Prospective associations between childhood exposure to living with adult alcohol misuse and major depressive disorder in adulthood: The role of child maltreatment

Abstract

Background:

Many children grow up with adult alcohol misuse in the home. A clearer understanding of this exposure’s long-term mental health consequences and the role of associated child maltreatment experiences and potential protective factors could guide relevant intervention strategies.

Objective:

To prospectively evaluate the association between living with adult alcohol misuse during childhood and major depressive disorder (MDD) during adulthood; whether child maltreatment explains the association; and whether sex, school bonding, or neighborhood bonding moderate the association.

Participants and setting:

This study used longitudinal data from 783 individuals followed from childhood to age 39.

Methods:

At grade 9, participants were asked whether they lived with adults who misused alcohol. Diagnostic assessments of MDD were conducted across three time-points during participants’ thirties and participants were categorized as having met diagnostic criteria 0, 1, or 2 or more times.

Results:

Ordinal logistic regressions found that children living with adult alcohol misuse showed greater chronicity of adult MDD (OR = 1.63; 95% CI: 1.05, 2.52). There was a 49% reduction in the odds ratio and the association was no longer statistically significant when child maltreatment was included in the model (OR= 1.32; 95% CI: 0.84, 2.07). No statistically significant moderation of associations was observed.

Conclusions:

Children exposed to adult alcohol misuse, and maltreatment often associated with this misuse, may be at risk for mental health challenges well into adulthood. Interventions that address childhood exposure to adult alcohol misuse and associated maltreatment may be important to mitigate long-term mental health challenges to exposed children.

Introduction

Depression is a major public health concern in the United States. One national study found the lifetime prevalence of major depressive disorder (MDD) to be 16.6% among U.S. adults (Kessler et al., 2005). Etiologic studies of depression have identified numerous risk factors including genetic propensity, personality traits, stressful life events, and environmental factors (Saveanu & Nemeroff, 2012). Less understood is the developmental timing of risk factors, specifically the role of childhood factors that may place individuals at higher risk for experiencing MDD later in life.

Adverse childhood experiences (ACEs) have consistently been found to place individuals at risk for life-long consequences including poor mental health in adulthood (Mwachofi et al., 2020; Reinherz et al., 2003; Lee & Chen, 2017). Additionally, considerable evidence suggests a positive relationship between number of ACEs and risk of poor mental health in adulthood (Hughes et al., 2016). ACEs typically include ten constructs: physical abuse, emotional abuse, sexual abuse, physical neglect, emotional neglect, parental separation, parental incarceration, domestic violence, household mental illness/suicide, and household substance abuse (Bethell et al., 2017). Associations between ACEs and poor mental health outcomes have been demonstrated, yet individual effects of specific ACEs, such as living with an adult who misuses alcohol, on adverse mental health outcomes in adulthood requires further examination, especially since the majority of studies are cross-sectional.

One in every four children in the U.S. is exposed to alcohol abuse or dependence in the family (Grant, 2000). Research suggests these youth experience short- and long-term consequences, such as poorer academic functioning; emotional, behavioral, and social problems; developmental delay; mental health conditions in childhood including depression; and later substance use problems in adolescence and early adulthood (Jones et al., 2016; Guttmannova et al., 2017; Hill et al., 1996; Vidal et al., 2012; M. Solis et al., 2012; Brummer et al., 2021). However, few studies have evaluated the specific association between living with adult alcohol misuse (rather than substance misuse more broadly) during childhood and depression in adulthood. A recent cross-sectional study of adults (mean age 54 years) found that those who retrospectively reported having lived with adult alcohol misuse during childhood experienced more days of poor mental health and were more likely to be diagnosed with any depressive disorder in their lifetime (Mwachofi et al., 2020). Further, a few prospective studies with limited follow-up length found that youth with a family history of adult alcohol misuse or substance use were more likely to have a depressive disorder diagnosis during adolescence (ages 12–18; Hill et al., 2008) and in the transition to adulthood (ages 18–27; Reinherz et al., 2003; Chassin et al., 1999; Marmorstein et al., 2012). We are aware of no prospective studies specifically linking exposure to adult alcohol misuse during childhood with MDD later in adulthood. As prevalence of MDD peaks in early adulthood and declines with age (Substance Abuse and Mental Health Services Administration, 2020), childhood experiences that place one at increased risk beyond one’s 20s may serve as particularly important targets for early intervention.

