Adverse Childhood Experiences and Current Health Status in a Community Sample of Runaway and Homeless Youth

Abstract

In recent years, researchers and policymakers have called attention to the importance of child and adolescent trauma for understanding adult health status. The primary aim of this study is to describe the adverse childhood events reported in a sample of runaway and homeless youths and examine their impact on these youths’ current health status. We utilize survey data collected from a community sample of runaway and homeless youths gathered in metro Atlanta. Using the Adverse Childhood Experiences Scale (ACEs scale), we examined the relationship between ACEs and several health status measures using OLS and logistic regression. We found that runaway and homeless youths endorsed experiencing many ACEs, especially sexual minority youths, youths who had prior involvement with child-serving social service systems, and youths who were homeless for more than a year. Black/African American youths were slightly less likely to report many adverse childhood experiences. Runaway and homeless youths who reported more ACEs had increased odds of experiencing significant current mental health and/or substance abuse problems. Our study suggests ACEs are an important factor shaping these youths’ health and underscores the potential value of trauma-informed care for youths experiencing homelessness.

1. Introduction

Adverse childhood experiences (ACEs) have significant effects on adult health in the general population in the United States [1,2,3]. Indeed, ACEs have been argued to be a significant factor in understanding health disparities that persist among many vulnerable populations, including proximal and longer-term adult health outcomes as well as experiencing violence and victimization [4,5]. While studies of people experiencing homelessness have documented significant childhood trauma, researchers have yet to estimate the prevalence of ACEs or their empirical relationship with health status in this very vulnerable population. In this paper, we summarize findings regarding the prevalence of ACEs from a large survey of runaway homeless youths and examine the relationship between reported adverse childhood experiences and current health status in this vulnerable group of youths.

Since the publication of the original Kaiser ACEs study [6], public health research interest in and concern regarding the prevalence and implications of ACEs has grown exponentially. The United States Centers for Disease Control and Prevention (CDC) incorporated a version of the Kaiser scale into an ACEs Module used in the Behavioral Risk Factor Surveillance Systems (BRFSSs) to encourage more research in this area [7]. In response, many states have incorporated the ACEs module into their BRFSS surveys [8], and many independent researchers have done so as well, using variations of the original Kaiser scale. As a result, many studies have been published demonstrating a strong association between ACEs and poorer subjective health and mental health, higher rates of STIs and HIV infection, as well as the prevalence of some longer-term physical and chronic health conditions [5,9]. Studies further suggest that several vulnerable populations experience ACEs more frequently, prompting researchers to ask whether these childhood experiences may, at least partially, explain greater health disparities in physical and mental health, substance use, and other problems across other socially disadvantaged populations. For example, Austin et al. [10] and others [11] find that higher ACE scores among lesbian, gay, and bisexual adults explain much of the elevated rates of health disparities across a range of physical and psychological health conditions. Further, Edaliti et al. [12] found that higher ACE scores are associated with a greater odds of criminal justice involvement and victimization among homeless adults with mental illness. As interest in the scale has grown, researchers also have begun to think more systematically about the psychometric properties of the ACEs as well as some of their limitations that include retrospective reporting and the assumption that each adverse event experienced is equally important to an individual [13]. Nevertheless, the ACEs scale provides a powerful tool for improving our understanding of key social determinants of health in the general population and specific vulnerable sub-populations, such as people experiencing homelessness generally and runaway and homeless youths in particular.

At the same time, there is parallel literature on the high rates and health implications of compounding and specific forms of trauma among many vulnerable populations, especially women, racial and ethnic groups, sexual and gender minorities, and people experiencing homelessness. For example, Black and Latinx youths report exposure to multiple ACEs more often than White youths [14]. Across racial and ethnic groups, youths living in poverty endorse more ACEs than youths not living in poverty [14], leading several researchers to posit that contextual and systemic-level factors are important to consider in the profound inequity in the distribution of ACEs experienced among minoritized groups [15,16,17]. Additionally, research has shown that youths who experience certain types of traumas, such as childhood sexual abuse, in comparison to youths who do not, are up to twice as likely to be exposed to other ACEs such as parental substance abuse, physical neglect, domestic violence, and emotional/physical abuse [18]. Briggs et al. [19] further highlight this important research and argue that as the total number of ACEs increases for minoritized populations, there is a possible “synergy” that occurs, whereby specific types of ACEs experienced within aversive environments lead to disproportionate risk for negative health outcomes. As research continues to expand the original ACEs framework to include contextual and environmental factors such as poverty and homelessness [20,21], examining the experiences of diverse runaway and homeless youths will provide a vital lens on how the toll of homelessness during childhood is potentially exacerbated by concurrent ACEs that perpetuate health inequalities.

