Living situation and substance use may be key predictors of risky sexual behavior among newly homeless young people. In a sample of Los Angeles County youth who had recently left home, the odds of reporting multiple sexual partners were lower for males who were housed in institutional settings than for those living in nonfamily settings, and were elevated for both females and males who used alcohol or other drugs. Females who lived in an institution or with family were more likely than others to report consistent condom use, while those who used drugs had reduced odds of engaging in this protective behavior; for males, none of the characteristics studied were associated with condom use.1

The young people in the study were recruited at shelters, drop-in centers and street hangouts in 2001–2002; to be eligible, they had to be 12–20 years old and to have been away from home for at least one night but for no more than six months. Participants completed interviews, conducted by specially trained research staff with the aid of audio computer-assisted self-interviewing technology, when they entered the study and then at specific intervals for up to 24 months. Demographic data were collected at baseline; follow-up interviews covered current housing situation (categorized as family, institutional or other, “nonfamily,” setting) and emotional distress (assessed with a standard scale), as well as substance use and sexual risk behaviors (number of partners, types of partners and consistency of condom use) in the past 90 days. Questions on sexual behavior referred to experience with vaginal and anal intercourse. The researchers used longitudinal random intercept effects models to identify characteristics associated with participants’ reports of multiple sexual partners and of consistent condom use; they constructed separate models for females and males.

A total of 261 young people enrolled in the study—156 females and 105 males, whose average age was 15.5. The participants came from diverse racial and ethnic backgrounds: Thirty-three percent were U.S.-born Latinos, 30% were black, 23% were white or Asian, and 15% were foreign-born Latinos. At baseline, 78% were living in institutional settings (e.g., shelters, boarding schools, group homes), 10% with family members and 12% in other situations. Three-quarters of the sample were sexually experienced, reporting an average of 1–2 partners. However, one-half had not had a partner in the past three months. Only about a third of those reporting a recent partner said that they always used condoms. Seventy percent of the young people reported using alcohol or drugs in the past 90 days; 16% showed signs of depression or anxiety.

In the multivariate analyses, females’ odds of reporting multiple partners increased with length of participation in the study and with age (odds ratio, 1.1 for each), and were elevated among those reporting use of various substances (1.3–1.6). The likelihood of having had multiple partners was lower among Latinas than among white or Asian females (0.8 for immigrant Latinas and 0.9 for those who were U.S.-born). By contrast, only living situation and substance use were associated with men’s reports of multiple partners: The odds were reduced among those living with relatives and those housed in an institutional setting (0.8 for each), and were elevated among those who said they had recently used alcohol or other drugs (1.4–2.3).

A number of characteristics were identified as predictors of consistent condom use for females, but none were significant in the analyses for males. Duration of study participation was positively associated with females’ reports of consistent use (odds ratio, 1.5), as was living with family members (4.0) or in an institutional setting (6.5). Females reporting substance use had reduced odds of saying that they always used condoms, whether they had a serious, monogamous partner or a casual one (0.3–0.4).

While acknowledging the limitations of their sample and the possibility that unmeasured factors affected the results, the researchers conclude that living situation and substance use “appear to be the most salient [predictors of] sexual risk” among newly homeless young people. They speculate that the findings regarding young people’s being housed in institutional or family settings reflect the supervision and social support that may be available in these environments. Therefore, they recommend that interventions targeting newly homeless youth help them find housing with such support, as well as help them reduce their levels of substance use.


1. Solorio MR et al., Predictors of sexual risk behaviors among newly homeless youth: a longitudinal study, Journal of Adolescent Health, 2008, 42(4):401–409.