An association between living with adult alcohol misuse during childhood and depression in adulthood is supported by the social development model (SDM). The SDM organizes risks and protective factors into etiologic pathways that predict behavior across development, theorizing that youth learn patterns of behavior from socializing units including families, schools, and peers (Catalano & Hawkins, 1996). The model hypothesizes that involvement with positive socializing units promotes prosocial bonds and values which facilitate positive youth development, and that involvement with socializing units that engage in antisocial behavior, such as adults who misuse alcohol, promotes antisocial bonds and values that increase poor health outcomes. Empirical support for the SDM has been found for several health and behavioral outcomes including depression (Mason et al., 2010).

As child maltreatment is more common in households where adult substance misuse is present (Clemens et al., 2020), it could be an important explanatory factor linking living with adult alcohol misuse during childhood to later mental health. However, this potential mechanism has not been well-studied. Recent cross-sectional research found that child maltreatment partially explained the relationship between childhood exposure to household substance misuse and depression as well as other outcomes including anxiety, physical health problems, and decreased life satisfaction (Clemens et al., 2020; Clemens et al., 2019). Yet, to our knowledge, no longitudinal studies have examined the extent to which child maltreatment is a key mechanism linking living with adult alcohol misuse during childhood and MDD in adulthood.

It is also important to consider protective factors that could mitigate long-term harms on mental health associated with living with adult alcohol misuse. Prior research suggests that social factors outside the family (e.g., community engagement, school attachment) may be important protective factors for children of parents who misuse alcohol (Velleman & Templeton, 2016), especially since children who live with adult alcohol misuse may not be receiving adequate social support at home. Further, a recent systematic review found that high academic ability, social support, and presence of teachers and mentors who promote academic skills, self-esteem, and confidence are protective against adverse outcomes for children of parents with alcohol use disorder (Park & Schepp, 2015). Given the potential for long-term, adverse effects, it is particularly important to understand if and how mutable factors, such as school and neighborhood environments, might buffer against deleterious effects of childhood exposure to adult alcohol misuse and thereby serve as intervention targets (Hawkins et al., 2005). Evaluating protective factors in the school and neighborhood environment is also supported by the SDM framework as bonding to positive socializing units can facilitate positive youth development (Catalano et al., 1996). No prior studies have evaluated whether school or neighborhood bonding have a protective effect on adulthood MDD following childhood exposure to adult alcohol misuse.

The association between living with adult alcohol misuse during childhood and depression in adulthood may differ by sex. Prior studies of children exposed to adult alcohol misuse in their household during childhood found that females were more vulnerable to internalizing symptoms compared to males who exhibited more externalizing symptoms such as aggressive behaviors (Park & Schepp, 2015). Although females are more likely to internalize symptoms, there is evidence that coping strategies for stressful situations during childhood differ between girls and boys due to learned gender norms (e.g., girls are more likely to seek social support, boys are more likely to use avoidant strategies) (Eschenbeck et al., 2007). Further, one prospective cohort study found that adults who lived with alcohol misuse during childhood and who coped effectively relied on significantly larger numbers of sources of support in their childhood and youth compared to those that did not, and that daughters turned to significantly larger numbers of caring adults during childhood and youth than sons (Werner & Johnson, 2004). Thus, differences in coping strategies by sex may manifest as downstream sex differences in adult mental health of those who experienced adult alcohol misuse in childhood. There have been mixed results from the limited studies that have examined sex differences in mental health among children who lived with adult alcohol misuse during childhood, and further prospective studies are needed.

This study used a prospective cohort design to evaluate the association between living with adult alcohol misuse during childhood and chronicity of MDD across three time points during participants’ thirties. Next, we evaluated the extent to which child maltreatment explained the hypothesized association. We also examined whether our association of interest was moderated by school and neighborhood bonding during childhood and by sex. We hypothesized that living with adult alcohol misuse during childhood would be associated with higher odds of meeting MDD diagnostic criteria in adulthood, that child maltreatment would partially explain the association, and that the association would be stronger among those with low relative to high neighborhood and school bonding during childhood and among females relative to males.

