Effects of Adverse Childhood Experiences and Chronic Health Conditions on Current Depression

This study investigated the ability of three adverse childhood experience (ACE) types (household dysfunction, emotional/physical abuse, and sexual abuse) to predict current depression among adults. We also determined the contribution of ACEs to current depression after controlling for covariates and chronic illnesses. Respondents to the 2010 Behavioral Risk Factor Surveillance System (n = 20,345) were divided into depressed and not depressed groups based on Patient Health Questionnaire-8 (PHQ-8) scores. Binary logistic regression determined the ability of ACE exposure to predict depression while controlling for effects of covariates and number of chronic illnesses. Hierarchical multiple linear regression determined the association of ACEs on depression scores after accounting for the covariate set and chronic illnesses. Sexual Abuse had the single strongest association with current depression of any ACE exposure. Exposure to three ACE types has the greatest association with current depression. ACEs accounted for about 7% of the variance in depression scores.

dysfunction that occur before the age of 18 years . ACEs in the form of physical abuse include being hit, beaten, or kicked. Emotional or psychological abuse includes being sworn at, insulted, or put down. Sexual abuse includes inappropriate touching and sexual activity. Household dysfunction includes living with a family or household member who was mentally ill or suicidal, substance abusing, ever incarcerated, living in a household with domestic violence, or with parents who divorced or separated after the child was born (Font & Maguire-Jack, 2016).
The negative psychological consequences of these childhood experiences manifest in childhood and adolescence (LeMoult et al., 2020) and have been associated with poorer mental, behavioral, and physical health outcomes in adulthood (Chapman et al., 2004;Dube et al., 2001;Fergusson, McLeod, & Horwood, 2013;Hillis et al., 2004). ACEs have been linked to a range of adverse physical health correlates such as obesity, heart disease, cancer and lung disease, diabetes, myocardial infarction, and premature death Brown et al., 1999;Campbell, Walker, & Egede, 2016;Felitti et al., 1998;Font & Maguire-Jack, 2016). Further, ACEs have also been linked to negative behavioral health correlates such as smoking and substance/alcohol abuse, suicide and high-risk sexual behavior, and anxiety, hallucinations, and sleep disturbances  Research on the long-term effects of childhood maltreatment has primarily focused on neglect and sexual or physical abuse. Researchers have continued to explore childhood maltreatment and an increasing number of studies have examined the impact of other forms of childhood abuse (Briere & Runtz, 1990;Brown, Cohen, Johnson, & Smailes, 1999;Liebschutz et al., 2002;Mullen, Martin, Anderson, Romans, & Herbison, 1996). These studies indicate that long-term effects of childhood maltreatment result not only from sexual or physical abuse, but also from other forms of maltreatment, such as emotional/psychological abuse and unstable, disordered family or household environments. Adverse childhood experiences (ACEs) are defined as occurrences of childhood maltreatment such as physical, emotional, or sexual abuse, as well as family or household Campbell et al., 2016;Felitti et al., 1998;Font & Maguire-Jack, 2016). In addition, several studies have reported a dose-response relationship between history of ACE exposure and negative mental health outcomes (Edwards et al., 2003;Felitti, 2002).
ACEs have specifically been associated with increasing the likelihood of experiencing depression in adulthood. Schilling et al. (2007) reported that ACEs were significantly related to depressive symptoms during the transition to adulthood. Women who reported experiencing household dysfunction, physical, verbal, and sexual abuse in childhood were more likely to report current depressive symptoms as adults (Remigio-Baker et al., 2014). Giano et al. (2021) investigated the relationship of ACE domains to personal history of depression diagnosis. Their findings indicated that sexual abuse and living with family members who had mental illness were independently and jointly associated with history of depression. Merrick et al. (2017) assessed the effect of ACE exposure on self-reported mental health outcomes and reported that greater exposure to ACEs was associated with adult depressed affect, substance abuse, and suicide attempts.
In 2009, the CDC included in the annual Behavioral Risk Factor Surveillance System (BRFSS) 11 items adapted from other instruments to measure exposure to adverse childhood experiences in the general population (ACE survey module, Centers for Disease Control and Prevention, 2009). , noting debate about how to properly sum ACE scale item scores for research purposes, assessed the factor structure of the 2009 Behavioral Risk Factor Surveillance System ACE Module (Centers for Disease Control and Prevention, 2009 ). Exploratory factor analysis of the 2009 BRFSS ACE Module data was followed with factor structure validation with confirmatory factor analysis using the 2010 BRFSS ACE Module data (Centers for Disease Control and Prevention, 2010). Results indicated that a 3-factor solution adequately fit the ACE Module data and represent the three construct areas of household dysfunction, emotional/ physical abuse, and sexual abuse .
Since the inclusion of the ACE survey module in the BRFSS, several research studies have explored the relationship of ACEs with the other health and mental health outcomes also assessed in the survey. For example, Ege, et al. (2015) found that ACE repeated exposure to any adverse childhood experiences, and a single instance or repeated instances of sexual abuse, were all significantly related to depression in geriatric BRFSS survey respondents. Remigio-Baker et al. (2014) conducted a study similar to ours, but focused on women in Hawaii. Our study brings the analysis of the relationship of ACE exposure and current depression to the U.S. general population. Relatively little is known about the distinct contribution of Ford and colleagues' factor-derived ACE subtypes to specific adult mental and physical health outcomes. The purpose of this study is to determine the associations of the ACE subtypes, separately and in combination, to current major depression. The primary hypothesis is that the different ACEs have additive effects and that multiple ACEs will have greater association with current depression than individuals who experienced few or no ACEs. A second hypothesis is that ACE total score is associated with current depression independent of a set of covariates and chronic health conditions.

