Associations among Childhood Household Dysfunction, Sexual Orientation, and DSM-5 Alcohol, Tobacco and Other Substance Use Disorders in Adulthood: Evidence from a National U.S. Survey

Abstract

Objectives:

This study examined the associations between childhood household dysfunction and adulthood past-year DSM-5 alcohol, tobacco, and other substance use disorders across sexual orientation subgroups (e.g., lesbian/gay, bisexual, and heterosexual).

Methods:

Prevalence estimates were based on National Epidemiologic Survey on Alcohol and Related Conditions (NESARC-III) data collected from structured diagnostic face-to-face interviews in a nationally representative sample of 36,309 U.S. adults. Multivariable regression was used to examine associations between childhood household dysfunction and past-year substance use disorders in adulthood.

Results:

Sexual minorities, particularly sexual minority women, reported higher rates of childhood household dysfunction (e.g., parental/household history of substance-related problems) and adulthood DSM-5 alcohol, tobacco, and substance use disorders. Results of multivariable analyses indicated that childhood histories of parental/household substance-related problems were associated with greater odds of past-year substance use disorders among sexual minorities than heterosexuals, and that such histories may moderate differences among sexual orientation subgroups. The risk of substance use disorders among sexual minority women relative to exclusively heterosexual women (i.e., heterosexual-identified women without same-sex attraction or behavior) remained high, even when accounting for household dysfunction. In contrast, there were no such differences between sexual minority men and exclusively heterosexual men.

Conclusions:

Sexual minorities are more likely to have childhood household dysfunction which in turn is associated with a higher risk of developing DSM-5 alcohol, tobacco, and substance use disorders in adulthood, especially among sexual minority women. Healthcare providers who care for individuals raised in dysfunctional households should carefully assess risk for substance use disorders and intervene as needed.

1. INTRODUCTION

An assumption in developmental psychology, psychiatry, and related fields is that adverse childhood experiences often set the stage for poorer adult functioning (Caspi and Silva, 1995; Cicchetti and Rogosch, 2002; McLeod and Almazan, 2003). Adverse childhood experiences (e.g., physical abuse, household substance-related problems, and other household dysfunction) have been found to contribute to increased risk of substance use and mental health disorders in adulthood (Anda et al., 2002; Bontempo and D’Augelli, 2002; Dube et al., 2002; Dube et al., 2003; Hughes et al., 2010; McLaughlin et al., 2012).

Growing evidence indicates that sexual minorities experience greater childhood adversity than heterosexuals (Hughes et al., 2010; McLaughlin et al., 2012; Alvy et al., 2013; Zou and Andersen, 2015). While measures of childhood household dysfunction vary across studies, these factors have been shown to serve as robust risk factors for substance use disorder (SUD) in adulthood (Anda et al., 2002; Dube et al., 2003). Experiences of early life stressors and household dysfunction may place sexual minorities at an increased risk of depressive symptoms, suicide, heavy drinking, cigarette smoking, and SUD (Hughes et al., 2010; McCabe et al., 2013; Medley et al., 2016; Kerridge et al., 2017) due to both environmental and inherited influences (Frisell et al., 2010). For example, some have posited that substance use behaviors are a means of self-medicating because these behaviors can serve as a short-term strategy to relieve painful feelings and cope with the stress of trauma, neglect, or other forms of family and household dysfunction (Khantzian, 1997; Dube et al., 2003; Felitti et al., 2019).

Sexual orientation is a complex construct that includes multiple dimensions (e.g., identity, attraction, and behavior). Among sexual minorities, risk of SUDs differ considerably across sexual orientation dimensions (McCabe et al., 2013; Demant et al., 2016). Much of prior research has focused solely on one dimension of sexual orientation (e.g., either behavior or identity). Findings from a growing number of studies highlight the importance of considering multiple dimensions of sexual orientation, and the need for a better understanding of variations in risk across sexual orientation dimensions (McCabe et al., 2013; Demant et al., 2016). There is preliminary evidence that U.S. sexual minorities have more extensive childhood histories of parental/household substance-related problems than their heterosexual counterparts (McCabe et al., 2013; Demant et al., 2016). However, relatively little is known about how risk for SUDs among those exposed to early household dysfunction (e.g., substance-related problems) might vary across sexual orientation subgroups (e.g., lesbian, gay, bisexual, heterosexual) or sexual orientation dimensions (e.g., identity, attraction, or behavior), and whether such exposure contributes to increased risk of SUD among sexual minorities.

Individuals who identify as heterosexual, but report same-sex attraction and/or same-sex behavior (i.e., sexual orientation discordance) are more likely to report risk factors for SUD such as higher levels of depression, perceived stress, alcohol misuse, and cigarette smoking (Drabble et al., 2005; Bauer et al., 2010; Talley et al., 2015; Lourie and Needham, 2017). Further, the association between sexual orientation discordance and risk factors associated with SUD appear to be stronger for women than men (Gattis et al., 2012; McCabe et al., 2018), although additional research is needed. Moreover, experiences of childhood adversity vary considerably by sex and sexual orientation. Given the aforementioned research, a more comprehensive examination of childhood adversity, mental health, sexual orientation dimensions, and the role that these play in SUD is necessary, including an examination of potential sex differences and implications for practitioners.

To our knowledge, no studies have assessed the associations of childhood household history of substance-related problems or other dysfunction with subsequent DSM-5 alcohol use disorder, tobacco use disorder, and SUDs in adulthood as a function of sexual orientation. The primary aims of the present study were to:

1. Estimate the prevalence of childhood household history of substance-related problems and other household dysfunction by sexual identity, sexual attraction, and sexual behavior; and

2. Examine associations between childhood household dysfunction and adulthood past-year SUDs across sexual orientation subgroups.

