Mental Health Outcomes Among Homeless, Runaway, and Stably Housed Youth

We used data from the 2016 minnesota Student Survey (MSS), a statewide, school-based surveillance system coordinated by the Departments of Education, Health, Human Services, and Public Safety.²⁵  The MSS anonymously assessed students’ health, health-related behaviors, and sociodemographic characteristics. It was administered to fifth-, eighth-, ninth- and 11th-grade students at public, tribal, and charter schools across Minnesota. All public schools in Minnesota were invited to participate in the MSS; 85% of schools participated in 2016. Parents were notified of MSS administration and could opt out if they did not want their children to participate. Students could decline, skip questions, or discontinue the survey at any time; 83% of ninth- and 70% of 11th-graders participated. Surveys were conducted in English. Details about the MSS can be found elsewhere.²⁵  The University of Minnesota Institutional Review Board determined that this secondary analysis was exempt from review by a human research ethics committee.

We restricted our sample to students in the ninth and 11th grades who answered questions regarding housing status and mental health outcomes. To focus our analysis on the influence of unaccompanied homelessness (ie, homelessness without an adult family member) on mental health outcomes, we excluded those who reported experiencing family homelessness without unaccompanied homelessness. Our analytic sample included 68 785 students who had complete data for all covariates, including abuse history and demographics.

We classified students’ housing status during the past 12 months using 2 questions. We first created an indicator of unaccompanied homelessness using the question, “During the past 12 months, have you stayed in a shelter, somewhere not intended as a place to live, or someone else’s home because you had no other place to stay?” We categorized those who answered “yes, I was on my own without any adult family” as having experienced unaccompanied homelessness. We then created an indicator of runaway status using the question, “During the last 12 months, how often have you run away from home?” We categorized students who answered “once or twice,” “3 to 5 times,” “6 to 9 times,” or “10 or more times” as having run away. We then combined these indicators to create 4 mutually exclusive groups: (1) stably housed youth (those who had neither run away nor experienced unaccompanied homelessness), (2) unaccompanied homeless youth (those who had experienced unaccompanied homelessness but had not run away), (3) runaway youth (those who had run away but denied unaccompanied homelessness), and (4) youth who had both run away and been homeless. The runaway group could therefore include any youth who had run away from home briefly, without staying away overnight, and might also include youth who stayed with other family or friends and therefore may not have considered themselves to have “no other place to stay.”

We examined 4 dichotomous (yes or no) mental health outcomes: self-injury, suicidal ideation, suicide attempts, and depressive symptoms. The survey assessed self-injury with the question, “During the last 12 months, how many times did you do something to purposely hurt or injure yourself without wanting to die, such as cutting, burning, or bruising yourself on purpose?” Drawing on a previously established approach to the nonnormal distribution of this variable,²⁶  we classified students as having self-injurious behavior if they answered as having done so 1 or more times.

We assessed suicidal ideation using the question, “Have you ever seriously considered attempting suicide?” We classified students as reporting suicidal ideation if they answered “yes, during the past year.” We chose this time frame to be consistent with our housing-related measure. We used a similar approach to create an indicator of suicide attempt using responses to the question, “Have you ever actually attempted suicide: mark all that apply.” We classified students as having attempted suicide if they responded “yes, during the past year.”

Surveys included the Patient Health Questionnaire-2 (PHQ-2), a commonly used screen that assesses depressive symptoms with 2 questions: “Over the last 2 weeks, how often have you been bothered by: (a) little interest or pleasure in doing things and (b) feeling down, depressed, or hopeless.” Response options included not at all (0 points), several days (1 point), more than half the days (2 points), and nearly every day (3 points). The 2 responses were then summed with a maximum score of 6. Although a score of 3 or greater is commonly used as a cut point in primary care with general populations of youth, given the well-documented high burden of mental illness among runaway and homeless youth,^4,18,24,27  we classified students with a score of 4 or higher as having depressive symptoms to increase specificity for major depression (sensitivity 54%; specificity 89%).^28,29 

The MSS assessed sociodemographic and health-related variables that have well-established associations with housing status and/or mental health.^1,14,30  Students answered questions regarding grade, race and/or ethnicity, biological sex, sexual orientation, and area of residence (residing in the 7-county Twin Cities area or outside the metropolitan area). The survey assessed both biological sex and identification as gender nonconforming (eg, transgender or genderfluid). We used biological sex in our analysis given the low percentage of students identifying as gender nonconforming (2.4%). We used student report of receiving free or reduced-price lunch as an indicator of poverty.^26,31  Four dichotomous (yes or no) questions assessed different types of intrafamilial and extrafamilial abuse. Students indicated whether they had ever experienced verbal or physical abuse by a parent or other adult in their household, sexual abuse by an older or stronger member of their family, or sexual abuse by an adult or other person outside their family.