Method
This study was approved as exempt from review by the Institutional Review Board of Auburn University Montgomery. In 2008, the CDC's National Center for Chronic Disease Prevention and Health Promotion developed questions similar to those used in prior studies of ACEs (Anda et al., 2010;Ford et al., 2014) for incorporation into the BRFSS survey. In 2009, the Centers for Disease Control and Prevention (CDC) began annually administering 11 items adapted from the ACE Study questionnaire (ACE survey module) to samples of adults participating in the BRFSS survey. ACE data were collected as part of the annually administered BRFSS questionnaire, a random digit dialing telephone survey. The BRFSS tracks the state-specific prevalence of behavioral health risks in all 50 states and US territories. BRFSS interviewers ask respondents a series of questions regarding their physical, mental, and behavioral health.

Participants
The data used in this study were from the 2010 Behavioral Risk Factor Surveillance System annual survey (Centers for Disease Control and Prevention, 2010). The study required data from survey respondents who had been administered the BRFSS ACE and the BRFSS Depression and Anxiety modules, both of which are optional. The portion of the survey data that met the criteria for our study were drawn from the states of Hawaii, Nevada, Vermont, and Wisconsin. The sample consisted of N = 20,345 adults. The total sample was divided into two groups based on scores obtained on the 8-item Patient Health Questionnaire (PHQ-8) (Kroenke et al., 2009), a measure of depression symptoms and described in the next section. The mean age of the total sample is 56.4 years (SD = 16.1 years, range = 18-99 years). Sociodemographic characteristics (sex, marital status, employment status, education level, chronic illness status, and ACE type) for each sample are summarized in Table 1.

Criterion Variable
The Patient Health Questionnaire-8 (Kroenke et al., 2009) consists of 8 items (Cronbach's α = 0.86) from the Patient Health Questionnaire-9 depression scale (Kroenke & Spitzer, 2002). The PHQ-8 questions form part of the optional Depression and Anxiety survey module that may be   Table 1 for distribution of ACE types).

Chronic Health Conditions
Chronic health conditions were measured by totaling the number of chronic illnesses reported by survey respondents including diabetes, asthma, heart attack, heart disease, and/ or stroke. The chronic health conditions variable was the total number of health conditions reported. There were relatively few respondents who reported more than 3 health conditions (n = 102 individuals reporting 4 or 5 chronic health conditions, or 0.33% of the weighted sample), so this variable was constructed to range from 0 (no chronic health conditions) to 3 (3 or more health conditions).

Sociodemographic Covariates
Sociodemographic variables were selected and included in the regression models in order to control for the effects of their known associations with depression. Model covariates were age, sex, employment status (7 categories), marital status (7 categories), race (6 categories), and educational level (6 categories). Research indicates that the first incidence of MDD tends to appear in emerging adulthood, suggesting that the likelihood of experiencing depression increases with age (Rohde et al., 2013). Women are more likely to be depressed than men (Parker & Brotchie, 2010) and unemployment is also associated with depression (Dooley et al., 2000;Kessler & Bromet, 2013). Divorced and separated individuals are more likely to experience depression than married individuals (Bulloch et al., 2009). In the United States, racial and ethnic minorities tend to have lower incidence of depression compared to whites (Dunlop et al., 2003;Ettman, Cohen, Abdalla, & Galea, 2020). Finally, lower levels of educational attainment are associated with depression (Peyrot et al., 2015).

