05. Our two-tailed analyses were programmed in SAS V9.2 and SAS-callable SUDAAN software (SAS Institute Inc., Cary, NC).
Results
Sample characteristics
There were 90,549 Californians aged 18 and older in our sample. This study focused on Californians who identified as gay, heterosexual, or lesbian—to be consistent with both the CHIS sexual identity question as well as the 2011 Institute of Medicine's sexual identity definition.9(p28)
To create this subsample, we excluded participants who identified as nonsexual/celibate (n = 456) and other (n = 53). We also excluded participants who self-identified as bisexual (n = 985) and participants who provided no response to the sexual identity question (n = 51) because the interview schedule did not ask systematically for their legal partner status. As CHIS asked no sexual identity questions of participants older than 70 (n = 19,826), we excluded them, leaving a sample size of 69,178 for our analyses. The retained respondents identified as gay, lesbian, homosexual (n = 1432), or heterosexual (n = 67,746).
Table 1 shows 2.45% of the adult (aged 70 or younger) male and 1.28% of the adult female California sample identified as gay or lesbian, respectively. The percentage of participants who reported being married using the standard marital wording was 54.06% for heterosexual women, 52.93% for heterosexual men, 22.2% for lesbians, and 9.7% for gay men. However, when participants were asked the legally partnered or legally married same-sex question, the prevalence increased to 38.83% for lesbians and 23.56% for gay men. The percentage of lesbians reporting having a child younger than 18 years was almost twice that of gay men, although heterosexuals overall were more likely to report children at home.
Lesbians differed somewhat demographically from heterosexual women. Although lesbians were approximately the same average age as heterosexual women, they were more likely to be white, to report higher educational attainment, to be in the labor force (employed or unemployed and looking for work), to have a partner in the labor force, and to have higher overall household income, all P values <0.01. More gay than heterosexual men reported being white (51%–43%). Gay men were significantly more likely than heterosexual men to report holding at least one college degree (55%–34%), to have a working partner (81%–59%), and to have higher household incomes (67%–48%), all P values ≤0.01.
Comparisons of self-reported health and healthcare use between respondents in legally recognized partnerships and other respondents
We explored the relationship between legally recognized partnership status and health-related variables. We compared legally partnered versus other respondents within gender and sexual identity strata (legally partnered vs. other lesbians; legally partnered vs. other gay men). These results appear in Table 2, in the “Association of Legal Recognition” column. AORs and their 95% confidence intervals (CIs) for statistically significant relationships among gay and lesbian respondents also appear in Table 2.
Participants in legally recognized partnerships reported higher odds of consistent health insurance over the past 12 months; odds were significantly higher among legally partnered lesbians (AOR = 1.58; 95% CI: 0.80, 3.11) and gay men (AOR = 2.67; 95% CI: 1.26, 5.68) than among other lesbians and gay men. Gay and heterosexual men in legally recognized partnerships were more likely to report a doctor's visit in the past 12 months (AOR = 3.21; 95% CI: 1.35, 7.64 among gay men); however, this legal status was not related to visiting a doctor for women of either sexual identity. Heterosexual women in legally recognized partnerships were more likely to report a mammogram within the past 2 years; this association was not significant for legally partnered lesbians. We found a higher OR for legally partnered heterosexuals to report good or excellent health and lower odds for delay in prescriptions compared to unmarried heterosexuals. Odds of delay in medical care were lower among legally partnered heterosexuals.
Comparisons of self-reported health and healthcare use between legally recognized same-sex and heterosexual couples
We explored relationships among legally recognized same-sex and heterosexual couples and health insurance coverage, and healthcare use. These results appear in the Table 2 column “Association of Sexual Identity.” AORs and their 95% CIs are included for statistically significant relationships.
Lesbians in legally recognized partnerships were substantively less likely than their heterosexual counterparts to report good or excellent health (AOR = 0.45; 95% CI: 0.25, 0.81), and to be insured consistently over the past 12 months (AOR = 0.72; 95% CI: 0.42, 1.24). Moreover, these lesbians were much more likely to report delays in filling prescriptions (AOR = 2.19; 95% CI: 1.26, 3.83). In addition, lesbians in legally recognized partnerships had lower odds of reporting a usual source of care, any healthcare visits, and mammograms.
We found significantly higher odds for gay men in legally recognized partnerships (than heterosexual men in legally recognized partnerships) to have continuous health insurance coverage (AOR = 1.62; 95% CI: 0.82, 3.18), a usual source of medical care (AOR = 2.78; 95% CI: 1.27, 6.05), and at least one doctor's visit over the past 12 months (AOR = 5.27; 95% CI: 2.34, 11.85). Yet, only legally partnered heterosexual men were significantly more likely to report good through excellent health.
