Abstract
This study uses data from the state-level Youth Risk Behavioral Survey to evaluate mental health and substance use outcomes among homeless and nonhomeless adolescents in 2019.
In the US, 1 in 30 adolescents experienced homelessness between 2016 and 2017.1 Homelessness among adolescents occurs in the context of overwhelming stress, trauma, and deprivation, all of which are associated with worse mental health and substance use outcomes.2
However, reliable estimates of the psychiatric burden among homeless adolescents are unavailable to inform policies and interventions to reduce health inequities experienced by this population.3 Therefore, we evaluated mental health and substance use outcomes among homeless and nonhomeless adolescents in 2019.
Methods
The state-level Youth Risk Behavior Survey (YRBS) was used to obtain data on the characteristics, housing status, and mental health and substance use outcomes of high school students in 2019. The YRBS used a 2-stage cluster sampling design to produce a representative sample of students in 44 states (response rate of 65%).4 Students responded anonymously and voluntarily using computer-scannable booklets at school. Housing status information was available for 22 states. Homeless adolescents were identified as “individuals who lack a fixed, regular, and adequate nighttime residence” (McKinney-Vento definition).2
We assessed 4 mental health outcomes and 11 substance use outcomes. Participants were asked about persistent sadness or hopelessness for 2 or more weeks and whether they had seriously considered, planned, or attempted suicide during the past year. Any alcohol use, cigarette use, e-cigarette use, marijuana use, and binge drinking during the prior 30 days and any lifetime use of cocaine, methamphetamine, heroin, ecstasy, and injection drugs or prescription opioid misuse also were ascertained.
Outcomes among homeless and nonhomeless adolescents were compared using multivariable logistic regression models that adjusted for age, sex, race and ethnicity, sexual orientation, and state of residence. The results from the models are presented as adjusted risk ratios (RRs) and risk differences. Survey weights provided by the YRBS were applied to account for nonresponse and generate representative state-level estimates. Analyses were performed using Stata version 15 (StataCorp) and a 2-sided α level of .05. The US Centers for Disease Control and Prevention’s institutional review board approved the YRBS. Further institutional review board approval was not sought for the secondary analysis of these publicly available, deidentified data, as per institutional policy.
Results
We identified 4523 homeless adolescents and 105 864 nonhomeless adolescents across 22 states. The overall weighted prevalence of homelessness was 5.6% (95% CI, 4.3%-6.9%), and ranged from 2.2% (95% CI, 1.9%-2.5%) in New Hampshire to 13.5% (95% CI, 9.4%-17.7%) in Louisiana. Homeless adolescents were more likely than nonhomeless adolescents to be male (62.3% vs 50.0%, respectively; P < .001), to be African American or Black (20.2% vs 12.6%; P < .001), to be Hispanic (38.1% vs 25.0%; P = .01), and to identify as gay or lesbian (7.6% vs 2.5%; P < .001) (Table 1).
Table 1. Characteristics of Homeless and Nonhomeless Adolescents in the US, 2019.
| Characteristics | No. (weighted %)a | |
|---|---|---|
| Homeless adolescents (n = 4523)b | Nonhomeless adolescents (n = 105 864)b | |
| Age, y | ||
| ≤12 | 247 (2.8) | 320 (0.2) |
| 13 | 83 (0.9) | 294 (0.3) |
| 14 | 509 (11.1) | 17 758 (14.9) |
| 15 | 1010 (21.1) | 28 816 (24.7) |
| 16 | 1056 (26.4) | 26 771 (25.6) |
| 17 | 969 (21.6) | 23 144 (23.7) |
| ≥18 | 612 (16.1) | 8507 (10.6) |
| Sex | ||
| Female | 1675 (37.7) | 53 925 (50.0) |
| Male | 2567 (62.3) | 51 221 (50.0) |
| Race and ethnicityc | ||
| African American or Black | 706 (20.2) | 12 633 (12.6) |
| American Indian or Alaska Native | 143 (1.5) | 1893 (1.3) |
| Asian | 168 (2.2) | 4854 (6.1) |
| Hispanic | 1151 (38.1) | 17 304 (25.0) |
| Native Hawaiian or Other Pacific Islander | 253 (1.9) | 1873 (0.6) |
| White | 1403 (32.4) | 58 413 (49.9) |
| Multiple races (non-Hispanic) | 271 (3.5) | 6268 (4.4) |
| Sexual orientationd | ||
| Heterosexual | 2497 (70.7) | 81 255 (84.4) |
| Gay or lesbian | 398 (7.6) | 2445 (2.5) |
| Bisexual | 487 (11.1) | 8650 (8.9) |
| Not sure | 484 (10.7) | 4074 (4.2) |
The between-group comparisons made using χ2 tests were statistically significant at P < .001. The adolescents were living in Arkansas, California, Connecticut, Hawaii, Idaho, Kansas, Kentucky, Louisiana, Maine, Maryland, Massachusetts, Michigan, Montana, New Hampshire, New Mexico, North Carolina, North Dakota, Pennsylvania, Rhode Island, South Carolina, South Dakota, and Virginia.