Procedure
A binary logistic regression model was constructed with the criterion variable of PHQ-8 depression score (depressed-not depressed) and included the covariates, chronic illness status, and ACE type as predictors. The SAS Surveylogistic module (SAS Institute Inc., 2009) was used to conduct the logistic regression analysis. Survey design specifications for clustering and stratification were included in the analysis, items ask the number of days in the past 2 weeks in which respondents had experienced a particular depressive symptom (Kroenke et al., 2001). Individual item scores range from 0 to 14, which are summed to produce the PHQ-8 total score. PHQ-8 total scores range from 0 to 112 and, in this form, scores are known as the PHQ-8 days scores (Dhingra et al., 2011). PHQ-8 days scores of 55 or higher are consistent with Major Depression as defined in DSM-5 (American Psychiatric Association, 2013). The depressed group in this study was formed by selecting cases with PHQ-8 Days scores of 55 or higher, and the not depressed group consisted of individuals scoring below 55 on the PHQ-8. There were n = 918 scoring in the depressed range on the PHQ-8. This group has a mean age = 52.1 years (sd = 13.7 years, range = 18-94 years). There were n = 19,427 respondents scoring below the depressed range on the PHQ-8 with a mean age = 56.7 years (sd = 16.2 years, range = 18-99 years). PHQ-8 days scores have a sensitivity of 0.91 and a specificity of 0.92 for identifying major depression based on DSM-IV diagnostic criteria (Dhingra et al., 2011).

Adverse Childhood Experiences
The ACE module included in the 2010 BRFSS (Centers for Disease Control and Prevention, 2010) consists of 11 items that measure self-reported exposure to the following 9 types of childhood adversities: emotional abuse, physical abuse, sexual abuse, household member mental illness, household member substance abuse, witnessing domestic violence, parental separation or divorce, and incarcerated family members. Subscales used in this study were adapted from  and Cronbach's alpha coefficients for these subscales were calculated using the 2010 BRFSS ACE module response data. The Household Dysfunction subscale consists of 5 items (Cronbach's alpha = 0.60) to which participants responded yes (1) or no (0) to if they lived with anyone who is mentally ill, used alcohol or drugs, or spent time in a correctional facility, and parents' marital status. The Emotional/ Physical abuse subscale consists of 3 items (Cronbach's alpha = 0.70) with response options of never (0), once (1), or more than once (2) to questions about physical or verbal abuse by a parent or witness of abuse between parents. The Sexual Abuse subscale consists of 3 items (Cronbach's alpha = 0.84) with response options of never (0), once (1), or more than once (2) to if they had been touched sexually, forced to touch another sexually, or forced to have sex.
The ACE Type variable was created by grouping participants according to type of abuse reported, which yielded 7 categories: (1) No reported ACEs; (2) Household and the analysis was weighted using the appropriate weight variable for the 2010 BRFSS data. Accounting for these design features in the analysis improves precision of estimates and better approximates population parameters.
Hierarchical regression was performed as a second statistical analysis using SAS Surveyreg. In this analysis, the PHQ-8 total score, ranging from 0 to 112, was used as the criterion variable. The purpose of this analysis was to determine the proportion of variance in PHQ-8 total score explained by the sociodemographic covariates alone, the covariate set plus number of chronic illnesses, and finally, ACE total score after accounting for covariates and number of chronic illnesses. ACE total scores range from 0 to 17 (5 items scored 0-1, 6 items scored 0-2) with a coefficient alpha = 0.77. As before, statistical analysis accounted for sample stratification and clustering, and was weighted to better approximate population values.