2. METHODS

2.1. Study Design

We analyzed data from the 2012–2013 National Epidemiologic Survey on Alcohol and Related Conditions (NESARC-III) that was collected using structured diagnostic face-to-face interviews with a large nationally representative sample of 36,309 U.S. adults ages 18 years and older. The NESARC-III is the primary source of information regarding DSM-5 SUDs (including alcohol, tobacco, and other drug use disorders) among the general civilian, non-institutionalized population of U.S. adults. The NESARC-III included the National Institute on Alcohol Abuse and Alcoholism Alcohol Use Disorder and Associated Disabilities Interview Schedule-5 (AUDADIS-5), a fully structured diagnostic interview of mental health, substance use, and other health conditions conducted with individuals in households. The household, person, and overall response rates were 72%, 84%, and 60.1%, respectively. The NESARC-III sample design, response rates, and weighting procedures are described elsewhere (Grant et al., 2015a). All NESARC-III study procedures received full human subjects review and IRB approval; the current study received IRB approval from the University of Michigan Health Sciences and Behavioral Sciences Institutional Review Board. Supplemental Table 1 presents estimated sociodemographic characteristics of the target NESARC-III population.

2.2. Measures

Sociodemographic characteristics

Sociodemographic characteristics included age (18–29 years, 30–44 years, 45–64 years, ≥65 years), sex (male, female), race (White, African American, Hispanic, other), educational attainment (high school degree or less, some college, college degree or higher), personal income, employment status, relationship/marital status, health insurance status, metropolitan statistical area (urban, rural), and U.S. Census geographical region (Northeast, South, Midwest, West).

Major dimensions of sexual orientation.

Sexual attraction was assessed by asking the following question: People are different in their sexual attraction to other people. Which category on the card best describes your feelings? (1) only attracted to females, (2) mostly attracted to females, (3) equally attracted to females and males, (4) mostly attracted to males, or (5) only attracted to males. Sexual behavior was assessed with the following two items: Have you had sex in the last 12 months? and During the last 12 months, did you have sex with only males, only females, or both males and females? Responses to the two questions about behavior were grouped into four past-year sexual behavior categories (only same sex, only other sex, both sexes, and did not have sex). Sexual identity was assessed by asking: Which of the categories on the card best describes you? (1) heterosexual (straight), (2) gay or lesbian, (3) bisexual, or (4) not sure. These three major dimensions of sexual orientation (attraction, behavior, and identity) were initially examined separately. Subsequent analyses examined results when the heterosexual-identified participants were stratified into (1) “exclusively heterosexual” defined as adults who identified as heterosexual and reported only opposite-sex sexual attraction and behavior and (2) heterosexual-identified adults with any same-sex attraction or behavior.

Household dysfunction

Household dysfunction was assessed by asking respondents about their experiences in their home before they were 18 years of age. These included the following six experiences: (1) had a parent or other adult living in your home who was an alcoholic or a problem drinker, (2) had a parent or other adult living in your home who had similar problems with drugs, (3) had a parent or other adult living in your home go to jail or prison, (4) had a parent or other adult living in your home treated or hospitalized for a mental illness, (5) had a parent or other adult living in your home attempt suicide, and (6) had a parent or other adult living in your home commit suicide. The survey defined an alcoholic or problem drinker as “a person who had physical or emotional problems because of drinking; problems with a spouse, family, or friends because of drinking; problems at work or school because of drinking; problems with the police because of drinking – like drunk driving; or a person who spent a lot of time drinking or being hungover.” Each item was coded “yes” or “no” and summed to yield a scale score ranging from zero to six (intraclass correlation coefficient [ICC] = 0.69, 95% confidence interval [CI] = 0.66, 0.73 and Alpha = 0.72) (Ruan et al., 2008). Consistent with prior research, we created a variable that combined the two questions about household history of alcohol and/or drug problems to form a measure of household history of any substance-related problems (Dube et al., 2002, 2003; Felitti et al., 2019).

Past-year DSM-5 alcohol use disorder

Past-year DSM-5 alcohol use disorder was assessed using the AUDADIS-5 and DSM-5 criteria. Consistent with the DSM-5, a past-year alcohol use disorder diagnosis was based on the presence of at least two of the 11 DSM-5 criteria (e.g., craving, tolerance, withdrawal) at the same time within the 12 months prior to the date of the survey (American Psychiatric Association, 2013). Reliability and validity of the DSM-based diagnosis of alcohol use disorder has been examined previously (Grant et al., 2015b; Grant et al., 2015c). DSM-5 alcohol use disorder criteria scales have demonstrated excellent reliability (ICC = 0.9) (Grant et al., 2015b). Similarly, past-year DSM-5 tobacco use disorder diagnosis was based on the presence of at least two of the 11 DSM-5 criteria at the same time within a 12-month period. Reliability and validity of the DSM-based diagnosis of tobacco use disorder has been established in prior psychometric studies (Grant et al., 2015c; Hasin et al., 2015).