Methods

Study Sample

Data for this study were from the Seattle Social Development Project (SSDP) (Catalano et al., 2021), a longitudinal study beginning in 1985 that studied health behavior and outcomes across development. All students entering the 5^th grade (mean age 10.01, SD=.55) in 18 public schools serving higher crime neighborhoods in Seattle, Washington, were eligible. Of the 1,053 eligible students, 808 (77%) youth and their parents consented to participate in the longitudinal study. Data were collected annually from ages 10–16 and in eight additional assessments from ages 18–39. Study subjects for this analysis included those enrolled in the longitudinal study and who participated in the grade 9 interview (n=783). The current study included data from grade 9 and ages 24, 30, 33, and 39 assessments. At grade 9 and age 24, assessments were conducted in person at the participant’s home and interviews were administered by trained staff. At ages 30, 33, and 39, assessments combined in-person surveys, password-protected web-based surveys, or by request a telephone or paper survey. Informed consent was obtained from all participants and the study was approved by the Human Subjects Review Committee at the University of Washington.

Study Measures

Living with adult alcohol misuse during childhood.

This was measured during the grade 9 (mean age 15.5, SD=.6) interview and based on the question “Have you ever lived with an adult who in your judgment was an alcoholic or a problem drinker (while you were living with them)?” [1=yes; 0=no].

Major Depressive Disorder (MDD) diagnoses during adulthood.

DSM-IV-based diagnosis of past year MDD was assessed at age 30, 33, and 39 interviews using the Diagnostic Interview Schedule (DIS-IV) (Robins et al., 1981; American Psychiatric Association, 1994). The DIS is frequently used in epidemiological studies of psychiatric disorders among adults and has shown strong reliability and validity when assessed against a clinical diagnosis (Newman et al., 1996). The outcome was operationalized based on chronicity of MDD across three waves as follows: 0=no MDD diagnosis; 1= one MDD diagnosis across the three waves; 2=two or more MDD diagnoses across the three waves.

Child maltreatment.

At the age 24 interview (1999), a 25-item Retrospective Child Nurturance and Abuse Questionnaire was included to capture information on subjects’ experience of physical abuse, emotional abuse, sexual abuse, physical neglect, and emotional neglect prior to age 18. Responses to items were measured on a five-point Likert scale (0=never true; 1=rarely true; 2=sometimes true; 3=often true; 4=always true). One item from the questionnaire was excluded as it indirectly inquired about substance misuse in the home, and therefore overlapped with our exposure of interest. A continuous child maltreatment variable was created by summing the scores from the included 24 items, with higher scores representing more chronic and severe child maltreatment experiences (range: 0–96).

School and neighborhood bonding.

At the grade 9 assessment, school bonding was measured on a 4-point Likert scale (1=NO!; 2=no; 3=yes; 4=YES!) in response to the following five statements: 1) “I like school”; 2) “I like my social studies teacher this year”; 3) “I like my other teachers this year”; 4) “Most mornings I look forward to going to school”; and 5) “I like my classes this year.” This scale showed good internal consistency (Cronbach’s α = 0.76). The scores of the five school bonding items were summed (range: 4–20) and then dichotomized for ease of interpretation in analyses (1=high level of bonding [total score $\ge$15]; 0=low level of bonding [total score <15]). Neighborhood bonding was measured using the same 4-point Likert scale in response to the following four statements: 1) “I like my neighborhood”; 2) “If I had to move, I would miss the neighborhood I now live in”; 3) “I want to stay in my neighborhood no matter what”; and 4) “I’d like to get out of my neighborhood” (reverse coded). This scale also showed good internal consistency (Cronbach’s α = 0.84). The scores of the four neighborhood bonding items were summed (range: 4–16) then dichotomized (1=high level of bonding [total score $\ge$12]; 0=low level of bonding [total score <12]).

Confounding variables.

The following potential confounders were selected a priori for adjustment in study analyses and were measured at grade 9 (1990) unless otherwise specified: parent report of highest level of education completed (1=some high school or less, 2=high school grad or GED, 3=some college or trade/business school, 4=college grad), parent report of annual household income (1=under $9,999, 2=$10,000 - $19,999, 3=$20,000 - $29,999, 4=$30,000 - $39,999, 5=$40,000 or more), parent report of marital status (1=married, 0=not married), indicator variables for participant’s race based on most commonly reported race from school records and multiple self-reports over the course of study (Caucasian [reference], African American, Native American, Asian American), and participant’s report of family history of depression report (1=yes, 0=no; based on the interview question “Has anyone in your family had any of the following illnesses or conditions? Depression?” asked at age 33). Race was included as a potential confounder in this study as a proxy for exposure to racial discrimination, which may lead to increased alcohol use as a coping strategy (Gilbert & Zemore, 2016).