In the case of runaway and homeless youths, childhood trauma and family problems have been documented as important factors in youths running away and/or becoming homeless [21,22,23,24,25]. Several studies have found that runaway and homeless youths have experienced significant traumas throughout their life and face significant health problems [22,26,27]. Yet, runaway and homeless youths’ histories of trauma have largely been documented through qualitative data. For example, qualitative research has documented numerous challenges and suggests that lesbian, gay, bisexual, transgender, and questioning runaway and homeless youths have significant histories of family conflict [21,24,27,28]. While trauma and difficulties with abuse and neglect are consistently important factors in youths becoming homeless and experiencing health problems (i.e., greater HIV risk, mental health problems), research in this population is largely anecdotal, focused on specific childhood events as single or a discrete set of trauma variables, or examined clinically from a post-traumatic stress disorder perspective [24,29,30,31].

In this paper, we summarize data from a large systematic community survey of runaway and homeless youths, in which youths were asked to complete the ACEs scale. In addition to summarizing past ACEs, we examine the relationship between ACEs and current health status, specifically focusing on the odds that runaway and homeless youths experience serious mental illness and/or significant substance use problems as well as their subjective health status and the number of self-reported major health conditions.

2. Materials and Methods

2.1. Sample

The data for this study come from the 2018 Atlanta Youth Count, a community-based survey of youths experiencing homelessness and housing insecurity. More details and the study documentation are available below in the data availability statement.

2.1.1. Ethical and Human Subject Considerations

The study was designed and conducted to estimate the number and needs of homeless youths between the ages of 14 and 24 in the metro Atlanta area to improve public awareness and services for this exceptionally vulnerable population. The research team considered a wide range of ethical issues, federal and state law, and international scientific standards for the conduct of research involving human subjects in designing the project. Because the investigators were aware that many, if not most, homeless and runaway youths have left home because they are estranged from their family and/or have troubled family situations, the researchers designed the study to ensure all participants’ confidentiality and to provide some tangible benefit for participation.

The United States’ Code of Federal Relations, Part 46, Subpart D, §46.408(c), specifically permits the involvement of children (i.e., youths less than 18 years of age) and further allows for a waiver of the normal requirement of parental or guardian consent when seeking such consent could pose significant risk to the subjects. Studies of homeless youths have suggested applying the concept of “mature minors” to adolescents older than 14 who are unable to consent to research that poses minimal risk [32]. Meade and Slesnick [33] further submit that homeless youths, who are no longer living with their caregivers whether by force or voluntarily, should have access to the resources of research (i.e., mental health referrals, monetary incentives, or food vouchers) that they may otherwise not obtain, as was the case in this study.

Consequently, the study protocol was designed to openly and clearly inform all participants of the nature of the research and guarantee their confidentiality throughout the entire research process. The study protocol and field procedures were reviewed and approved by the authors’ Institutional Review Board (see below).

2.1.2. Field Work and Data Collection

For approximately six months prior to the initiation of survey data collection, senior members of the research team conducted extensive field work at multiple homeless youth-serving agencies and in areas where homeless youths were believed to live or congregate. The aim of this phase of the project was to better understand the social patterns and movement of runaway and homeless youths in the area and construct detailed maps of the focal counties encompassing metro Atlanta (i.e., Fulton, DeKalb, Cobb, Clayton, and Gwinnett) to guide our sampling and data collection.

In accordance with capture–recapture methods [34,35,36,37], we developed a formal one-month survey field plan that systematically identified timeframes and locations for sampling to maximize opportunities for contacting runaway and homeless youth. While like traditional convenience and/or snowball sample methods for surveying invisible or hard-to-reach populations, the systematic selection of areas and timeframes and allowing individuals to be “recaptured” (i.e., surveyed more than once) at different locations/timeframes are central and critical methodological tools designed to yield better estimates of hidden populations in a defined geographic area [38,39,40]. To accomplish this task, the research team constructed a confidential unique identifier concatenated from each youth’s responses to seven non-varying, non-identifiable questions. This procedure allowed the researchers to maintain the youths’ anonymity and reasonably estimate if a given survey represented a youth who may have participated more than once (i.e., “recaptured”). Upon completion of data collection, we reviewed all cases with duplicate identifiers and eliminated duplicate surveys so that only one survey was included in the final sample for each unique youth (i.e., unique identifier), regardless of the date, time, or location we encountered them during the month-long survey period. The final sample included 564 unduplicated youths.

