This study documents differential health risks among youth who were homeless with or without a family member (i.e., unaccompanied) during the past year in the context of ACEs. Youth experiencing homelessness were more likely to have health problems, ranging from 10% increased odds of physical health conditions among youth who were homeless with a family member, to over 2 times the odds of co-occurring chronic physical and mental conditions among unaccompanied-homeless youth. Cumulative psychosocial risk was much higher among these youth compared to those with consistent housing – a majority of youth who were homeless with a family member had 1 or more ACEs, and nearly 1 in 5 of youth who were homeless unaccompanied had over 4 ACEs. Each additional ACE, regardless of housing status, increased the odds of low overall health by 40%; chronic physical health conditions by 16%; chronic mental health conditions by 65%; and both chronic mental and physical health conditions by 74%. However, the health risks of homelessness remained elevated to a similar degree to the risks of ACEs even after controlling for ACEs along with sociodemographic factors that are associated with both homelessness and health. Furthermore, past-year homelessness compounded the health risks of ACEs in differential ways depending upon past-year housing context (consistently-housed vs. homeless with or without family) and broad category of health outcome (overall health vs. chronic physical and/or mental health).
The finding that homelessness is associated with health problems is consistent with prior research,24,25 and is the first to our knowledge to directly compare health between youth who were consistently housed and those who were homeless in the past year with family or unaccompanied. Similarly, the finding that health problems during adolescence are more prevalent with increasing ACEs aligns with research on pediatric obesity, behavioral health, and substance abuse,13,26 as does the finding that health problems are associated with negative life events among young children experiencing concurrent family homelessness.27–29 Our findings build upon this prior knowledge by establishing that ACEs and housing status pose independent and interacting problems for children’s health. Since health problems partially mediate long-term homelessness for adults who had ACEs,30,31 our results imply that more intensive prevention, identification, and intensive treatment of chronic health problems in youth experiencing homelessness and ACEs could help “break the cycle” of homelessness across the lifespan.
Our results also highlight the complex interactions between housing, health, and ACEs. Few studies to date have compared risk gradients among subtypes of youth homelessness. While we found that being homeless with family is associated with health risks, we also found that these youth are also generally better off from a health perspective than unaccompanied-homeless youth, confirming that youth with ACEs and unaccompanied homelessness have the highest odds of co-occurring health conditions. These results are consistent with a recent study of the intersection between youth who ran away from home and those who experienced homeless, showing that those who had both run away and been homeless were at greater odds of mental health problems than either group alone (those who ran away had intermediate odds).32 Overall, our findings affirm the protective nature of caregivers and families33–37 and underscore the importance of ensuring that youth who have suffered adversity (e.g., maltreatment, neglect, or family dysfunction) are protected from becoming separated from supportive family members when faced with evictions or unsafe housing, humanitarian crises, disasters, war, or refugee/immigration policies that simultaneously threaten the integrity of families and the homes that allow health to thrive.
The strengths of this study include its use of a large, representative statewide sample, increasing statistical power to test hypothesized interactions between psychosocial factors among hard-to-reach populations of youth (i.e., those with high ACEs and/or homelessness). Additionally, the fact that these youth are enrolled and attending school provides one clear route for providing services and support. Limitations include the inability to draw causal inferences due to the cross-sectional sample; for example, health problems may increase the chances that youth experience homelessness. Furthermore, although MSS items assessing health and housing status have established associations with various outcomes in other studies,19,38 these items are self-reported and have not yet been assessed for reliability and external validity. These items also do not assess other aspects of housing such as duration or frequency of homelessness, rent instability, or housing safety. Nevertheless, assessing homelessness using a broad definition, as done with the MSS, may be more accurate than head counts.39–41 Similarly, the MSS sample selects for youth who attend school, so while it gives voice to many youth who are traditionally excluded from public health research (such as those experiencing homelessness), some of these youth who are most at risk (e.g., high ACEs; living on the street) might also be those most likely to be absent from school at the time of survey administration.42 Given that our cohort was confined to one state (and reporting on only one year of housing status), longitudinal research of national samples is needed to better ascertain the directionality and generalizability of interactions between life experiences and health (e.g., before, during, and after episodes of homelessness). We are also limited to questions asked on the Minnesota Student Survey, limiting the covariates we could include in our models that have been associated with ACEs and/or mental health, such as juvenile criminal offending.43 Finally, some of the stratified logistic regression models examining moderation of health by housing status had small sample sizes for youth with high ACEs, increasing the chances of type II errors (failing to reject false null hypotheses). We thus recommend that longitudinal nationally-representative surveys of youth health include items on both housing context and ACEs to inform program and policy development more precisely.
Our results have important clinical and policy implications for youth at high risk due to histories of maltreatment, neglect, and family dysfunction. Few pediatric clinicians ask about ACEs or document childhood adversity in the medical record.44 Notably, ACEs surveys do not typically include housing status, so even if ACEs surveillance became a routine part of pediatric or community healthcare, housing experiences and/or needs might not be considered without additional screening. While some clinical systems have piloted methods to screen for social determinants of health, including housing and ACEs, efforts are widely variable and lack consistent protocols across the field.45–48 As such efforts advance, our results – and those of others showing that even unstable or precarious housing is associated with health risks for children49,50 – suggest that the context of housing will be important to specify (i.e., within social determinants of health screening protocols that may include ACEs) because it is differentially associated with health risks.
Additionally, clinicians cite being hesitant to ask about some psychosocial topics due to lack of training in these areas as well as uncertainty as to how to resolve issues that arise from these questions.51,52 For example, within typical primary care settings, social work staff may be unavailable or overwhelmed, limited options may be available for on-site mental health care and related supports, and community-based options such as housing advocacy may not exist. Increasingly, efforts to link services such as behavioral health, food banks, housing services, and medical-legal partnerships within pediatric primary care attempt to bridge these gaps. Increased public support for such linkages and cross-sector collaborations such as school mental health and school-based clinics, along with public-private partnerships, will be essential in helping pediatric health care organizations integrate housing and ACEs screening, surveillance, and remediation into their processes that aim to address these and other social determinants of health.
In addition to trauma-informed care, advocacy is needed for “resilience-informed care” that incorporates protective factors both in clinical practice and social policies.37,53 Traditionally, policy work and medical care are focused on identifying and treating problems. Less attention is paid to what is going well, particularly in contexts of known risk, such as homelessness or ACEs. For instance, our results identify important nuances in health within the context of homelessness (i.e. level of ACEs and unaccompanied homeless versus homeless with family). These results illuminate that even within challenging situations, efforts to keep families together and ameliorate toxic stress can have important impacts on child health. Policies should aim to reflect this and recognize the protective impact of family relationships as well as other known facilitators of resilience.