Statistical Analyses

Ordinal logistic regression was used to evaluate the three study aims. Odds ratios estimated describe the proportional change in the odds of a one-unit increase in the outcome of interest (e.g., one MDD diagnosis compared to none, two or more MDD diagnoses compared to one) among those who lived with adult alcohol misuse during childhood relative to those who did not. For our primary aim, the model included our binary exposure and confounding variables as covariates (Model 1). For our second aim, the model included Model 1 variables plus the continuous child maltreatment variable (Model 2). We evaluated the extent to which the odds ratio associated with the exposure was attenuated in Model 2 compared to the corresponding odds ratio in Model 1 (unadjusted for child maltreatment) as an indication of how much child maltreatment explained, at least in part, the hypothesized association. For our third aim, we ran three models that separately examined moderation of our exposure-outcome association by school bonding, neighborhood bonding, and sex. Moderation was assessed by Wald’s tests for the interaction terms between our exposure and each potential moderator.

Among the 783 children who participated in the grade 9 interview, 94% completed the interview at age 24, 90% completed the interview at ages 30 and 33, and 85% completed the interview at age 39. Although participant retention was fairly high considering follow-up length, missingness could still lead to important biases. To account for missingness, we used multiple imputation by chained equations (Graham, 2009). The imputation model included our exposure, outcome, confounding variables, and six auxiliary variables: child report of alcohol use in grade 9 (yes/no); child report of marijuana use in grade 9 (yes/no); DSM-IV-based diagnosis of alcohol dependence or abuse at ages 30, 33, and 39 (yes/no); and subject’s annual income at age 33 (continuous). The imputation model also specified relevant interactions (e.g., exposure-x-moderators). Twenty datasets were imputed. Regression models were run within each of the twenty datasets and parameter estimates and their standard errors were combined across the datasets using Rubin’s rules that account for uncertainty between and within imputations (Rubin, 2004).

In elementary grades, part of the SSDP sample received the Raising Healthy Children preventive intervention (Hawkins et al., 2005). To assess whether the intervention influenced differences in the exposure-outcome association, we tested the exposure*intervention interaction in sensitivity analyses. Statistically significant moderations were not observed; and, thus, we report full sample results.

All statistical analyses were conducted in R version 4.0.3 (R Core Team, 2020) and the mice package was used for multiple imputation (Van Buuren & Groothuis-Oudshoorn, 2011). Statistical significance was evaluated at an alpha=0.05 level.

Results

Frequency of responses to each child maltreatment item are shown in Table 1. Distribution of sample characteristics by childhood exposure to adult alcohol misuse as averaged across the twenty imputed datasets are shown in Table 2. The percentage of the study sample who reported living with adult alcohol misuse at grade 9 was 28.8%. In bivariate analyses, those who reported living with adult alcohol misuse compared to those who did not were significantly more likely to: be female, White, and Native American; and have lived in a household with lower annual income, family history of depression, unmarried parents, lower combined parental education attainment, lower school and neighborhood bonding, higher child maltreatment scores, and have at least one report of MDD during their thirties (Table 2). Similar prevalence estimates of sample characteristics by exposure status were observed using the non-imputed data.

Table 1.

Prevalence^a of child maltreatment indicators by subtype from the Retrospective Child Nurturance and Abuse Questionnaire^b.