Discussion
Before same-sex marriage became legal nationwide in June 2015,13 clinical researchers recommended marriage equality as one way to promote LGBT health.19,20 Concurrently, scientists demonstrated the lack of comprehensive, representative datasets to explore this opinion rigorously.3,4,9,10,21 Several population-level surveys ask about sexual identity, but do not include same-sex marriage or legal partnership as options to answer marital status questions. These datasets may miss marriage as an important moderator or confounder in their analyses.
Although California began to provide legal partnership protections for same-sex couples in 1999,22 in 2008, its Supreme Court guaranteed marriage equality and associated benefits for same-sex couples, including couples married earlier and out-of-state. As same-sex legal partnerships and marriage emerged in California and CHIS quickly implemented corresponding questions before most other United States, we thought an immediate examination of available CHIS data could refine current knowledge and facilitate future representative U.S. studies to explore health and well-being benefits associated with legal same-sex marriage and partnerships.
Our findings replicate previous findings that heterosexuals in legally recognized partnerships report better health, health insurance coverage, and use of medical services compared to their counterparts who are not in such partnerships. Among CHIS participants in legally recognized same-sex partnerships, however, we found significant positive results only in regard to health insurance coverage for both LG same-sex couples. Our analysis did not find uniformly positive results. For example, although gay men in legally recognized partnerships were more likely than other gay men to report a doctor's visit in the past 12 months, we did not find similar results among lesbians in this sample.
Subsequently, our findings did not support our second hypothesis that lesbians and gay men in legally recognized partnerships would report equally good health outcomes as married heterosexual couples. We found gay men in legally recognized partnerships to be more likely than legally partnered heterosexual men to have continuous health insurance coverage, a usual medical care source, and at least one provider visit over the past 12 months. However, we found statistically significant poorer health status outcomes among lesbians in legally recognized partnerships compared to their heterosexual female counterparts.
Minority stress theory23–25 posits that homoprejudicial experiences over the life course influence actual and perceived mental and physical health among lesbians and gay men who, nonetheless, also display stores of strength and resilience.9 An analysis of 2009 CHIS data found that respondents in legally recognized same-sex partnerships reported lower levels of psychological distress than their single LG counterparts, although married heterosexuals in this sample reported the lowest levels of psychological distress.26 Moreover, recent studies have found that people in legally partnered same-sex relationships have higher relationship stability, more financial resources, and better health outcomes than couples who cohabit without legal recognition.27,28 In addition, LeBlanc et al. have hypothesized that lesbians and gay men may not experience the health benefits of marriage exactly as their heterosexual counterparts, as the individuals in these same-sex partnerships may experience individual- and/or couple-level stressors related to internalized homophobia, homoprejudice, and/or the lack of universal support for same-sex rather than opposite-sex marriage.29 It is not possible to explore these ideas with this small sample; nevertheless, active studies as described above and future work may continue to find disparate health and other outcomes for legally recognized same- and opposite-sex partners—perhaps even future analyses with a larger sample that includes recently released CHIS data. Indeed, this framework and corresponding analyses may be extended to analyze the health outcomes for other marginalized couples, for example, interfaith, interracial, or age-discordant couples.29,30
Limitations
Our decision to combine legal partnerships and legal marriage as one convenient status, legally recognized, is a limitation as these discrete legal statuses did not convey identical rights and privileges between 2009 and 2012, a period that included concurrent, limited periods of city-, county-, or state-based same-sex marriages in California. We conflated legal partnerships and marriage as a small proportion of sampled Californians identified as gay or lesbian, and an even smaller proportion reported being in a legally recognized partnership. To place our findings in context, we included heterosexuals as a comparison group given the data-driven tenet that marriage provides property and health benefits for heterosexuals and the recent societal argument, now law, that marriage can provide similar benefits for same-sex couples. We could have conducted separate regressions for each sexual identity, but would not have been able to compare the relationship between marriage and the dependent variable for persons of different sexual identities. The interaction term allows us to formally test if marriage relates to a health outcome differently for lesbian and gay male CHIS participants—an interaction we could not test given the structure of the questionnaire and corresponding dataset.
Another limitation is the need to exclude bisexually identified Californians because the survey's order and skip pattern complicate notions of self-identification and self-report of sexual behavior. Thus, a respondent who identified as bisexual may not have had sex with a same-sex partner in the last 12 months; consequently, the CHIS questionnaire would not have prompted such a participant to answer the question of legal partnership/marriage with someone of the same sex.
Conclusions
Our exploration of CHIS data demonstrates that the survey schedule provides a set of questions that researchers can adapt rapidly to conduct population-based research on the health and well-being of married same-sex couples and their families. The rate of legally partnered or married same-sex individuals almost doubled by including legally recognized partnerships instead of only the standard marital status question.