Past-year DSM-5 any substance use disorder (SUD)

Past-year DSM-5 any substance use disorder (SUD) was assessed using the AUDADIS-5 and included substance-specific diagnoses for ten substances: alcohol, cannabis, cocaine, heroin, hallucinogens, inhalants, prescription opioids, sedatives/tranquilizers, stimulants, and other drugs (e.g., ecstasy, ketamine). Substance-specific diagnoses were made for the past-year and prior-to-past-year timeframes. Each SUD diagnosis required 2 or more of the 11 DSM-5 criteria at the same time in the 12 months preceding the interview. The test-retest reliability and validity of each AUDADIS-5 DSM-5 SUD diagnoses range from fair to good, and dimensional criteria scales range from fair to excellent (ICC = 0.5 – 0.9, respectively) (Grant et al., 2015b; Grant et al., 2015c; Hasin et al., 2015; Grant et al., 2016).

2.3. Data Analysis

All analyses in this paper were design-based, in that they explicitly accounted for the complex sample design features of the NESARC-III (weights, stratum codes, and cluster codes) when 1) computing weighted estimates for the target NESARC-III population and 2) estimating the sampling variances of the population estimates with respect to the complex sample design (Heeringa et al. 2017, Chapter 3). We began with descriptive analyses, estimating population distributions for each dimension of sexual orientation for men and women separately. We then estimated the prevalence of all measures of childhood household dysfunction separately for men and women, including prevalence of household history of alcohol or drug problems, incarceration, mental illness, suicide attempts, and suicide, as a function of each sexual orientation dimension. The associations between each sexual orientation dimension, the covariates, and each household history indicator were tested for men and women separately using design-based Rao-Scott tests of significance. Given the number of tests of association performed, we used p < 0.005 as the level of significance.

Next, we fitted logistic regression models to the household history indicators to examine whether significant bivariate associations remained significant when adjusting for relevant covariates, including age, race/ethnicity, educational level, personal income, employment status, relationship status, health insurance status, geographic location, and metropolitan statistical area. When fitting these models, we explicitly accounted for the complex sampling features of the NESARC-III, using the final survey weights to estimate the model coefficients and Taylor Series Linearization to estimate the sampling variances of the estimated coefficients based on the stratified cluster sample design (Heeringa et al., 2017). These models were also fitted separately for men and women.

Finally, we examined associations between past-year SUDs and both sexual orientation and the household dysfunction history indicator. We first analyzed these associations in a descriptive fashion, followed by regressing the SUD indicators on sexual orientation, the household history indicators, and all covariates. Design-adjusted Archer-Lemeshow goodness-of-fit tests were used to evaluate the quality of fit for each logistic regression model. Multiple comparisons of descriptive quantities across the five groups defined by sexual orientation employed Bonferroni adjustments, concluding that there were significant pairwise differences with p < 0.01. All analyses were performed using the Stata software (Version 15.1). Syntax is available upon request.

3. RESULTS

3.1. Sexual Identity

Table 1 presents estimated sample distributions for the three major dimensions of sexual orientation. The majority of both women and men identified as heterosexual (96%−97%) and approximately 3.0% of men and 3.6% of women identified as non-heterosexual (Table 1). When comparing the subgroups of individuals defined by sexual identity in terms of distributions on the other sociodemographic measures considered in this study (see Supplemental Table 1), we found that sexual minority women were significantly more likely to be older, racial/ethnic minorities, and living in urban areas, and they were significantly less likely to have health insurance or be married. Bisexual-identified women in particular tended to have lower educational attainment, employment, and personal income than lesbian-identified or heterosexual-identified women. Similar associations of sexual identity with the other covariates emerged for men, with the exceptions of race/ethnicity, health insurance, and urbanicity, where no associations were statistically significant. These associations underscore the need to adjust for sociodemographic characteristics when comparing sexual orientation subgroups.

Table 1.

Estimated distributions of sexual orientation dimensions among U.S. adults

	Women		Men
Sexual Orientation Dimensions	% (SE)a	Sample Sizeb	% (SE)a	Sample Sizeb
Sexual Identity
Lesbian/Gay	1.2% (<0.1)	265	1.8% (0.1)	321
Bisexual	1.9% (0.1)	422	0.8% (<0.1)	144
Not sure	0.6% (<0.1)	130	0.4% (<0.1)	69
Heterosexual	96.4% (0.2)	19,454	97.0% (0.2)	15,190
Sexual Attraction
Only same sex	2.6% (0.2)	530	2.8% (0.2)	492
Mostly same sex	0.5% (<0.1)	115	0.8% (<0.1)	141
Equally both sexes	2.0% (0.2)	444	0.7% (<0.1)	138
Mostly other sex	4.5% (0.2)	933	2.7% (0.2)	429
Only other sex	90.4% (0.3)	18,228	93.0% (0.3)	14,524
Sexual Behavior (past-year)
Only same-sex	2.0% (0.1)	426	3.2% (0.2)	545
Both sexes	0.6% (<0.1)	171	0.3% (<0.1)	56
Did not have sex	30.3% (0.5)	6,427	20.6% (0.4)	3,385
Only other sex	67.0% (0.5)	12,954	76.0% (0.5)	11,571

Source: 2012–2013 National Epidemiologic Survey on Alcohol and Related Conditions (NESARC-III)

^aWeighted estimates; standard errors estimated using Taylor Series Linearization..

^bUnweighted sample sizes.

As shown in Table 2, there was a higher prevalence of a childhood history of parental/household substance-related problems among sexual minority subgroups, especially sexual minority women. Although associations were weaker among men, the same general patterns emerged: sexual minority men reported a greater number of adverse parental/household histories than heterosexual men.

Table 2.