Child Maltreatment Indicator	Total N (%)
Physical Abuse
I got hit so hard by my mother or father (or other primary caregiver) that I had to see a doctor or go to the hospital.
Never True	643 (92.5)
Rarely True	28 (4.0)
Sometimes True	14 (2.0)
Often True	7 (1.0)
Always True	3 (0.4)
My mother or father (or other primary caregiver) hit me so hard that it left me with bruises or marks.
Never True	538 (77.3)
Rarely True	78 (11.2)
Sometimes True	56 (8.0)
Often True	20 (2.9)
Always True	4 (0.6)
I was punished with a belt, a board, a cord, or some other hard object.
Never True	324 (46.4)
Rarely True	125 (17.9)
Sometimes True	160 (22.9)
Often True	60 (8.6)
Always True	30 (4.3)
I believe that I was physically abused.
Never True	589 (84.5)
Rarely True	35 (5.0)
Sometimes True	40 (5.7)
Often True	16 (2.3)
Always True	17 (2.4)
I got hit or beaten so badly that it was noticed by someone like a teacher, neighbor, or doctor.
Never True	649 (93.0)
Rarely True	23 (3.3)
Sometimes True	16 (2.3)
Often True	4 (0.6)
Always True	6 (0.9)
Emotional Abuse
My mother or father (or other primary caregiver) called me things like “stupid” or “lazy” or “ugly”.
Never True	357 (50.6)
Rarely True	155 (22.0)
Sometimes True	140 (19.9)
Often True	43 (6.1)
Always True	10 (1.4)
My mother or father (or other primary caregiver) said hurtful or insulting things to me.
Never True	366 (52.4)
Rarely True	152 (21.7)
Sometimes True	130 (18.6)
Often True	41 (5.9)
Always True	10 (1.4)
I believe that I was emotionally abused.
Never True	490 (70.1)
Rarely True	67 (9.6)
Sometimes True	79 (11.3)
Often True	40 (5.7)
Always True	23 (3.3)
Sexual Abuse
My mother or father (or other primary caregiver) tried to touch me in a sexual way, or tried to make me touch them.
Never True	567 (82.1)
Rarely True	52 (7.5)
Sometimes True	42 (6.1)
Often True	20 (2.9)
Always True	10 (1.4)
My mother or father (or other primary caregiver) threatened to hurt me or tell lies about me unless I did something sexual with them.
Never True	648 (93.5)
Rarely True	17 (2.5)
Sometimes True	17 (2.5)
Often True	10 (1.4)
Always True	1 (0.1)
My mother or father (or other primary caregiver) tried to make me do sexual things or watch sexual things.
Never True	592 (85.9)
Rarely True	39 (5.7)
Sometimes True	39 (5.7)
Often True	16 (2.3)
Always True	3 (0.4)
My mother or father (or other primary caregiver) molested me.
Never True	578 (84.0)
Rarely True	44 (6.4)
Sometimes True	27 (3.9)
Often True	10 (1.5)
Always True	29 (4.2)
I believe that I was sexually abused.
Never True	588 (85.2)
Rarely True	33 (4.8)
Sometimes True	22 (3.2)
Often True	11 (1.6)
Always True	36 (5.2)
Physical Neglect
There was someone to take me to the doctor if I needed it.
Never True	7 (1.0)
Rarely True	11 (1.6)
Sometimes True	28 (4.0)
Often True	95 (13.5)
Always True	562 (79.9)
I had to wear dirty clothes.
Never True	608 (87.1)
Rarely True	46 (6.6)
Sometimes True	33 (4.7)
Often True	5 (0.7)
Always True	6 (0.9)
I knew that there was someone to take care of me and protect me.
Never True	13 (1.8)
Rarely True	27 (3.8)
Sometimes True	46 (6.5)
Often True	115 (16.3)
Always True	503 (71.4)
I didn’t have enough to eat.
Never True	606 (87.1)
Rarely True	41 (5.9)
Sometimes True	37 (5.3)
Often True	8 (1.1)
Always True	4 (0.6)
Emotional Neglect
I felt that I was loved.
Never True	6 (0.9)
Rarely True	20 (2.8)
Sometimes True	76 (10.8)
Often True	143 (20.3)
Always True	458 (65.1)
I thought that my parents wished I had never been born.
Never True	599 (85.9)
Rarely True	45 (6.5)
Sometimes True	38 (5.5)
Often True	12 (1.7)
Always True	3 (0.4)
There was someone in my family who helped me feel that I was important or special.
Never True	27 (3.8)
Rarely True	23 (3.3)
Sometimes True	105 (15.0)
Often True	211 (30.1)
Always True	336 (47.9)
I felt that my mother or father (or other primary caregiver) hated me.
Never True	520 (74.3)
Rarely True	70 (10.0)
Sometimes True	62 (8.9)
Often True	28 (4.0)
Always True	20 (2.9)
People in my family looked out for one another.
Never True	24 (3.4)
Rarely True	35 (5.0)
Sometimes True	95 (13.5)
Often True	221 (31.5)
Always True	327 (46.6)
People in my family felt close to each other.
Never True	28 (4.0)
Rarely True	53 (7.5)
Sometimes True	151 (21.4)
Often True	255 (36.2)
Always True	217 (30.8)
My family was a source of strength and support.
Never True	40 (5.7)
Rarely True	66 (9.3)
Sometimes True	130 (18.4)
Often True	197 (27.9)
Always True	273 (38.7)