During the month of survey data collection, small research teams composed of two or more student data collectors and one or two professional homeless service providers canvassed the sampling sites at the selected times. To maximize rapport with runaway and homeless youths, the field staff consisted primarily of university students who were similar in age to our target study population and thoroughly trained in the study protocol and the ethics of collecting data from human subjects. In the field, the field staff approached youths and asked them if they would be willing to answer a short screening questionnaire to determine if they were eligible for a research study. To be eligible for inclusion, youths had to indicate that they (1) were between the ages of 14 and 25; (2) were living on their own without significant family support; and (3) did not have a stable residence of their own for at least one day in the prior 30 days.

All eligible youths were then invited to participate. Prior to administering the survey, the data collection team member provided an overview of the study, reviewed the informed consent statement, and answered questions the youth had about the research. All willing participants were offered the opportunity to complete a paper copy of the survey on their own or have the data collection team member read the survey and record their responses for them. Upon completion of the survey, participants were provided with a USD 10 gift card and a booklet listing available support services in the metro area. Non-eligible youths and those who were not interested in participating were also provided with the same service provider listing. All youths encountered in the field were encouraged to talk with service providers and offered other resources the teams carried with them (e.g., granola bars, hygiene supplies, blankets).

2.2. Measures

The survey included questions regarding the youths’ demographic characteristics, history of homelessness, physical and mental health status, and ACEs. The specific measures utilized in this study are described in the following sections.

2.2.1. Adverse Childhood Experiences (ACEs) Study Questionnaire

The respondents were presented with the 11-item ACE module, as developed and used by the United States Centers for Disease Control and Prevention [2,6,38]. Consistent with prior research [2], we created a series of binary (yes = 1/no = 0) variables and combined them to create 8 categories of childhood experiences most reported in ACE studies. We also constructed a total ACE count that measures the cumulative exposure to adverse childhood experiences (range 0–8) by summing these binary variables; only respondents who provided valid data on all 8 categories were included, yielding a study sample of 451.

2.2.2. Health Measures

The survey also included four measures of current health status. First, we asked each respondent about their general subjective health status by asking the question: “Would you say that in general, your health is (1) “poor”, (2) “fair”, (3) “good”, (4) “very good”, or (5) “excellent”. We constructed a dummy variable to focus on individuals who indicated that their current health status was “poor” or “fair” and coded these cases as 1. Those who indicated they were in good, very good, or excellent health were coded as 0. After posing this question, respondents were asked to indicate if they currently had a series of 7 major or chronic health problems, such as asthma, diabetes, dental or skin conditions, and hepatitis, and whether they had been diagnosed with HIV, or any other STI other than HIV. To facilitate analysis, after counting the number of chronic health conditions reported (range 0–7, mean = 0.667, SD = 0.945), we created a dummy variable to identify those youths with any current self-reported chronic health condition (1, n = 191, 42.4%) versus those that reported no problems (0, n = 260, 57.6%).

Because mental health and substance abuse are significant concerns in this population, the survey instrument included reliable and valid screening tools to assess the probable presence of clinical problems in these areas. To assess mental health issues, we used the Kessler 6 scale [39]. This six-item screening tool asks respondents to indicate how frequently over the past 30 days they felt “nervous”, “hopeless”, “restless or fidgety”, “so depressed that nothing could cheer you up”, “everything was an effort”, and “worthless”. The response categories ranged from “none of the time” (0) to “all of the time” (5). Following recommendations from Kessler and colleagues [41], we dichotomized responses to the six questions before summing their responses (range 0–24, alpha reliability = 0.841). We then created a binary variable to identify respondents who had scores of 13 or more (1 = scores of 13 or more; 0 = scores of 12 or less), the recommended cut point indicative of a probable severe mental illness. Similarly, we used the widely adopted nine-item CRAFFT scale [42,43], specifically designed for adolescents and older youths. The CRAFFT scale asks youths whether, in the past year, they used alcohol or other substances and, more importantly, whether or not they engaged in any risky substance use-related behaviors recently, including driving a Car themselves or with someone else under the influence, using substances to Relax or Alone, Forgetting things they did while using, whether Family or Friends told them to cut down on their use, and whether they ever got into Trouble while using alcohol or drugs. Based on the youth’s responses to the six yes/no risk items (alpha reliability = 0.680), youths who reported engaging in 2 or more risk behaviors were coded as having a probable substance abuse problem, based on extensive clinical research using this tool [42,43].