Prevalence estimates of parental/household history of alcohol and other drug problems by sexual identity, sexual attraction, and sexual behavior

	Women			Men
	Parental/household history of alcohol problems % (SE), n	Parental/household history of other drug problems % (SE), n	Parental/household history of alcohol or drug problemsa % (SE), n	Parental/household history of alcohol problems % (SE), n	Parental/household history of other drug problems % (SE), n	Parental/household history of alcohol or drug problemsa % (SE), n
Sexual identity
Lesbian/Gay	32.1% (3.7), 265	8.8% (1.7), 263	34.7% (3.8), 265	29.7% (2.7), 319	8.0% (1.9), 319	32.1% (2.7), 320
Bisexual	30.4% (3.4), 418	18.2% (2.7), 416	34.0% (3.5), 417	32.1% (4.4), 144	10.7% (4.3), 144	34.8% (5.0), 144
Not sure	32.7% (4.6), 128	12.7% (2.7), 129	37.1% (4.9), 130	28.5% (7.1), 68	3.0% (1.5), 68	28.6% (7.1), 68
Heterosexual	24.7% (0.5), 19394	6.0% (0.2), 19354	26.2% (0.5), 19384	21.4% (0.4), 15125	5.2% (0.2), 15113	22.8% (0.4), 15116
p-value	0.017	<0.001	0.001	<0.001	0.050	<0.001
Sexual attraction
Only same sex	24.4% (2.6), 529	7.8% (1.5), 528	27.8% (2.8), 529	24.0% (2.2), 490	6.5% (1.4), 487	26.0% (2.3), 490
Mostly same sex	30.8% (5.1), 113	9.7% (3.3), 113	33.9% (5.3), 114	32.8% (5.6), 141	7.4% (2.9), 141	36.2% (5.7), 141
Equally both sexes	25.7% (2.5), 443	13.2% (2.0), 441	28.6% (2.7), 442	25.4% (4.3), 136	6.3% (2.3), 137	26.2% (4.1), 137
Mostly other sex	32.0% (1.6), 931	13.2% (1.5), 926	34.9% (1.7), 930	25.1% (2.3), 426	8.1% (1.8), 426	27.4% (2.4), 425
Only other sex	24.6% (0.5), 18168	5.8% (0.2), 18132	26.0% (0.5), 18160	21.4% (0.4), 14462	5.2% (0.2), 14453	22.7% (0.5), 14455
p-value	<0.001	<0.001	<0.001	0.027	0.144	0.006
Sexual behavior
Only same sex	21.8% (2.6), 425	7.7% (1.4), 423	25.0% (2.8), 425	25.5% (2.3), 545	6.7% (1.3), 541	27.9% (2.3), 544
Both sexes	38.7% (4.6), 169	29.5% (4.3), 169	45.5% (4.7), 170	36.6% (9.6), 55	19.8% (7.7), 56	43.4% (9.2), 56
Did not have sex	22.1% (0.8), 6399	3.3% (0.2), 6397	22.7% (0.8), 6396	19.0% (0.8), 3371	3.8% (0.4), 3367	20.2% (0.9), 3370
Only other sex	26.2% (0.6), 12920	7.5% (0.3), 12884	28.1% (0.6), 12914	22.2% (0.5), 11521	5.6% (0.3), 11515	23.5% (0.5), 11514
p-value	<0.001	<0.001	<0.001	0.002	<0.001	<0.001

Source: 2012–2013 National Epidemiologic Survey on Alcohol and Related Conditions (NESARC-III)

P values based on Rao-Scott chi-square tests examining the bivariate associations between sexual orientation and each parental/household history outcome. Sample sizes not exactly equal to Table 1 due to unknown responses on the parental/household history items.

^aParental history of any substance problems included alcohol or other drug problems.

Table 3 presents estimated relationships of sexual identity with the odds of a childhood history of parental/household substance-related problems based on separate multivariable logistic regression models. With heterosexuals as the reference group in each model, many of the bivariate associations remained robust after adjusting for several other relevant covariates. Bisexual-identified women had nearly two times greater odds than heterosexual-identified women of reporting a childhood history of parental/household drug-related problems (AOR = 1.9, 95% CI = 1.3, 2.7). Consistent with the bivariate results, we observed fewer significant associations for men than women in the multivariable models. For example, gay-identified men had higher odds than heterosexual-identified men of reporting a parental/household history of any type of alcohol or drug-related problems (AOR = 1.4, 95% CI = 1.1, 1.9).

Table 3.

Estimated relationships of sexual orientation with the odds of parental/household history of alcohol and other drug problems, based on separate multiple logistic regression analyses for women and men