We used χ² analyses to assess differences in demographics, abuse, and mental health by housing status. We then assessed the relationship between housing status and each of the mental health outcomes using separate multivariable logistic regression models controlling for age, grade, race and/or ethnicity, biological sex, sexual orientation, poverty status, area of residence, and abuse history. We used Stata version 15.1 (Stata Corp, College Station, TX) to conduct all analyses. Statistical tests were 2 tailed with a critical α of .05.

Most youth in our sample (95%) reported being stably housed in the past year; 0.5% had experienced unaccompanied homelessness, 4% had run away from home, and 0.6% reported both homelessness and having run away (Table 1). Overall, roughly half of students reported being in ninth grade (54%) or being male (49%). The sample was racially and ethnically diverse. With respect to sexuality, 23% of unaccompanied homeless youth, 22% of runaways, and 30% in the both group identified as lesbian, gay, or questioning compared with 10% of stably housed peers. Housing groups differed (P < .001) with regard to most characteristics except for area of residence; youth in the Twin Cities metropolitan and Greater Minnesota areas had similar rates of experiencing unaccompanied homelessness and running away.

Self-injury, suicidal ideation, suicide attempts, and depressive symptoms were frequently reported among youth who experienced homelessness and/or running away (Fig 1). Across mental health outcomes, youth who had experienced both unaccompanied homelessness and had run away had the highest prevalence of each outcome, followed by those who had run away, then those who had experienced unaccompanied homelessness, with the lowest prevalence seen among those who had been stably housed. For example, one-third (33%) of youth in the both group reported suicide attempts in the past year; the percentages of youth in the runaway, unaccompanied homeless, and stably housed groups who reported suicide attempts were 20%, 11%, and 2%, respectively.

Findings from multivariable analyses controlling for demographics and history of abuse followed a similar pattern. Compared with youth who were stably housed, youth in the unaccompanied homeless, runaway, and both groups had greater odds of experiencing self-injury, suicidal ideation, and suicide attempts (all P < .01; Table 2). Youth who had run away and youth who had experienced both unaccompanied homelessness and running away had greater odds of having depressive symptoms compared with stably housed youth, but there was no difference in depressive symptoms between young people who had experienced homelessness only and those who were stably housed.

Findings from this statewide study demonstrate important differences in mental health outcomes between youth based on their experiences of homelessness and running away. Youth who both ran away from home and experienced homelessness in the past year had >7 times greater odds of attempting suicide compared with their housed peers who had not run away, and this risk remained elevated (albeit to a somewhat lesser degree) for youth who had run away but had stable housing and for unaccompanied homeless youth who had not run away. These findings extend previous work regarding the heterogeneity of youth who experience unaccompanied homelessness. Importantly, the vast majority of youth in our data who reported running away did not also report unaccompanied homelessness, suggesting that these are distinct, but sometimes overlapping, groups. Although other studies suggest that youth who have run away are at particularly high risk for poor mental health outcomes,^18,19,32  to our knowledge, this study is the first to demonstrate that running away confers significant elevated mental health risk independent from that associated with unaccompanied homelessness.

Previous research has explored typologies of housing instability and homelessness, highlighting the heterogeneity of experiences of unaccompanied homeless and runaway youth,^8,10,11,33  but there have been few recent studies that separate and compare health outcomes among these populations.^1,10  Previous literature has classified youth by motive for leaving the home.^8,10,11  Others have proposed a needs-based typology of homeless youth, advocating for multifaceted service programs to address the needs of homeless youth.³³  Voices of Youth Count¹  data provide much-needed information about “explicit homelessness” (including running away) and couch surfing, highlighting the importance of a more inclusive definition of homelessness. Our findings reveal that youth with different housing experiences may have unique unmet mental health needs, suggesting that it may be important to account for housing experience when designing interventions for this population.

Consistent with other studies,¹⁰  we found significant sociodemographic differences between housing groups, including differences in biological sex, race and ethnicity, sexual orientation, and abuse history, providing further support that runaway and homeless youth may represent unique populations. Lesbian, gay, bisexual, transgender, and queer youth in particular were overrepresented among all unstably housed groups, with particularly high prevalence (30%) in the both group, similar to previous literature documenting elevated levels of homelessness among lesbian, gay, bisexual, transgender, and queer youth.^1,34,35  However, even when controlling for these differences, our analyses revealed significant differences in the experience of mental health outcomes between groups.