Our findings and limitations lead us to strongly recommend that researchers adapt CHIS questions on same-sex marriage and skip patterns to ensure complete and accurate responses. For example, the interview schedule for the CHIS data we analyzed placed same-sex partnership/marriage questions only after questions about same-sex sexual encounters within the preceding 12 months. This questionnaire structure precludes asking legal partnership status of all bisexual participants as it is possible that some participants had no same-sex relationships within that limited time period. Also likely is that same-sex couples may not have had sex within the preceding 12 months due to distance, illness, and other myriad reasons. Any LG Californians who reported no such sex during this time period would have lacked the opportunity to answer these legal partnership/marriage questions as well—thus limiting the representativeness of this study.
We also recommend that CHIS and other researchers ask marriage- and sexual identity-related questions en bloc to ensure the representativeness of a sample. The interview schedule that collected the data we analyzed asked about same-sex marriage and partnerships only after respondents confirmed having sex with a same-sex partner within the preceding 12 months; it asked about presumed heterosexual marriage in the sociodemographics section. En bloc placement of these questions ensures more robust representation, analyses, recommendations, and policy resolutions.
Although sexual identities other than heterosexual have been commonplace in the United States for decades, the National Center for Health Statistics finalized the development of a valid and reliable question on this topic in this decade31 and began to ask Americans this question in 2013.12 The 2015 U.S. Supreme Court decision had direct implications for American lesbians and gay men. This change also affects healthcare providers and policymakers who expect researchers to provide data-driven recommendations. We can begin to provide these recommendations expeditiously by beginning with tested measures. Nevertheless, to explore the influence of legally recognized partnership on the health and well-being of all people, we recommend proposing the question consistently regardless of participants' sexual identity.
Acknowledgments
The UCLA Center for Health Policy Research conducts the CHIS. The National Institutes of Health (NIH) Office of Behavioral and Social Sciences Research (OBSSR) provided financial support for design and implementation of previous iterations of CHIS. NIH/OBSSR supported access to the dataset created and analyzed for this project. The authors thank William Byne, Julia Chapman, Dana Wolff-Hughes, and this journal's reviewers for their input on previous versions of this article.
Disclaimer
This article represents only the authors' views and perspectives, not the positions of the National Institutes of Health or the U.S. Government.
Author Disclosure Statement
No competing financial interests exist.
References
- 1.Tatum AK: The influence of same-sex marriage access on sexual minority identity, mental health, and subjective well-being. J Homosex 2016;1–16 [Epub ahead of print]; DOI: 10.1080/00918369.2016.1196991 [DOI] [PubMed] [Google Scholar]
- 2.Kealy-Bateman W, Pryor L: Marriage equality is a mental health issue. Australas Psychiatry 2015;23:540–543 [DOI] [PubMed] [Google Scholar]
- 3.Ponce NA, Cochran SD, Pizer JC, Mays VM: The effects of unequal access to health insurance for same-sex couples in California. Health Aff (Millwood) 2010;29:1539–1548 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.King B, Bartlett A: What same sex civil partnerships may mean for health. J Epidemiol Community Health 2006;60:188–191 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Waite LJ, Gallagher M: The Case for Marriage: Why Married People Are Happier, Healthier, and Better Off Financially. New York: Doubleday, 2001 [Google Scholar]
- 6.Goldman N, Korenman S, Weinstein R: Marital status and health of the elderly. Soc Sci Med 1995;40:1717–1730 [DOI] [PubMed] [Google Scholar]
- 7.Kiecolt-Glaser JK, Newton TL: Marriage and health: His and hers. Psychol Bull 2001;127:472–503 [DOI] [PubMed] [Google Scholar]
- 8.Kornblith E, Green RJ, Casey S, Tiet Q: Marital status, social support, and depressive symptoms among lesbian and heterosexual women. J Lesb Stud 2016;20:157–173 [DOI] [PubMed] [Google Scholar]
- 9.Institute of Medicine (US) Committee on Lesbian, Gay, Bisexual, and Transgender Health Issues and Research Gaps and Opportunities: The Health of Lesbian, Gay, Bisexual, and Transgender People: Building a Foundation for Better Understanding. Washington, DC: The National Academies Press, 2011 [PubMed] [Google Scholar]