	Women			Men
	Parental/household history of alcohol problems AOR (95% CI)b	Parental/household history of other drug problems AOR (95% CI)b	Parental/household history of alcohol or drug problemsa AOR (95% CI)b	Parental/household history of alcohol problems AOR (95% CI)b	Parental/household history of other drug problems AOR (95% CI)b	Parental/household history of alcohol or drug problemsa AOR (95% CI)b
Sexual identity
Lesbian/Gay	1.4 (1.0–2.1)*	1.2 (0.8–1.8)	1.5 (1.0–2.1)*	1.4 (1.1–1.8)*	1.4 (0.8–2.3)	1.4 (1.1–1.9)**
Bisexual	1.2 (0.9–1.7)	1.9 (1.3–2.7)**	1.3 (0.9–1.7)	1.6 (1.0–2.4)*	1.5 (0.6–3.8)	1.6 (1.0–2.5)*
Not sure	1.5 (1.0–2.3)	1.6 (1.0–2.7)	1.6 (1.0–2.5)*	1.3 (0.7–2.6)	0.5 (0.2–1.4)	1.2 (0.6–2.4)
Heterosexual	Referent	Referent	Referent	Referent	Referent	Referent
GOF	P = 0.91	P = 0.92	P = 0.33	P = 0.94	P = 0.76	P = 0.83
Sexual attraction
Only same sex	1.0 (0.8–1.3)	1.3 (0.9–2.0)	1.1 (0.8–1.5)	1.1 (0.8–1.4)	1.2 (0.7–1.8)	1.1 (0.9–1.4)
Mostly same sex	1.4 (0.9–2.3)	1.3 (0.6–3.0)	1.4 (0.9–2.3)	1.6 (1.0–2.6)	1.4 (0.6–3.2)	1.7 (1.1–2.8)*
Equally both sexes	1.1 (0.8–1.4)	1.7 (1.2–2.5)**	1.1 (0.9–1.4)	1.1 (0.7–1.8)	1.0 (0.4–2.2)	1.1 (0.7–1.6)
Mostly other sex	1.4 (1.2–1.6)***	1.9 (1.4–2.4)***	1.4 (1.2–1.7)***	1.2 (0.9–1.6)	1.5 (0.9–2.4)	1.3 (1.0–1.6)
Only other sex	Referent	Referent	Referent	Referent	Referent	Referent
GOF	P = 0.88	P = 0.79	P = 0.40	P = 0.83	P = 0.50	P = 0.35
Sexual behavior
Only same sex	0.8 (0.6–1.1)	0.9 (0.6–1.4)	0.8 (0.6–1.1)	1.1 (0.9–1.4)	1.1 (0.7–1.6)	1.2 (0.9–1.5)
Both sexes	1.8 (1.2–2.7)**	2.9 (1.9–4.6)***	2.0 (1.4–2.9)**	1.6 (0.7–3.9)	2.5 (0.8–7.3)	2.0 (0.9–4.5)
Never had sex	0.9 (0.8–1.0)*	0.6 (0.5–0.8)***	0.8 (0.7–0.9)**	0.8 (0.7–0.9)***	0.8 (0.6–1.0)	0.8 (0.7–0.9)***
Only other sex	Referent	Referent	Referent	Referent	Referent	Referent
GOF	P = 0.44	P = 0.39	P = 0.33	P = 0.51	P = 0.92	P = 0.91

Source: 2012–2013 National Epidemiologic Survey on Alcohol and Related Conditions (NESARC-III).

Notes: AOR = Adjusted odds ratio; CI = Confidence interval; GOF = Goodness of fit.

^aParental history of any substance problems included parental alcohol or other drug problems.

^bAOR indicates odds ratios adjusted for race, age, educational level, personal income, employment status, relationship status, health insurance status, geographic location, and metropolitan statistical area; the results for these variables are not shown.

^*p < 0.05,

^**p < 0.01,

^***p < 0.001.

3.2. Sexual Attraction

We estimated that 7.0% of men and 9.6% of women reported sexual attraction other than “only other sex.” Fewer differences in childhood history of parental/household substance-related problems emerged for sexual minorities based on sexual attraction than on sexual identity. As shown in Table 2, the associations were much more robust for women, with none of the bivariate tests for men showing significance at the p < 0.005 level. Consistent with the bivariate results, we observed fewer significant associations for men in the multivariable models involving sexual attraction (see Table 3). Most notably, women who were attracted “mostly to the other sex” (indicating some same-sex attraction) had greater odds of reporting a childhood history of parental/household substance-related problems (AORs ranged from 1.4 to 1.9) compared to women who were attracted to “only the other sex.”

3.3. Sexual Behavior

We estimated that 3.5% of men and 2.6% of women reported any same-sex sexual behavior in the past 12 months. Adults who reported not having had sex in the past year had lower odds of having a childhood history of parental/household substance-related problems. As shown in Table 3, women who were engaged in sexual behavior with both women and men in the past 12 months had two times greater odds of reporting a childhood history of parental/household substance-related problems (AOR = 2.0, 95% CI = 1.4, 2.9) compared to women who engaged in sexual behavior with only men. In contrast, women who reported not having had sex in the past 12 months consistently had lower odds of reporting a history of parental/household alcohol and/or drug problems compared to women who engaged in sexual behavior with only men. Similarly, men who have not had sex in the past 12 months had significantly lower odds of reporting parental/household history of alcohol or drug-related problems (AOR = 0.8, 95% CI = 0.7, 0.9) as compared to men who engaged in sexual behavior with only women. We found each of these models to have a good fit to the observed data.

3.4. Associations of Sexual Orientation with Substance Use Disorders and Household Dysfunction

Table 4 shows that among women, those identifying as bisexual had significantly higher probabilities of reporting a parental/household history of incarceration, parental/household history of mental illness, or any other household dysfunction measure relative to heterosexual women. These patterns did not hold for men; there were no differences in these individual measures of household dysfunction between bisexual-identified men relative to exclusively heterosexual men. We found that men identifying as gay had a significantly higher probability of reporting any household dysfunction than men who were exclusively heterosexual (i.e., heterosexual identified men who reported only opposite-sex sexual attraction and behavior).

Table 4.