2.2.3. Demographic Variables

To describe the sample and better understand the relationship between ACEs and the health status of runaway and homeless youths, we used information from several demographic questions for this analysis, including age; race/ethnicity; education (>HS); whether it was the first time they were homeless (Yes vs. No); and how long they had been homeless at the time of the interview (less than 1 month or 30 days; 1–2 months; 3–6 months; 6 months to 1 year; 1 year or more). We also asked each youth to describe their current gender and sexual identity. Because of the small number of individuals who identified as a gender minority (n = 33, 6.5%), we created a binary variable that combined youths who identified as a sexual (lesbian, gay, bisexual) and/or gender (transgender, gender non-binary, genderfluid) minority and operationally defined them as a sexual or gender minority for use in our multivariate analyses (n = 155, 29.8%).

2.3. Analyses

The data were analyzed using IBM SPSS (version 28.0). In this paper, we focus on the runaway and homeless youths who completed the ACEs (N = 451). We summarize the youths’ responses to the individual ACE items and analyze them using the eight ACE categories as well as the total number of categories endorsed. To better understand the number of ACEs, we use OLS regression to identify which groups of runaway and homeless youths reported higher ACE scores. We applied logistic regression methods to estimate the odds that different sub-groups of runaway and homeless youths report poor or fair health, one or more chronic health conditions, a probable severe mental illness, or a substance abuse problem. In addition to including the youth’s age (in years), we included a standard series of binary variables in all of the multivariate models to examine the effects of background factors, including race (Black vs. all other races), gender (cisgender woman vs. cisgender man), being a sexual or gender minority (vs. cisgender straight identified), education (having a high school diploma or more education vs. less education), and whether the youth was involved with a child-serving system such as child welfare, foster care, or juvenile justice (vs. no such contact). Lastly, we included two variables to model different experiences of being homeless, i.e., if it was the first time a youth was homeless (vs. if they were homeless more than once before the current episode) and whether the youth had been homeless for more than one year (vs. less than one year).

3. Results

3.1. Sample Characteristics

Table 1. Demographic profile of the analytic sub-sample of runaway and homeless youth, 2018 Atlanta Youth Count (N = 451).

	N	%
Age
17 Or Younger 18–19 Years Old 20–21 Years Old 22–23 Years Old 24–25 Years Old	6 103 134 88 120	1.3 22.8 29.7 19.6 26.6
Race/Ethnicity
Black; African American Multiracial White; Caucasian Other Race; Ethnicity	244 144 36 18	55.2 32.6 8.1 4.1
Gender Identity
Cisgender Man Cisgender Woman Transgender Or Other Gender Identity	265 132 27	62.5 31.1 6.4
Sexual Identity
Straight (Heterosexual) Lesbian, Gay, Or Bisexual (LGB)	325 116	73.7 26.3
Education (Hs Diploma Or Better)	311	69.0
System Involvement As Child (Before Age 18)	289	64.1
First Time Homeless (Yes)	172	39.1
Time Homeless (Current Episode)
Less Than 1 Month/30 Days 1–2 Months (30–89 Days) 3–6 Months (90–180 Days) 6 Months-1 Year (181–364 Days) 1 Year Or More (More Than 365 Days)	80 79 97 68 110	18.4 18.2 22.4 15.7 25.3

presents the demographic characteristics of the analytic sample of runaway and homeless youths (N = 451). The youths in the sample were on average 21.43 years old (SD = 2.35). The majority indicated that they were Black or African American (55.2%) or Multiracial (32.6%) and cisgender men (62.5%). Regarding sexual and gender minority status, 26.3% reported identifying as lesbian, gay, or bisexual, and 6.4% of individuals self-identified as transgender or a gender identity other than a cisgender identity; when combined, 28.3% identified as a sexual or gender minority.

3.2. Adverse Childhood Experiences (ACEs)

Table 2. Responses to the Adverse Childhood Experiences Scale (ACEs) reported by runaway and homeless youth respondents, 2018 Atlanta Youth Count (N = 451).