- 10.Liu H, Reczek C, Brown D: Same-sex cohabitors and health: The role of race-ethnicity, gender, and socioeconomic status. J Health Soc Behav 2013;54:25–45 [DOI] [PubMed] [Google Scholar]
- 11.Denney JT, Gorman BK, Barrera CB: Families, resources, and adult health: Where do sexual minorities fit? J Health Soc Behav 2013;54:46–63 [DOI] [PubMed] [Google Scholar]
- 12.Ward BW, Dahlhamer JM, Galinsky AM, Joestl SS: Sexual orientation and health among U.S. adults: National Health Interview Survey, 2013. Natl Health Stat Report 2014;77:1–12 [PubMed] [Google Scholar]
- 13.Supreme Court of the United States: James Obergefell et al., Petitioners.v. Richard Hodges, Director, Ohio Department of Health, et al. Docket No. 14-556. June 26, 2015. Available at www.supremecourt.gov/Search.aspx?FileName=/docketfiles/14-556.htm Accessed July10, 2015
- 14.Brick JM, Edwards WS, Lee SH: Sampling telephone numbers and adults, interview length, and weighting in the California Health Interview Survey cell phone pilot study. Public Opin Q 2007;71:793–813 [Google Scholar]
- 15.California Health Interview Survey: CHIS 2009 Methodology Report Series: Report 4–Response Rates. Los Angeles, CA: UCLA Center for Health Policy Research, 2011. Available at http://healthpolicy.ucla.edu/Documents/Newsroom%20PDF/CHIS2009_method4.pdf Accessed December1, 2015 [Google Scholar]
- 16.California Health Interview Survey: CHIS 2009 Adult Survey. Los Angeles, CA: UCLA Center for Health Policy Research, November 2012. Available at http://healthpolicy.ucla.edu/chis/data/public-use-data-file/Documents/puf09_adult_datadic.pdf Accessed December1, 2015 [Google Scholar]
- 17.California Health Interview Survey: CHIS 2011–2012 Adult Survey. Los Angeles, CA: UCLA Center for Health Policy Research, June 2014. Available at http://healthpolicy.ucla.edu/chis/data/public-use-data-file/Documents/CHIS%202011-2012%20Data%20Dictionary%20PUF%20-%20Adult.pdf Accessed December1, 2015 [Google Scholar]
- 18.Dallal GE: The little handbook of statistical practice (Version 1.10). Amazon Kindle, 2012. Available at www.amazon.com/Little-Handbook-Statistical-Practice-ebook/dp/B00847SM6A Accessed April18, 2016 [Google Scholar]
- 19.Campion EW, Morrissey S, Drazen JM: In support of same-sex marriage. N Engl J Med 2015;372:1852–1853 [DOI] [PubMed] [Google Scholar]
- 20.Gonzales G: Same-sex marriage—A prescription for better health. N Engl J Med 2014;370:1373–1376 [DOI] [PubMed] [Google Scholar]
- 21.Ducharme JK, Kollar MM: Does the ‘marriage benefit’ extend to same-sex union? Evidence from a sample of married lesbian couples in Massachusetts. J Homosex 2012;59:580–591 [DOI] [PubMed] [Google Scholar]
- 22.California Legislative Information: California Family Code, Division 2.5. Domestic Partnership Registration [297–299.6]. Available at http://leginfo.legislature.ca.gov/faces/codes_displayText.xhtml?lawCode=FAM&division=2.5&title=&part=1.&chapter=&article=Accessed December1, 2015
- 23.Meyer IH: Prejudice as stress: Conceptual and measurement problems. Am J Public Health 2003;93:262–265 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Meyer IH: Prejudice, social stress, and mental health in lesbian, gay, and bisexual populations: Conceptual issues and research evidence. Psychol Bull 2003;129:674–697 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Wight RG, LeBlanc AJ, Meyer IH, Harig FA: Internalized gay ageism, mattering, and depressive symptoms among midlife and older gay-identified men. Soc Sci Med 2015;147:200–208 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Wight RG, LeBlanc AJ, Badgett MVL: Same-sex legal marriage and psychological wellbeing: Findings from the California Health Interview Survey. Am J Public Health 2013;103:339–346 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Baumle AK: Same-sex cohabiting elders versus different-sex cohabiting and married elders: Effects of relationship status and sex of partner on economic and health outcomes. Soc Sci Res 2014;43:60–73 [DOI] [PubMed] [Google Scholar]
- 28.Manning WD, Brown SL, Styles JB: Same-sex and different-sex cohabiting couple relationship stability. Demography 2016;53:937–953 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.LeBlanc AJ, Frost DM, Wight RG: Minority stress and stress proliferation among same-sex and other marginalized couples. J Marriage Fam 2015;77:40–59 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Lehmiller JJ, Agnew CR: Marginalized relationships: The impact of social disapproval on romantic relationship commitment. Pers Soc Psychol Bull 2006;32:40–51 [DOI] [PubMed] [Google Scholar]
- 31.Miller K, Ryan JM: Design, Development and Testing of the NHIS Sexual Identity Question. Hyattsville, MD: National Center for Health Statistics, October 2011. Available at wwwn.cdc.gov/qbank/report/Miller_NCHS_2011_NHIS%20Sexual%20Identity.pdf Accessed August30, 2016 [Google Scholar]