Housing status was assessed with the following question: “During the past 30 days, where did you usually sleep?” Responses included (1) “in my parent’s or guardian’s home,” (2) “in the home of a friend, family member, or other person because I had to leave my home or my parent or guardian cannot afford housing,” (3) “in a shelter or emergency housing,” (4) “in a motel or hotel,” (5) “in a car, park, campground, or other public place,” (6) “I do not have a usual place to sleep,” (7) “somewhere else,” or (8) “in a dormitory or other school housing” (Maine only). Students who responded with items 2 through 6 were identified as experiencing homelessness.
Self-reported by participants and classified based on the 1997 US Office of Management and Budget revised standards and included given previously observed associations among race and ethnicity, homelessness, and study outcomes. Two questions were used to assess race and ethnicity: (1) “Are you Hispanic or Latino?” and (2) “What is your race?” The second question permitted the selection of multiple responses.
Assessed with the following question: “Which of the following best describes you?” Response choices included “heterosexual (straight),” “gay or lesbian,” “bisexual,” or “not sure.”
After adjusting for sociodemographics, homeless adolescents were significantly more likely than nonhomeless adolescents to report persistent sadness or hopelessness (53.0% vs 37.2%, respectively; RR, 1.41 [95% CI, 1.26-1.57]), having seriously considered suicide (44.4% vs 19.2%; RR, 2.15 [95% CI, 1.88-2.45]), planned suicide (41.8% vs 16.1%; RR, 2.34 [95% CI, 2.01-2.71]), and attempted suicide (28.0% vs 8.0%; RR, 2.70 [95% CI, 2.14-3.40]) (Table 2). Similar patterns were observed for substance use outcomes, including current cigarette use (28.7% for homeless adolescents vs 5.3% for nonhomeless adolescents; RR, 4.64 [95% CI, 4.12-5.24]), marijuana use (31.6% vs 18.4%, respectively; RR, 1.60 [95% CI, 1.20-2.12]), and binge drinking (21.4% vs 10.3%; RR, 2.04 [95% CI, 1.74-2.40]).
Table 2. Mental Health and Substance Use Among Homeless and Nonhomeless Adolescents in the US, 2019.
| Outcomesa | Weighted estimates, % (95% CI) | Adjusted risk difference, % (95% CI)b | Adjusted risk ratio (95% CI)b | |
|---|---|---|---|---|
| Homeless adolescents | Nonhomeless adolescents | |||
| Mental health | ||||
| Persistent sadness or hopelessnessc,d | 53.0 (44.9-61.0) | 37.2 (35.6-38.8) | 15.2 (8.9-21.5)e | 1.41 (1.26-1.57)e |
| Seriously considered suicided | 44.4 (32.6-56.8) | 19.2 (17.2-21.4) | 22.3 (14.8-29.9)e | 2.15 (1.88-2.45)e |
| Planned suicided | 41.8 (29.9-54.6) | 16.1 (14.0-18.3) | 21.7 (14.0-29.4)e | 2.34 (2.01-2.71)e |
| Attempted suicidef | 28.0 (24.2-32.2) | 8.0 (7.4-8.7) | 13.5 (8.6-18.4)e | 2.70 (2.14-3.40)e |
| Substance use | ||||
| Currentg | ||||
| Alcohol | 37.0 (30.0-44.7) | 24.1 (23.0-25.2) | 10.9 (4.5-17.2)e | 1.46 (1.21-1.76)e |
| Cigarette | 28.7 (25.6-32.0) | 5.3 (4.9-5.8) | 18.9 (16.0-21.8)e | 4.64 (4.12-5.24)e |
| e-Cigarette | 45.2 (40.2-50.4) | 22.6 (21.1-24.2) | 21.2 (15.8-26.5)e | 1.93(1.68-2.21)e |
| Marijuana | 31.6 (23.6-40.9) | 18.4 (17.2-19.7) | 11.1 (2.7-19.4)h | 1.60 (1.20-2.12)h |
| Binge drinkingi | 21.4 (16.9-26.7) | 10.3 (9.6-11.1) | 10.4 (7.0-13.8)e | 2.04 (1.74-2.40)e |
| Lifetimej | ||||
| Cocaine | 32.2 (28.7-36.0) | 2.9 (2.6-3.2) | 18.8 (15.9-21.6)e | 8.17 (6.94-9.62)e |
| Methamphetamine | 36.0 (27.2-45.8) | 3.0 (1.6-5.5) | 19.3 (15.3-23.3)e | 7.05 (3.86-12.85)e |
| Heroin | 28.0 (24.8-31.5) | 1.3 (1.1-1.5) | 13.8 (11.2-16.4)e | 13.14 (10.16-17.00)e |
| Ecstasy | 32.9 (27.7-38.4) | 3.6 (2.6-4.9) | 17.3 (13.6-21.0)e | 5.81 (3.89-8.66)e |
| Injection drug | 28.1 (22.8-34.1) | 2.4 (1.4-4.2) | 14.6 (10.6-18.7)e | 6.65 (3.49-12.69)e |
| Prescription opioid misuse | 31.3 (24.7-38.8) | 12.9 (11.8-14.1) | 14.8 (7.4-22.2)e | 2.18 (1.65-2.87)e |
Missingness did not exceed 10% for any outcome. Missing data were deleted listwise on analytic variables.