Category	Item	No %	Yes %
	The next set of questions are going to ask you about some things that may have happened to you during your childhood and your life. Before you were 18…
Household Dysfunction	Did you live with anyone who was depressed, mentally ill or suicidal?	61.9	38.1
	Did you live with anyone who was a problem drinker or alcoholic?	63.6	36.4
	Did you live with anyone who used illegal street drugs or who abused prescription medication?	68.3	31.7
	Did you live with anyone who served time or was sentenced to serve time in a prison, jail or other correctional facility?	58.8	41.2
		No %	Yes %	Parents Never Married %
	Were your parents separated or divorced? *	27.1	39.7	33.3
		Never %	Once %	More than once %
Physical or Emotional Abuse	Before you were 18, how often did PARENTS in your home ever…
	Slap, hit, kick, punch or beat each other up?	61.0	15.3	23.7
	Hit, beat, kick or physically hurt you in any way? (do not include spanking)	52.8	14.6	32.0
	Swear at you, insult you or put you down?	37.7	12.2	50.1
Sexual Abuse	Before you were 18, how often did anyone at least 5 YEARS OLDER THAN YOU or AN ADULT…
	Touch you sexually?	69.1	9.8	21.1
	Try to make you touch them sexually?	70.5	12.0	17.5
	Force you to have sex?	73.6	9.8	16.7

* In pilot testing of the ACES with runaway and homeless youths, respondents frequently expressed confusion regarding the original wording of this item; in response, we added the response category “parents never married” as an option. To maintain consistency with the standard ACES coding, we treated “parents never married” as a “no”.

presents the responses to the individual items of the ACEs scale in our sample. The most frequently endorsed individual ACE items included someone in the household having been incarcerated (41.2%); parental divorce (39.7%); and a childhood household member being depressed, mentally ill, or suicidal (38.1%). When we collapsed the individual items into the eight ACE categories commonly examined in the literature [5], emotional abuse (62.3%) was the most common, followed by physical abuse (47.2%) and household substance abuse (46.1%; see

Table 3. Responses to the Adverse Childhood Experiences Scale (ACEs) categories reported by runaway and homeless youth respondents in the 2018 Atlanta Youth Count (N = 451).

ACE Categories Mentioned	n	%
Physical Abuse	213	47.2
Sexual Abuse	159	35.3
Emotional Abuse	281	62.3
Parental Separation or Divorce	179	39.7
Adult Mental Illness in Household	172	38.1
Adult Substance Abuse in Household	208	46.1
Incarcerated Household Member	186	41.2
Violence Between Adults in Household	176	39.0
Number of ACE Categories Mentioned	n	%
0	58	12.9
1	51	11.3
2	48	10.6
3	70	15.5
4	73	16.2
5	60	13.3
6	36	8.0
7	31	6.9
8	24	5.3
	M (SD)
	3.49 (2.31)

).

Table 4. Regression analysis of the cumulative number of adverse childhood experiences (ACEs) reported by youths experiencing homelessness, 2018 Atlanta Youth Count (n = 451).

	b
Age (in Years)	−0.030
Race Group (Black)	−0.741 ***
Gender (Cisgender Woman)	0.245
Sexual/Gender Minority (LGBT)	0.756 **
Education Level (High School Diploma or More)	−0.266
Prior System Involvement (Yes)	1.083 ***
First Time Homeless (Yes)	−0.377
Homeless for More Than 1 Year (Yes)	0.571 *
Model F Statistic	9.418 ***
Standard Error of the Equation (SEE)	2.140
R2	0.159

* p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001.

presents an OLS regression of the number of ACEs based on demographic characteristics among the runaway and homeless youths. This model suggests that Black or African American runaway and homeless youths (b = −0.741) were significantly less likely to report more ACEs. In contrast, runaway and homeless youths who were a sexual or gender minority (b = 0.756), as well as those with prior involvement with child-serving systems (b = 1.083) and who had been homeless for more than a year (b = 0.571), were significantly more likely to report higher ACE scores.

In

Table 5. Logistic regression analyses of individual adverse childhood experience (ACE) categories among youths experiencing homelessness, 2018 Atlanta Youth Count (n = 451).