Weighted prevalence estimates of past-year substance use disorders and parental/household dysfunction based on sexual orientation

	Women					Men
	Lesbian (n = 265) % (SE)	Bisexual (n = 422) % (SE)	Not Sure (n = 130) % (SE)	Heterosexual w/ same-sex attraction or behavior (n = 1294) % (SE)	Exclusivelye heterosexual (n = 17845) % (SE)	Gay (n = 321) % (SE)	Bisexual (n = 144) % (SE)	Not Sure (n = 69) % (SE)	Heterosexual w/ same-sex attraction or behavior (n = 782) % (SE)	Exclusivelye heterosexual (n = 14228) % (SE)
Substance use disorders (past-year)
Alcohol use disorder	24.9 (2.5)ab	29.7 (3.0)a	26.3 (4.8)ab	19.3 (1.6)b	9.2 (0.3)c	26.6 (2.7)a	31.4 (5.5)ab	23.7 (7.0)ac	14.8 (1.5)c	17.5 (0.4)bc
Tobacco use disorder	27.3 (3.5)ab	36.3 (3.2)a	33.6 (5.1)ab	21.6 (1.4)b	16.2 (0.4)c	30.0 (3.6)ab	40.8 (5.5)a	27.5 (7.2)ab	19.2 (1.6)b	23.3 (0.6)b
Any substance use disorderd	28.2 (3.0)ab	33.0 (3.3)a	28.7 (5.0)ab	22.8 (1.7)ab	10.5 (0.4)c	28.5 (2.9)a	36.9 (5.2)ab	23.7 (7.0)ac	17.1 (1.4)c	19.6 (0.4)c
Household dysfunction
Alcohol problems	32.1 (3.7)	30.4 (3.4)	32.6 (4.6)	27.2 (1.5)	24.5 (0.5)	29.7 (2.7)a	32.1 (4.4)ab	28.6 (7.1)ab	22.4 (1.7)ab	21.4 (0.4)b
Drug problems	8.8 (1.7)ac	18.2 (2.7)b	12.7 (2.7)abc	10.2 (1.2)a	5.8 (0.2)c	8.0 (1.8)	10.7 (4.3)	3.0 (1.5)	6.0 (1.0)	5.2 (0.2)
Incarceration	13.1 (2.1)ab	17.1 (2.0)a	22.6 (4.1)a	11.2 (1.0)a	7.6 (0.2)b	11.1 (2.2)	12.7 (4.6)	8.8 (3.5)	7.6 (1.1)	7.1 (0.3)
Mental illness	8.6 (1.9)ab	12.3 (1.7)a	7.5 (2.0)ab	7.9 (1.0)ab	5.9 (0.2)b	4.9 (1.4)	6.6 (2.7)	10.4 (4.8)	6.6 (1.1)	4.5 (0.2)
Suicide attempt	5.8 (1.5)	6.3 (1.4)	5.1 (2.1)	5.5 (0.9)	3.6 (0.2)	6.3 (1.9)	6.9 (3.2)	1.7 (1.7)	3.3 (0.8)	2.5 (0.1)
Committed suicide	0.5 (0.3)	0.9 (0.5)	2.0 (1.3)	0.6 (0.3)	1.0 (0.1)	1.9 (1.1)	0.3 (0.3)	0.0 (0.0)	1.4 (0.5)	0.8 (0.1)
Any household dysfunction	43.1 (4.0)a	42.4 (3.4)a	43.3 (4.9)ab	35.3 (1.7)ab	30.6 (0.5)b	37.9 (3.0)a	38.8 (4.7)ab	36.2 (7.4)ab	30.5 (1.9)ab	27.0 (0.5)b
Household dysfunction (#)
None	57.9 (4.0)	58.7 (3.3)	57.8 (5.0)	65.7 (1.7)	70.0 (0.5)	63.2 (3.1)	61.4 (4.7)	63.8 (7.4)	70.4 (1.9)	73.6 (0.5)
One experience	27.6 (3.6)	19.4 (2.5)	22.1 (4.3)	19.6 (1.4)	19.7 (0.5)	24.0 (2.8)	22.9 (4.4)	23.7 (7.2)	19.8 (1.5)	17.9 (0.4)
2 or more experiences	14.4 (2.3)	21.9 (2.8)	20.1 (3.8)	14.7 (1.4)	10.4 (0.3)	12.8 (2.2)	15.7 (4.8)	12.5 (4.6)	9.8 (1.3)	8.6 (0.3)
Design-Adj Rao-Scott F: p-value	p < 0.001					p = 0.006

^{a,b c}Estimates with different superscripts for a given sex are significantly different from each other, after applying a Bonferroni correction for the number of comparisons being performed (p < 0.005). Estimates with the same superscript are not significantly different from each other after applying the Bonferroni correction.

^dAny substance use disorder include any non-tobacco substance use disorder including alcohol, cannabis, inhalant, cocaine, hallucinogen, heroin, prescription opioid, prescription sedative/tranquilizer, prescription stimulants, club drug, or other drug use disorder.

^eExclusively heterosexual refer to individuals who identified as heterosexual without same-sex attraction or same-sex behavior.

Table 4 also shows comparisons of the estimated probabilities of having past-year DSM-5 alcohol use disorder, tobacco use disorder, and any SUD across the sexual identity subgroups. Bisexual-identified women had significantly higher probabilities of past-year alcohol use disorder and tobacco use disorder than exclusively heterosexual women. In general, exclusively heterosexual women had significantly lower probabilities of past-year substance-related disorders than every other subgroup. We found only slight differences among men. Men identifying as gay had significantly higher probabilities of past-year alcohol use disorder and of any past-year SUD than those identifying as heterosexual, with or without same-sex attraction or behavior.