Associations of ACEs to Current Depression
The binary logistic regression model was constructed with depression status as the outcome variable (depressed-not depressed). Covariates and chronic illness status were included in the model. Model fit was assessed with the likelihood ratio test and the model including all the variables was superior to the null model (F (31, ∞) = 15,929, p < .0001). This indicates that model fit is improved by inclusion of the predictor variables compared to an intercept-only model. Table 2 presents the Type 3 Analysis of Effects, in which each individual effect is adjusted for the presence of all other variables in the model. Every variable in the model is significantly related to depression status except age, sex, and race. Table 2 shows that depression status is significantly related to type of reported adverse childhood experiences and chronic health conditions while controlling for the effects of the other variables in the model. Table 3 presents adjusted odds ratios and 95% confidence intervals for adverse childhood experiences, chronic   Ford and colleagues (2014). The results of this study contribute to understanding and validity of the factor-derived subscales by demonstrating the association of each ACE subscale with current depression. We hypothesized that additive effects of different types of ACEs and that reporting a greater number of different ACEs would be associated with greater association with current depression.
The results of the logistic regression partially supported the hypothesis that the additive effects of multiple ACEs would be greater than the effects of experiencing any single ACE subtype. All adverse childhood experiences, alone or in combination, are associated with an at least doubled risk of current depression. However, sexual abuse, experienced either alone or with household dysfunction or emotional/ physical abuse, is a relatively stronger risk factor for current depression than either household dysfunction or emotional/ physical abuse experienced alone. This is consistent with and further supports findings that, among geriatric respondents, a single incident of forced sexual contact in childhood was associated with an almost three-fold increase in risk of depression (Ege et al., 2015). The combination of household dysfunction and emotional/physical abuse is also as strongly associated with current depression as sexual abuse alone. Consistent with our hypothesis, individuals who experience all three forms of adverse childhood experiences including household dysfunction, emotional/physical abuse, and sexual abuse have an almost 9 times greater risk of current depression than individuals reporting no ACEs.
The hierarchical regression analysis indicated that adverse childhood experiences independently contributes to current depression over and above the associations of depression with the covariate set and chronic illness. Number of chronic illnesses accounts for a 2% increase in variance explained in PHQ-8 depression scores. Adding ACE type to the final regression model accounted for an additional 7% of variance explained in depression scores.
Overall, our findings are consistent with previous research indicating that adverse childhood experiences have a cumulative and negative impact on adult mental health (Schilling et al., 2007;Schilling, Aseltine, & Gore, 2008). The findings of this study provide additional evidence and support for the relationship between ACEs and depression. It is also clear from previous research that experiencing ACEs can lead to physical health problems and other mental/behavioral difficulties (Dube et al., 2001Felitti et al., 1998;Fergusson et al., 2013). Therefore, it seems plausible that the negative health outcomes that result from ACEs reinforce or exacerbate the development of depression either during adulthood or continuing into adulthood.
It is important to note that the outcome variable in this study is the experience of depressive symptoms in the two illnesses, and the covariates. Any number of chronic illnesses is associated with an increased association with depression compared to the reference group of survey respondents not reporting any chronic illnesses. Likewise, all adverse childhood experiences occurring singly or in combination are associated with an increased association with current depression compared to the group of respondents reporting no history of these experiences. Sexual abuse is the single ACE with the greatest association with depression (AOR = 4.8, 95% CI = 2.1-10.8). Household dysfunction and emotional/ physical abuse are associated with a more than double the risk of current depression. Finally, the combination of all three ACEs -household dysfunction, emotional/physical abuse, and sexual abuse -has the greatest association of current depression with an almost nine-fold increase in risk (AOR = 8.8, 95% CI = 5.6-13.8).

Contribution of ACEs to Depression Scores
Three sequential regression models were used to determine the independent contributions of chronic illnesses and ACE total score to variance in PHQ-8 depression scores over and above the variance explained by the set of covariates. All regression models included cluster and stratification design effects and were weighted to better estimate population parameters. In Model 1, all covariates were entered as a block and included age, sex, employment status, educational attainment, and marital status. Each covariate except race was significantly related to PHQ-8 depression scores (R 2 = 0.127, F 21, 20341 = 22.4, p < .0001). In Model 2, number of chronic diseases was added to the covariate set from Model 1. Again, each variable was significantly related to PHQ-8 scores except race (R 2 = 0.146, F 24, 20341 = 24.9, p < .0001, ∆R 2 = 0.019). Model 3 included covariates, number of chronic illness, and added ACE total score. All variables included in Model 3 were significantly related to PHQ-8 depression total score (R 2 = 0.218, F 25, 20341 = 33.2, p < .0001, ∆R 2 = 0.072). Number of chronic illnesses accounts for about 2% of the variance in PHQ-8 depression score after controlling for the covariate set. ACE total score accounts for about 7% of this variance after controlling for the covariate set and number of chronic illnesses.