Our study underscores the importance of recognizing these differences and their associated health implications. Although it is well established that youth who have experienced homelessness or have run away are at high risk of poor mental health outcomes, and that poor mental health may precipitate runaway episodes and persist into adulthood,^18,27,32  to our knowledge, there are no recent studies comparing mental health outcomes between runaway and homeless youth. We found that running away is associated with significant mental health risk even when it is not associated with unaccompanied homelessness. This was unsurprising given the well-documented association of runaway episodes with poor family function.^13,16,32,36  Strong family connections are a known protective factor for all adolescents,^31,37  including runaway youth.^24,32,38  Runaway episodes are also associated with increased risk of trauma and exploitation, which could contribute to the observed increased mental health risk.^{14,21,24,39–41}  Although we were able to control for some forms of abuse, these measures may not capture exploitation and other forms of trauma experienced while away from home. Further study is needed to elucidate the underlying factors contributing to the increased risk of poor mental health associated with running away.

Running away was extremely common within our school-based sample, with 5% of youth reporting that they had run away at least once in the past year. Previous studies describing the prevalence of runaway episodes are inconsistent, ranging from 6%¹³  to 5.3% per 1000.⁴²  These differences are likely due to variations in survey methodologies and definitions of running away. Importantly, running away is thought to be a precursor to homelessness.^43,44  Given the high prevalence of running away and its strong association with poor mental health outcomes, identifying youth who have run away may provide an opportunity to provide support and prevent progression to chronic homelessness.

Our findings have important policy and practice implications. Federal government agencies frequently group runaway and homeless youth together in services and legislation, yet these populations may benefit from more targeted interventions. Inclusive definitions of unaccompanied youth homelessness and runaway youth are critical, as is recognition of the nuanced differences between subgroups of unaccompanied homeless youth.^1,45  Research, policy, and practice should account for important differences between these groups.

Our study reveals that runaway youth are at particularly high risk even when they have not had traditional experiences of unaccompanied homelessness, such as living on the street or in shelters or couch surfing. Few targeted programs exist for youth who have run away, and fewer have been evaluated. The National Runaway Safeline, a 24/7 phone line for youth who have run away and their families, represents one example of a targeted national intervention and resource for runaway youth.⁴⁶  Minnesota’s nurse-led Runaway Intervention Program incorporates home visiting, intensive case management, and empowerment groups, providing an example of a novel and effective targeted intervention for runaway youth, which has shown promise for improving mental health outcomes.^47,48 

The high prevalence of runaway and homeless youth in this school-based sample suggests that professionals across sectors likely interact with young people experiencing these conditions. Intervening early in the cycle of youth homelessness, while youth remain connected with communities and services, may be critical to preventing progression to more chronic homelessness.⁴⁹  Schools and youth-serving community agencies may benefit from training on how to identify and intervene with young people with these experiences. Health care professionals also have a critical role to play in screening for runaway episodes and homelessness and working to support youth and families who are at risk. Further research is needed to clarify best practices for screening in these settings and intervening to address running away and unaccompanied youth homelessness.

Given the particularly high risk for adverse outcomes among youth who experienced both homelessness and running away, youth housing programs should work to connect runaway youth in shelters with mental health support and services. However, policies that allow homeless youth to access and consent to mental health care are needed to make these efforts a reality,⁵⁰  underscoring the need for intervention at multiple levels to support the health needs of vulnerable young people.

Our study is strengthened by the use of data from a statewide sample with adequate size to distinguish between different homeless and runaway experiences not found with other population-based data. Limitations include a cross-sectional design, which prevents causal inference, and self-reported measures, which may result in response bias. However, other research finds that youth can reliably report engagement in risk behaviors⁵¹  and that self-reported experience of homelessness may be more accurate than other methods of enumeration.^52,53  Although the MSS provides rich information regarding those youth who are enrolled in school, it does not capture youth who may not be enrolled or who do not attend, a population that is likely overrepresented among homeless and runaway youth. Thus, experiences of homelessness and running away may be underreported in our sample. Generalizability of findings to youth in other contexts (eg, those out of school, youth in juvenile corrections facilities, or those in other geographic locations) remains to be established. Furthermore, the MSS did not assess the duration of homelessness or runaway experience or other contextual information that could be important for understanding youth experiences.

The experience of running away is associated with increased risk of poor mental health above and beyond that of unaccompanied homelessness. Tailored clinical and community interventions to meet the unique needs of homeless and runaway youth are critical. Clinicians in primary care, school health, and community settings (eg, shelter-based health care) are well positioned to identify and intervene with runaway and homeless youth. Additional research is needed regarding clinical best practice for identifying and intervening to support youth and families who may be at risk for or are experiencing housing instability.

MSS

Minnesota Student Survey

PHQ-2

Patient Health Questionnaire-2