Descriptive results indicated that parental/household histories of substance-related problems may moderate differences among sexual orientation subgroups, with parental/household histories of alcohol or drug use tending to amplify the probability of a past-year SUD among sexual minority respondents compared to heterosexual respondents (results available upon request). Most of the descriptive findings remained robust in the multivariable models, each of which were found to have a good fit to the observed data (see Table 5). After adjusting for relevant covariates, in addition to the measures of household dysfunction, the larger differences observed among sexual minority women remained significant. For example, the odds were 110% higher for any SUD among bisexual women compared to exclusively heterosexual women (AOR = 2.1, 95% CI = 1.5, 3.0), and this pattern generally held for the sexual minority women subgroups compared to exclusively heterosexual women. In addition, the odds were 130% higher for any SUD among heterosexual women who reported same-sex attraction or same-sex behavior compared to exclusively heterosexual women (AOR = 2.3, 95% CI = 1.9, 2.7). In contrast, there were no statistically significant differences based on sexual orientation among men at the 0.01 level after adjusting for the covariates. In fact, the odds of any SUD for heterosexual-identified men with same-sex attraction or same-sex sexual behavior were lower than that of exclusively heterosexual men (AOR = 0.8, 95% CI = 0.6, 1.0).

Table 5.

Results of logistic regression models: relationships of sexual orientation and parental/household dysfunction with substance use disorders

	Women			Men
	Past-year alcohol use disorder AOR (95% CI)b	Past-year tobacco use disorder AOR (95% CI)b	Past-year substance use disordera AOR (95% CI)b	Past-year alcohol use disorder AOR (95% CI)b	Past-year tobacco use disorder AOR (95% CI)b	Past-year substance use disordera AOR (95% CI)b
Sexual orientation
Exclusively heterosexualc	Referent	Referent	Referent	Referent	Referent	Referent
Heterosexual same-sex attraction/behavior	2.1 (1.7–2.6)***	1.4 (1.2–1.6)***	2.3 (1.9–2.7)***	0.8 (0.6–1.0)	0.8 (0.6–0.9)*	0.8 (0.6–1.0)*
Lesbian/Gay	2.1 (1.6–2.8)***	1.8 (1.2–2.6)**	2.2 (1.6–3.0)***	1.1 (0.8–1.4)	1.2 (0.8–1.7)	1.0 (0.8–1.3)
Bisexual	2.1 (1.5–2.9)***	1.9 (1.4–2.5)***	2.1 (1.5–3.0)***	1.4 (0.9–2.4)	1.7 (1.0–2.8)*	1.5 (1.0–2.4)
Not sure	2.6 (1.5–4.6)**	2.1 (1.3–3.4)**	2.4 (1.4–4.2)**	1.2 (0.6–2.7)	0.9 (0.5–1.9)	1.0 (0.5–2.2)
Parental/household alcohol problems
No	Referent	Referent	Referent	Referent	Referent	Referent
Yes	1.5 (1.3–1.7)***	1.5 (1.3–1.7)***	1.5 (1.3–1.7)***	1.5 (1.3–1.8)***	1.5 (1.3–1.7)***	1.6 (1.4–1.8)***
Parental/household drug problems
No	Referent	Referent	Referent	Referent	Referent	Referent
Yes	1.1 (0.8–1.4)	1.2 (1.0–1.4)*	1.1 (0.9–1.4)	1.1 (0.9–1.3)	1.5 (1.3–1.9)***	1.1 (0.9–1.4)
Parental/household incarceration
No	Referent	Referent	Referent	Referent	Referent	Referent
Yes	1.4 (1.2–1.7)**	1.3 (1.1–1.5)**	1.3 (1.1–1.6)**	1.3 (1.0–1.5)*	1.2 (1.0–1.4)	1.4 (1.2–1.7)***
Parental/household mental illness
No	Referent	Referent	Referent	Referent	Referent	Referent
Yes	1.2 (1.0–1.5)	1.1 (0.9–1.4)	1.2 (1.0–1.5)	1.0 (0.7–1.3)	0.9 (0.7–1.2)	1.1 (0.8–1.5)
Parental/household suicide attempt
No	Referent	Referent	Referent	Referent	Referent	Referent
Yes	1.2 (0.9–1.7)	1.5 (1.1–1.9)*	1.2 (0.9–1.6)	1.1 (0.8–1.5)	1.6 (1.1–2.3)**	1.0 (0.8–1.4)
Parent/household committed suicide
No	Referent	Referent	Referent	Referent	Referent	Referent
Yes	0.9 (0.5–1.6)	0.7 (0.4–1.1)	1.0 (0.5–1.9)	0.8 (0.4–1.4)	0.7 (0.4–1.2)	0.8 (0.4–1.4)
Sample Size / GOF test p-value	19,659, p = 0.10	19,659, p = 0.08	19,659, p = 0.27	15,331, p = 0.04	15,331, p = 0.90	15,331, p = 0.63

^*p < 0.05,

^**p < 0.01,

^***p < 0.001.

Notes: AOR = Adjusted odds ratio; CI = Confidence interval; GOF = Goodness of fit.

^aPast-year substance use disorder refers to alcohol, cannabis, sedatives, tranquilizers, opioids, stimulants, cocaine, hallucinogens, inhalants, or heroin use disorder.

^bAOR indicates odds ratios adjusted for race, age, educational level, personal income, employment status, relationship status, health insurance status, geographic location, and metropolitan statistical area; the results for these variables are not shown.

^cExclusively heterosexual refer to individuals who identified as heterosexual without same-sex attraction or same-sex behavior.