	Physical Abuse	Sexual Abuse	Emotional Abuse	Parental Separation or Divorce	Adult Mental Illness in Household	Adult Substance Abuse in Household	Incarcerated Household Member	Violence Between Adults in Household
	O.R. (95% CI)	O.R. (95% CI)	O.R. (95% CI)	O.R. (95% CI)	O.R. (95% CI)	O.R. (95% CI)	O.R. (95% CI)	O.R. (95% CI)
Age (in Years)	0.997 (0.907– 1.097)	0.895 * (0.805– 0.996)	1.013 (0.921– 1.115)	1.018 (0.927– 1.118)	0.920 (0.835– 1.015)	1.049 (0.956– 1.150)	0.955 (0.870– 1.049)	0.994 (0.904– 1.094)
Race Group (Black)	0.589 * (0.386– 0.899)	0.669 (0.418– 1.071)	0.606 * (0.395– 0.929)	0.653 * (0.433– 0.986)	0.485 *** (0.315– 0.745)	0.620 * (0.413– 0.931)	0.710 (0.470– 1.072)	1.056 (0.691– 1.613)
Gender (Cisgender Woman)	0.894 (0.563– 1.419)	3.508 *** (2.146– 5.737)	1.155 (0.726– 1.837)	0.548 * (0.345– 0.870)	1.581 (0.996– 2.509)	1.060 (0.681– 1.651)	1.057 (0.675– 1.656)	0.983 (0.622– 1.554)
Sexual/ Gender Minority (LGBT)	1.521 (0.953– 2.427)	3.887 *** (2.374– 6.365)	2.277 *** (1.380– 3.758)	0.809 (0.511– 1.281)	1.425 (0.895– 2.270)	1.205 (0.770– 1.884)	1.310 (0.834– 2.058)	1.399 (0.886– 2.208)
Education Level (HS or better)	0.742 (0.465– 1.183)	0.984 (0.590– 1.641)	1.103 (0.686– 1.772)	0.955 (0.605– 1.508)	0.951 (0.596– 1.518)	0.775 (0.494– 1.217)	0.729 (0.464– 1.145)	0.786 (0.497– 1.243)
Prior System Involvement (Yes)	2.760 *** (1.747– 4.359)	2.447 *** (1.445– 4.146)	1.498 (0.963– 2.329)	1.170 (0.754– 1.816)	2.282 *** (1.416– 3.677)	1.614 * (1.045– 2.492)	1.659 * (1.063– 2.588)	2.226 *** (1.396– 3.549)
First Time Homeless (Yes)	0.584 * (0.381– 0.896)	1.146 (0.713– 1.842)	0.669 (0.439– 1.021)	1.206 (0.797– 1.827)	0.749 (0.483– 1.160)	0.915 (0.607– 1.379)	0.759 (0.500– 1.153)	0.672 (0.438– 1.032)
Homeless for More Than 1 Year (Yes)	1.521 (0.931– 2.486)	1.710 * (1.005– 2.909)	1.094 (0.664– 1.805)	1.332 (0.826– 2.148)	1.456 (0.889– 2.385)	1.242 (0.774– 1.993)	1.577 (0.978– 2.541)	1.254 (0.773– 2.033)
−2 Log Likelihood	512.310	439.583	513.598	535.174	500.256	545.631	533.740	518.351
Overall X2	52.626 ***	88.591 ***	28.196 ***	17.237 *	43.513 ***	19.011 *	23.512 **	24.440 ***
Nagelkerke R2	0.161	0.269	0.091	0.056	0.137	0.061	0.075	0.079

* p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001.

, we present the corresponding logistic regression analyses of the individual ACE categories with the adjusted odds ratios (and 95% Confidence Interval) for our standard set of predictor variables. Overall, the most consistent predictor across these models was prior child-serving system involvement. In fact, this predictor was associated with a higher odds ratio across all the models, although the coefficients were not statistically significant in the models of past emotional abuse or parental separation/divorce. We also observed that being a cisgender woman increased the odds of reporting prior sexual abuse by 3.5 times (O.R. = 3.508). Sexual/gender minority youths also appear to be at an increased risk of having experienced sexual abuse (O.R. = 3.877), and past emotional abuse (O.R. = 2.277). Older youths had a significantly lower odds of reporting past sexual abuse (O.R. = 0.895). Youths who were homeless for the first time were significantly less likely to share that they had experienced physical abuse (O.R. = 0.584); however, those who were homeless for more than one year had an increased odds of reporting sexual abuse (O.R. = 1.710). Interestingly, Black or African American youths had significantly lower odds ratios for reporting past physical abuse (O.R. = 0.589), emotional abuse (O.R. = 0.606), parental separation or divorce (O.R. = 0.653), and adult mental illness (O.R. = 0.485) or substance abuse (O.R. = 0.620) in the household.

3.3. The Impact of ACEs on Current Health Status of Runaway and Homeless Youths

In

Table 6. Logistic regression analyses of health status indicators and the cumulative number of adverse childhood experiences (ACEs) among youths experiencing homelessness, 2018 Atlanta Youth Count (n = 451).