^aPrevalence from the observed dataset reflect participants who completed the item; missing values ranged from 9.8% to 12.1% across items.

^bRetrospective Child Nurturance and Abuse Questionnaire was administered at age 24 interview and participants reported how often each of the child maltreatment indicators happened before the age of 18.

Table 2.

Prevalence^a of participant characteristics by childhood exposure of adult alcohol misuse.

Characteristicb		Have you ever lived with an adult who in your judgment was an alcoholic or a problem drinker (while you were living with them)?
	Total	Yes		No
	N = 783 %	n = 226 %	n = 557 %
Female sex	49.0	55.1	46.6
Race/Ethnicity
Caucasian	47.5	57.4	43.5
African American	25.2	22.4	26.3
Native American	5.2	10.1	3.3
Asian American	22.1	10.1	27.0
Annual parental income, $
< 9,999	14.1	18.2	12.5
10,000 – 19,999	20.5	24.3	19.0
20,000 – 29,999	19.5	17.0	20.5
30,000 – 39,999	17.3	15.9	17.9
40,000 or more	28.5	24.6	30.1
Family history of depression	38.9	54.8	32.5
Parents married	60.3	49.4	64.8
Highest level of parental education completed
Some high school or less	15.1	15.6	14.8
High school grad or GED	21.7	24.4	20.7
Trade or business school or some college	37.1	39.4	36.2
College graduate	26.1	20.6	28.3
School bonding, high	61.0	54.5	63.6
Neighborhood bonding, high	67.2	58.6	70.7
Child maltreatment, mean	12.6	17.5	11.4
Number of MDD diagnoses during 30s
0	78.1	67.9	82.2
1	14.0	17.9	12.4
≥2	7.9	14.2	5.4

Abbreviations: MDD, major depressive disorder

^aEstimates averaged across 20 imputed datasets

^bDifferences in covariates across exposure status were all statistically significant at an alpha=0.05 level based on the chi-square test for categorical covariates and a t-test for the continuous child maltreatment covariate

Odds ratios from multivariable ordinal logistic regression models are shown in Table 3. Results from models adjusting for confounding variables (Model 1) showed living with adult alcohol misuse during childhood was associated with 63% higher odds for a one-unit increase in reported MDD during study follow-up (OR = 1.63, 95% CI: 1.05, 2.52). To aid in interpretation, the model-predicted prevalence estimates of reporting 0, 1, and 2 or more MDD diagnoses among those exposed were 84%, 11%, and 5%, respectively, compared to 89%, 8%, and 3%, among those unexposed. When child maltreatment was included in the model (Model 2), there was a 49% reduction in the odds ratio compared to Model 1 results and the association was no longer statistically significant (OR= 1.32, 95% CI: 0.84, 2.07).

Table 3.

Odds ratios^a from ordinal logistic regression models for number of adult MDD diagnoses in one’s thirties according to covariates.

	Model 1			Model 2
Covariate	OR	95% CI	p-value	OR	95% CI	p-value
Living with an adult misusing alcohol	1.63	1.05, 2.52	0.030	1.32	0.84, 2.07	0.233
Race
Caucasian (ref)	---	---	---	---	---	---
African American	1.84	1.14, 2.97	0.012	1.52	0.93, 2.50	0.092
Native American	1.84	0.86, 3.95	0.115	1.73	0.80, 3.74	0.164
Asian American	0.70	0.36, 1.39	0.311	0.64	0.32, 1.28	0.206
Parental income	0.92	0.77, 1.12	0.436	0.91	0.76, 1.10	0.345
Family history of depression	3.33	2.21, 5.01	<0.001	2.75	1.81, 4.17	<0.001
Parental education	0.80	0.63, 1.00	0.055	0.84	0.66, 1.06	0.138
Parental marriage status	0.89	0.55, 1.46	0.651	0.94	0.57, 1.55	0.807
Child maltreatment	---	---	---	1.04	1.02, 1.05	<0.001