For both men and women, a history of household alcohol-related problems significantly increased the odds of alcohol use disorder, tobacco use disorder, and any SUD in adulthood (p < 0.001). A history of parental/household incarceration also emerged as a robust predictor of past-year SUDs for both men and women. We also found that a parental/household suicide attempt increased the odds of a past-year tobacco use disorder for both women and men. Finally, we found evidence that any household dysfunction increased the probability of each type of SUD, for both men and women; there were no differences in the odds of each type of SUD between those with only one experience of household dysfunction versus those with multiple experiences of household dysfunction (results available upon request).

4. DISCUSSION

Parental substance-related problems and other household dysfunction during childhood often co-occur in families and households (Anda et al., 2002). In this study, higher rates of childhood household dysfunction were associated with greater prevalence of past-year SUDs among adult sexual minorities. Greater exposure to substance-related problems in childhood among sexual minorities has important implications for SUDs due to both inherited and environmental influences that warrant more attention. The odds of tobacco use disorder and any SUD among bisexual-identified women were about two times greater compared to heterosexual-identified women. This pattern generally held among all sexual minority women (whether based on attraction, behavior, or identity) compared to heterosexual women, aligning with previous research (McCabe et al., 2009; Demant et al., 2016; Medley et al., 2016; Kasza et al., 2017; Kerridge et al., 2017). These findings are also consistent with previous evidence that sexual minority women show higher rates of alcohol and tobacco use than heterosexual women (Hughes et al., 2016).

Results of multivariable analyses showed that sexual minority women had a higher risk of alcohol use disorder, tobacco use disorder, and any SUD when compared to exclusively heterosexual women, even when adjusting for parental/household history of substance-related problems and other household dysfunction. These findings suggest that factors other than household dysfunction may also contribute to SUD among sexual minority women, such as other forms of childhood abuse, victimization, non-adherence to traditional female roles, and discrimination in adulthood (Balsam et al., 2005; Hughes et al., 2010; McCabe et al., 2010). Notably, sexual orientation differences in alcohol use disorder, tobacco use disorder, and SUD risk between sexual minority men and exclusively heterosexual men were minimal after adjusting for the covariates. The sex differences in the findings highlight the limitations in studies that combine sexual minority women and men in analyses.

The risks of alcohol use disorder, tobacco use disorder, and any SUD were elevated among heterosexual-identified women with same-sex attraction or same-sex sexual behavior (all associations, p < 0.001). In contrast, no such increased risks were found among heterosexual-identified men with same-sex attraction or same-sex sexual behavior relative to exclusively heterosexual men. Indeed, the odds of tobacco use disorder and any SUD for heterosexual-identified men with same-sex attraction or same-sex sexual behavior were significantly lower than those of exclusively heterosexual men. The increased risk of tobacco use disorder is highly important given that tobacco use remains the leading preventable cause of death in the U.S. These new findings add to a growing literature indicating that sexual orientation discordance among women is a more salient risk factor for alcohol use disorder, tobacco use disorder, and other SUDs than it is among men (Gattis et al., 2012; Talley et al., 2015; McCabe et al., 2018). These divergent results by sex could stem from gendered differences in exposure to proximal and distal stressors. For example, identity concealment and other proximal internal stressors could operate differently by sex and discordance could create greater stress in women than in men (Pachankis, 2007; Saladin et al., 2012).

There were some notable strengths and limitations that should be carefully considered when evaluating the implications of these findings. A major strength was the nationally representative sample of U.S. adults that was large enough to stratify by sex and sexual orientation subgroup. Another strength was the use of reliable and valid measures of alcohol use disorder, tobacco use disorder, and other SUDs based on DSM-5 criteria along with household histories of substance-related problems and other types of household dysfunction. In terms of limitations, causal inferences were not possible given the cross-sectional design of the study. The risks of adverse experiences were not assessed longitudinally in the NESARC-III, but these risks may accumulate over time. This warrants prospective designs to understand the temporal ordering of household dysfunction, stress, substance use, and sexual orientation. In addition, the NESARC-III did not assess gender identity or sexual minority-specific factors, such as internalized homophobia, degree of “outness”, and family rejection that may influence risk of SUDs (Pachankis et al., 2015). Future research is needed that includes gender identity measures and more detailed sexual minority-specific factors. In addition, there is evidence that risk factors for SUD among sexual minorities differ across age groups (Evans-Polce et al., in press) and future studies could consider oversampling sexual minorities, enabling larger samples for age-stratified analysis. Finally, the prevalence of SUDs was likely underestimated in the NESARC-III sample because small but high-risk groups of currently institutionalized individuals, such as incarcerated adults, were not included.

5. CONCLUSIONS

In conclusion, the novel results of the present study have important clinical implications, reinforcing that childhood household history of substance-related problems and other childhood household dysfunction are important factors to consider in the context of assessment, prevention and treatment of SUDs. The major clinical implications of this work are driven by the compelling finding that sexual minorities are more likely than heterosexuals to experience parental/household history of substance-related problems and other household dysfunction during childhood. In turn, this increased likelihood of childhood household dysfunction among sexual minorities is associated with a higher risk of developing alcohol, tobacco, and other substance use disorders in adulthood, especially among sexual minority women.

These findings should be taken into account by health care providers, particularly health professionals who treat individuals with alcohol, tobacco, and other substance use disorders. At a minimum, sexual orientation (attraction, behavior, and identity), household dysfunction, and other adverse childhood experiences can be assessed during initial health examinations while collecting family and personal history information. Healthcare providers who care for individuals raised in dysfunctional households should carefully assess risk for mental health and intervene as needed.