	Poor/Fair Subjective Health	Current Self-Reported Health Condition	Probable Severe Mental Illness	Probable Substance Abuse Problem
	O.R. (95% CI)	O.R. (95% CI)	O.R. (95% CI)	O.R. (95% CI)
Age (in Years)	1.152 * (1.023–1.298)	1.059 (0.964–1.162)	0.975 (0.875–1.049)	1.066 (0.969–1.172)
Race Group (Black)	0.410 *** (0.239–0.704)	0.699 (0.462–1.059)	0.644 (0.395–1.049)	1.124 (0.734–1.720)
Gender (Cisgender Woman)	1.439 (0.825–2.509)	1.186 (0.759–1.854)	2.182 ** (1.333–3.609)	0.504 * (0.318–0.801)
Sexual/Gender Minority (LGBT)	1.424 (0.820–2.474)	1.255 (0.797–1.976)	1.633 (0.986–2.703)	0.838 (0.523–1.341)
Education Level (HS or better)	0.749 (0.429–1.309)	0.841 (0.5533–1.327)	1.118 (0.658–1.900)	0.768 (0.483–1.225)
Prior System Involvement (Yes)	1.109 (0.613–2.007)	1.182 (0.754–1.853)	1.330 (0.767–2.306)	1.209 (0.767–1.906)
First Time Homeless (Yes)	0.691 (0.401–1.192)	1.465 (0.965–2.222)	0.912 (0.558–1.491)	1.111 (0.727–1.698)
Homeless for More Than 1 Year (Yes)	0.460 * (0.238–0.887)	1.222 (0.757–1.973)	0.689 (0.386–1.216)	0.709 (0.431–1.166)
Number of Adverse Childhood Events	1.047 (0.927–1.182)	1.133 ** (1.030–1.246)	1.148 * (1.026–1.231)	1.300 *** (1.176–1.438)
−2 Log Likelihood	372.996	539.206	422.190	523.493
Overall X2	28.449 ***	23.672 **	32.857 ***	43.498 ***
Nagelkerke R2	0.107	0.075	0.115	0.135

* p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001.

, we present the adjusted odds ratios of the impact of ACEs and our background characteristic measures on our four health outcome variables among runaway and homeless youths. Interestingly, the runaway and homeless youths’ subjective health assessment was not associated with cumulative ACEs in their background. Being homeless for more than a year (O.R. = 0.460) as well as being Black or African American (O.R. = 0.410) decreased the odds of reporting only poor or fair health. Older youths, however, were more likely (O.R. = 1.152) to state they were in only fair or poor health. Higher ACE scores significantly increased the odds of reporting a current health condition (O.R. = 1.133) as well as having a probable severe mental illness (O.R. = 1.148) or substance abuse problem (O.R. = 1.300). Cisgender women were significantly more likely than cisgender men to report symptoms indicative of a probable mental illness (O.R. = 2.182) and less likely than cisgender men to report clinically significant substance use problems or abuse (O.R. = 0.504).

4. Discussion

Overall, we found a significantly high prevalence of individual ACEs and cumulative ACE scores in our sample of runaway and homeless youths. Youths with prior child-serving system involvement, sexual and gender minority youths, and youths who had been homeless for more than a year had an increased odds of reporting higher ACE scores. Having more trauma exposure in one’s background, as measured by the ACEs scale, is associated with significantly greater odds of reporting a current health problem and having a probable severe mental health and/or substance use problem. Having more traumatic experiences in one’s background was not related to an overall subjective assessment of health status.

Recent studies suggest that being homeless and ACEs have been independently shown to be associated with poor overall health and chronic health conditions among homeless youths [44]. While there does appear to be a general association between greater ACE scores and reporting a chronic or major health problem, the near-term effects of ACEs among the runaway and homeless youths in our study appear to be most impactful for their current mental health and/or substance abuse status. Given that most of these respondents are in the early stages of adulthood, we speculate that the chronic health problems so frequently connected with ACEs in general population samples [2] are possibly only beginning to emerge and raise intriguing theoretical questions about how the long-term impact of ACEs may influence the development of various health conditions and overall health status over the life course. For example, findings from a longitudinal study revealed that youths with a history of three or more ACEs reported significantly higher post-traumatic stress symptoms, anxiety, depression, and a poorer quality of life than youths with no history of ACEs [45].

At the same time, the more time-proximal impact on mental health/substance use is an important finding from a public health perspective. Indeed, this study underscores the need to significantly improve early access to mental health and substance abuse services among runaway and homeless youths while meeting their needs for housing and other social services. Early access to care has the potential to help runaway and homeless youths better manage the complexities of their lives and potentially prevent the development of other chronic health problems in the longer term. More generally, the strong relationship between ACEs and health further emphasizes that such services should be organized around a trauma-informed care perspective [22,46].