Multivariable logistic regression models controlled for age (≤12, 13, 14, 15, 16, 17, or ≥18 years), sex (female or male), race and ethnicity (African American or Black, American Indian or Alaska Native, Asian, Hispanic, Native Hawaiian or Other Pacific Islander, White, or multiple races [non-Hispanic]), sexual orientation (heterosexual, gay or lesbian, bisexual, or not sure), and state of residence.
Felt so sad or hopeless almost every day for 2 weeks or more in a row that the participant stopped doing some usual activities.
During the 12 months before the survey.
P≤.001.
One or more times during the 12 months before the survey.
Used 1 or more times during the 30 days before the survey.
P = .01.
Had 5 or more drinks (male) or 4 or more drinks (female) in a row within a couple of hours 1 or more times during the 30 days before the survey.
Used 1 or more times during lifetime.
In addition, lifetime cocaine use was significantly higher among homeless adolescents vs nonhomeless adolescents (32.2% vs 2.9%; RR, 8.17 [95% CI, 6.94-9.62]), as were methamphetamine use (36.0% vs 3.0%; RR, 7.05 [95% CI, 3.86-12.85]), heroin use (28.0% vs 1.3%; RR, 13.14 [95% CI, 10.16-17.00]), ecstasy use (32.9% vs 3.6%; RR, 5.81 [95% CI, 3.89-8.66]), injection drug use (28.1% vs 2.4%; RR, 6.65 [95% CI, 3.49-12.69]), and prescription opioid misuse (31.3% vs 12.9%; RR, 2.18 [95% CI, 1.65-2.87]). The adjusted risk differences appear in Table 2.
Discussion
This study found that homeless adolescents experienced poor mental health and substance use outcomes in the US compared with nonhomeless adolescents. Homeless adolescents were more likely to have seriously considered or attempted suicide than nonhomeless adolescents and were more likely to have used cocaine, methamphetamine, or heroin. The disproportionate representation of racial, ethnic, and sexual minority groups among homeless adolescents in this study may reflect enduring structural injustices such as systemic racism and homophobia.5
Study limitations include a survey response rate of 65% and lack of information on gender identity and whether the adolescent was part of a homeless family or alone.
Policy strategies that increase investment in social protections and child welfare systems, and that enable health care clinicians and schools to identify and refer homeless adolescents to evidence-based interventions,6 may help address upstream determinants of homelessness and improve mental health and substance use outcomes.3
Section Editors: Jody W. Zylke, MD, Deputy Editor; Kristin Walter, MD, Associate Editor.
References
- 1.Morton M, Dworsky A, Samules GM, Patel S; US Department of Housing and Urban Development . Voices of youth count comprehensive report: youth homelessness in America. Published 2018. Accessed January 4, 2022. https://www.huduser.gov/portal/sites/default/files/pdf/Voices-of-Youth-Report.pdf
- 2.Grant R, Gracy D, Goldsmith G, Shapiro A, Redlener IE. Twenty-five years of child and family homelessness: where are we now? Am J Public Health. 2013;103(suppl 2):e1-e10. doi: 10.2105/AJPH.2013.301618 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Embleton L, Lee H, Gunn J, Ayuku D, Braitstein P. Causes of child and youth homelessness in developed and developing countries: a systematic review and meta-analysis. JAMA Pediatr. 2016;170(5):435-444. doi: 10.1001/jamapediatrics.2016.0156 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Underwood JM, Brener N, Thornton J, et al. Overview and methods for the Youth Risk Behavior Surveillance System—United States, 2019. MMWR Suppl. 2020;69(1):1-10. doi: 10.15585/mmwr.su6901a1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ortega AN, Roby DH. Ending structural racism in the US health care system to eliminate health care inequities. JAMA. 2021;326(7):613-615. doi: 10.1001/jama.2021.11160 [DOI] [PubMed] [Google Scholar]
- 6.Wang JZ, Mott S, Magwood O, et al. The impact of interventions for youth experiencing homelessness on housing, mental health, substance use, and family cohesion: a systematic review. BMC Public Health. 2019;19(1):1528. doi: 10.1186/s12889-019-7856-0 [DOI] [PMC free article] [PubMed] [Google Scholar]