Abbreviations: MDD, major depressive disorder; OR, odds ratio; CI, confidence interval

^aEstimates averaged across 20 imputed datasets

We next examined moderation of the association of living with adult alcohol misuse during childhood and MDD diagnosis chronicity by school bonding, neighborhood bonding, and sex. Odds ratios from separate multivariable ordinal logistic regression models for number of adult MDD diagnoses in one’s thirties, by sex, school bonding, and neighborhood bonding are shown in Table 4. None of the three exposure-x-moderator interaction terms were statistically significant (exposure-x-school bonding OR=0.59, 95% CI: 0.21, 1.31; exposure-x-neighborhood bonding OR=0.93, 95% CI: 0.42, 2.05; exposure-x-sex OR=0.92, 95% CI: 0.42, 2.03).

Table 4.

Odds ratios^a from separate ordinal logistic regression models for number of adult MDD diagnoses in one’s thirties, by sex, school bonding, and neighborhood bonding.

Covariate	ORb	95% CI	p-value
Effect Modification by Sex
Living with an adult misusing alcohol	1.66	0.97, 2.84	0.07
Sex	0.70	0.44, 1.13	0.15
Living with an adult misusing alcohol x sex	0.92	0.42, 2.03	0.84
Effect Modification by School Bonding
Living with an adult misusing alcohol	2.11	1.15, 3.87	0.02
School bonding	0.71	0.43, 1.19	0.19
Living with an adult misusing alcohol x school bonding	0.59	0.27, 1.31	0.20
Effect Modification by Neighborhood Bonding
Living with an adult misusing alcohol	1.66	0.89, 3.10	0.11
Neighborhood bonding	0.82	0.49, 1.37	0.45
Living with an adult misusing alcohol x neighborhood bonding	0.93	0.42, 2.05	0.86

Abbreviations: MDD, major depressive disorder; OR, odds ratio; CI, confidence interval

^aEstimates averaged across 20 imputed datasets

^bOrdinal logistic regression Model 1 adjustments

Discussion

In this prospective cohort study following individuals from grade 9 (~15 years old) to age 39, we found children who reported living with adult alcohol misuse during childhood were significantly more likely to meet diagnostic criteria for MDD during their thirties, and that this association was explained, at least partially, by child maltreatment. We did not find evidence of moderation by school bonding, neighborhood bonding, or sex.

Depression into Adulthood

Our findings linking childhood exposure to adult alcohol misuse and MDD in adulthood are consistent with previous epidemiological studies (Mwachofi et al., 2020; Reinherz et al., 2003). This study builds on earlier studies by suggesting the mental health consequences of living with adult alcohol misuse during childhood may extend further into adulthood than previously shown. If adults are impacted into their late 30s—when many are partnered, raising children, and becoming more established in careers—this has important implications for later life problems, family functioning, and perhaps intergenerational impacts (Catalano et al., 2021; Bramlett & Mosher, 2002; Hill et al., 2020). Our results are also consistent with the SDM framework. According to the SDM, alcohol misuse by a child’s parents or any other important adult in the child’s life provides opportunities for unhealthy bonding and conveys health-risking beliefs and values; this may also hinder formation of familial prosocial bonds that promote positive development. Healthy bonding opportunities in other settings such as schools and neighborhoods may be particularly important for children living with adult alcohol misuse in the family, where such opportunities may be limited.