Discussion
In this study we investigated the individual and combined effects of different types of self-reported adverse childhood experiences on current depression. Types of adverse childhood experiences investigated were household dysfunction, emotional/physical abuse, and sexual abuse, which were measured using the factor-derived subscales of the BRFSS neglect and exposure to interpersonal violence. The ACE module embedded in the BRFSS survey instrument consisted of only 11 items in 2010. These items were adapted from earlier survey instruments and do not allow quantification of ACE exposure on dimensions of frequency, severity, and duration. The latest version of the BRFSS ACE Module does increase breadth of coverage of ACE events with the inclusion of two items focused on emotional and physical neglect (Centers for Disease Control and Prevention, 2021). Researchers would welcome continued refinement of the BRFSS ACE module.
The PHQ-8 scale is widely used in depression research because of its brevity, its correspondence with DSM diagnostic criteria, and its reliability and validity in assessing depression. However, the PHQ-8 scale does not allow identification of type and duration of depression symptoms. The PHQ-8 scale in the present study, and similar studies, is best viewed as providing an estimate of current level of depression symptoms in the population or among subgroups. This application cannot provide precise diagnoses of survey respondents who get elevated scores (e.g., bipolar disorder, major depressive episode). Although we set the cut score for forming our depression group to be consistent with major depressive disorder, this level of depression may be cooccurring with a variety of clinical conditions. The results bear on the experience of current depressive symptoms, but not on any particular diagnostic category.
This study affirms the association between all types of adverse childhood experiences and the experience of negative mental health outcomes, namely depressive symptomatology, in a sample representative of the U.S. adult general population. The negative psychological consequences may emerge contemporaneously with the abuse in childhood, or problems may emerge later in adolescence or adulthood. This study design and nature of the survey data do not permit the claim that adverse childhood experiences directly cause increased depressive symptomatology in adulthood. Nevertheless, the association between these experiences in childhood and later adult behavioral pathology is strong, consistent, and a dose-response relationship exists between number of ACEs experienced and later health and mental health problems (Centers for Disease Control and Prevention, 2019). Adverse childhood experiences are also plausible as remote causes of later disorders of emotional regulation. (Crow et al., 2021;Dvir, Ford, Hill, & Frazier, 2014;Michopoulos et al., 2015) Schilling et al. (2008 have offered compelling evidence that poorer mental health outcomes are associated with both number and severity of adverse childhood experiences, rather than the simple sum of different types of experiences that make up a particular exposure. The authors suggest that prevention focusing on the most severe forms of traumatic weeks prior to the telephone survey, rather than history of depression. In light of this distinction, the relationship demonstrated between ACE exposure and the PHQ-8 depression score is both interesting and revealing. The findings indicate that the individual with a history of ACE exposure, contacted at random, is more likely to be depressed at the time of that contact than another person who does not have the ACE exposure history. Knowing that ACE exposure is associated with past diagnosis of depression is important and useful (e.g., Giano et al., 2021). However, history of depression diagnosis as an outcome variable permits the interpretation that depression has resolved and is not a current problem. Because depression can and does resolve, current depression presents a rather narrow target as an outcome variable. Thus, the findings of our study may suggest that people who experience ACEs are more likely, as compared to those without ACE exposure history, to be struggling with depression for extended periods in adulthood.
The primary limitation of this study is that determination of ACE exposure is based on recall during a telephone survey and cannot be verified. Hardt and Rutter (2004) offered a thorough review of the research and problems that may affect the validity of adult recall of traumatic events experienced in childhood. The authors distinguish between serious adverse experiences with clear and understandable case definitions (e.g., severe physical abuse or sexual trauma) and more subtle and less serious negative aspects of family life. Recall of both types of adverse experiences are problematic from a validity standpoint, but the more serious variety of adverse event is relatively less subject to validity concerns. The authors point out that, among adults with verified histories of significant abuse in childhood, about a one-third do not report childhood abuse when asked about it. In a metaanalysis of studies including both prospective and retrospective measures of child maltreatment, it was found that 52% of individuals with prospectively established childhood abuse did not report it retrospectively. In addition, 56% of participants who retrospectively reported abuse did not have corresponding prospective records of abuse (Baldwin et al., 2019). The authors note that the low agreement between prospective and retrospective measures does not mean that retrospective recall of child maltreatment is invalid. If prospective measures of child maltreatment lack sensitivity and fail to identify a significant proportion of cases, then greater case prevalence using retrospective recall could reflect relatively greater sensitivity of retrospective methods.
Gardner, Thomas, and Erskine (2019) published a metaanalysis of data from population-representative studies of the association of ACE exposure and later adult depression. In their discussion, the authors noted that relatively few studies meeting their inclusion criteria for emotional abuse were available. The same was true of studies investigating exposure would yield the greatest benefit to adult mental health. Many adverse childhood experiences, often occurring within a complex network of family life, social, and economic disadvantage, may go undetected and unreported, thus allowing for their continuation over long periods during childhood. The early detection, reporting, and prevention of adverse childhood experiences are important goals in the efforts to improve mental health among children and adults. The Centers for Disease Control have offered a package of programs that support families, discourage violence, and teach skills that help prevent ACEs. (Centers for Disease Control and Prevention, 2019) Certainly, the prevention of ACEs holds out the possibility of avoiding the long-term negative physical and mental health effects. This study, and the abundance of studies yielding similar results, point to the critical importance of advancing early detection and prevention of adverse childhood experiences, and efforts to ameliorate their damaging effects, for the improvement of public health and mental health.