Given recent research documenting a higher prevalence of ACEs among Black/African American youths than among White youths and many other racial groups [15,47,48,49,50], our finding that Black/African American runaway and homeless youths report fewer ACEs compared to other runaway and homeless youths in other race groups was unexpected. In a more detailed analysis of the ACEs categories, this difference is attributable to Black youths being less likely to report experiences of past physical, sexual, and emotional abuse; parental separation or divorce; and adult household mental illness. This pattern of results is difficult to interpret. One possibility is that Black runaway and homeless youths represent a unique sub-population of youths in the U.S. Most studies of race differences in ACEs among youths in the U.S. take advantage of national probability samples and do not account for housing status [47,48,49]. These results also could suggest an unknown systematic bias in our sampling procedures. However, the research team took the necessary steps and great care to apply systematic capture–recapture to avoid many of the weaknesses of pure convenience samples in studies of people experiencing homelessness [39]. Taking the results at face value, it is potentially possible that intersectional components of oppression could be at play with this unique sample of homeless youths, and the experience of homelessness could exacerbate the influence of other marginalized identities like gender and class (i.e., low income) to offset solely the factor of race influencing greater reports of ACEs among Black Americans in this sample [51]. Lastly, it could also reflect systematic response bias in our sample due to the “hyper-policing” of the homeless [52] and, in particular, Black/African American runaway and homeless youths [24,53]. Mistrust of formal systems due to historical and ongoing disproportionality in system involvement and persistent inequities in treatment, access, and disparate outcomes may also impact whether Black runaway and homeless youths disclose ACEs [54]. Many of our respondents openly shared concerns about their confidentiality and the risk of being reported to the police and/or state child/family protection services. This pattern, we believe, underscores the importance of culturally sensitive approaches in future research and clinical interventions. Bernard et al. [15] suggest a culturally informed ACE framework (“C-ACEs”) that is more representative of the potentially traumatic events that youths of color uniquely experience such as racial discrimination, poverty, community violence, and child welfare and juvenile justice system involvement. In fact, these Black youths were more likely to report ACEs consistent with the expanded framework, while their White counterparts were more likely to report ACEs aligned with the conventional framework [55]. Regardless, this unique finding is notable and contrasts with the prevailing trends observed among general population samples of youths.

At a more general level, our research reveals that runaway and homeless youths with prior involvement in child-serving systems are more likely to be significantly exposed to greater ACEs. This makes sense given that youths typically enter child welfare, foster care, juvenile justice, and other child-serving systems because they have been identified by someone to have experienced some form of child maltreatment, abuse or neglect, or other household problems [56,57]. There is, however, growing evidence that system involvement, particularly considering the myriad of difficulties state-supported child-serving systems face in meeting the needs of the youths in their care, can further traumatize, or even “re-traumatize”, already vulnerable youths [46,47,58,59]. From a social policy or public health perspective, the transition “out of the system” is also a critical turning point for these youths because they are often left to transition into adulthood with little, if any, support from family or friends or adequate services [59]. In many cases, youths who transition out of child-serving systems are more likely to engage in high-risk behaviors, become victims of human trafficking, and/or end up homeless and on the streets [39,60].

5. Conclusions

The results of this study indicate that adverse childhood experiences are prevalent among runaway and homeless youths and increase the odds that they experience significant mental health symptoms and engage in substance abuse risk behaviors while they are homeless. Beyond the empirical results, we believe our study has important implications both for service providers and researchers. Our findings reinforce a growing view that trauma-informed care and specific clinical interventions aimed at addressing adverse childhood experiences are critical for providing effective services in systems of care for homeless youths [61,62]. At the same time, our results stand as a reminder that runaway and homeless youths are a heterogenous population and that they may experience and make sense of adverse childhood experiences in different ways, lending support for developing more tailored, culturally sensitive approaches to trauma-informed services, such as those suggested by Bernard [15] for Black and African American youths. Looking more “upstream”, our findings call attention, once again, to the potentially important preventive role child welfare and foster care providers could play in addressing the clinical impact of adverse childhood events and in helping to reduce the number of youths who become homeless when they leave their care [46]. Lastly, the high prevalence of ACEs among homeless youths in our study has implications for ACE research and lends credence to others’ recommendations [19,21] that researchers should expand the ACE framework and methods to explicitly recognize the traumatic consequences of experiencing homelessness and extreme poverty at this critical stage of youth development.