Role of Child Maltreatment

Our finding that child maltreatment partially explained the association are consistent with previous research (Clemens et al., 2020; Clemens et al., 2019), and provides evidence that child maltreatment may be an important mechanism through which living with adult alcohol misuse during childhood influences long-term mental health status. Parental and caregiver alcohol misuse has been shown to be associated with child maltreatment (Dube et al., 2001), and affects parenting capacities, relationships, and attachments with children (Velleman & Templeton, 2016). This could confer risk on later-life depression as child maltreatment is a key determinant of emotional dysregulation across the lifespan (Thompson et al., 2014), and emotional dysregulation is in turn associated with psychiatric disorders such as depression (Bradley et al., 2011). In this study, participants retrospectively reported their experience of child maltreatment before age 18. Evidence suggests there may be sensitive developmental periods that could be particularly harmful to a child if they experience maltreatment. One study found that experiencing child maltreatment during middle childhood (ages 6–10) was particularly harmful and led to worse mental health outcomes compared to other developmental periods (Dunn et al., 2018). Thus, future research should consider the role of child maltreatment across different developmental periods and child maltreatment subtypes (e.g., emotional abuse, physical abuse, sexual abuse, emotional neglect, physical neglect) in the association between living with adult alcohol misuse during childhood and MDD in adulthood.

Role of Sex

We did not find evidence of moderation by sex, which may be consistent with prior literature suggesting that the relationship between substance abuse in the home during childhood and poor mental health in adulthood does not differ by sex (Reinherz et al., 2003; Lee & Chen, 2017). However, one study found that poor mental health among children that experienced adult alcohol misuse differed according to whether the mother or father misused alcohol (Balsa et al., 2009). Thus, there may be specific mother vs. father and daughter vs. son combinations that are particularly impactful. This study could not consider this particular question because of a lack of specificity in our measure of exposure (i.e., could not distinguish adult-child relationship), but this may be an area of future research.

Role of School and Neighborhood Bonding

Also, we did not find evidence of moderation by either school bonding or neighborhood bonding. This may be due to limited power to detect interactions in this study. Regarding school bonding, although the interaction did not reach statistical significance at p<0.05, stratified point estimates were consistent with our hypothesis. Children that reported having lived with an adult who misused alcohol and had low school bonding had greater chronicity of MDD in adulthood (OR=2.08, 95% CI: 1.11, 3.89) compared to children with high school bonding (OR=1.25, 95% CI: 0.69, 2.27). While these stratified results should be interpreted with caution, the direction of findings aligns with prior studies suggesting perceived positive school environments during childhood is associated with reduced risk of poor mental health in adulthood (Mason et al., 2010; Velleman & Templeton, 2016), and supports the SDM framework. To better specify prevention implications, future well-powered studies should seek to identify modifiable factors that could protect against deleterious effects of childhood exposure to adult alcohol misuse.

Limitations

There were limitations of this study. First, findings may not be generalizable given our sample included youth originally from higher risk neighborhoods in one urban region. However, prevalence of living with adult alcohol misuse during childhood and child maltreatment in our sample was similar to national studies (Grant, 2000; Hussey et al., 2006). Secondly, living with adult alcohol misuse during childhood was measured at one time point, which may not capture this exposure at other important developmental periods, particularly earlier in childhood, or cumulatively over time. Thirdly, it was not possible to determine the relationship of the adult who misused alcohol (e.g., a parent, sibling, grandparent), which would allow for further insight into the association and implications for interventions. Fourthly, we could not definitively establish temporality of adult alcohol misuse and child maltreatment. It is possible that child maltreatment precedes adult alcohol misuse. Yet, prior longitudinal studies suggest adult alcohol misuse leads to child maltreatment rather than vice versa (Bramlett & Mosher, 2002; Laslett et al., 2012). Lastly, assessment of child maltreatment relied on retrospective self-report which may introduce measurement error and threaten internal validity. However, studies have found retrospective reporting of child maltreatment to be valid (Hardt & Rutter, 2004).

Strengths

The study also had important strengths, including the longitudinal study design, follow-up of participants into the 30s, novel examination of mechanisms and protective factors, and investigation of possible moderators. These features provide unique contributions to a better understanding of the link between childhood exposure to adult alcohol misuse and poor mental health in adulthood.

Conclusions

Our study suggests living with adult alcohol misuse during childhood may be associated with greater chronicity of MDD in the 30s, with child maltreatment substantially explaining this association. These results highlight the importance of interventions that reach families with alcohol misuse in order to mitigate long-term mental health challenges related to this misuse or associated child maltreatment (Hildebrandt et al., 2020). Future research should seek to identify modifiable environmental factors, including a possible role for opportunities for positive bonding in schools and neighborhoods, that could potentially mitigate the negative impacts of the childhood exposure to adult alcohol misuse on mental health in adulthood.