Homeless persons are more likely than other people to engage in behaviors that are associated with increased risk of acquiring or transmitting HIV including substance use, injection drug use (IDU) and needle sharing, risky sexual practices, and exchanging sex for money, drugs, or a place to stay.1-7 HIV rates are 3-9 times higher in homeless persons compared with those in stable housing.1,2,8 Homeless people living with HIV or AIDS (PLWHA) have poorer mental and physical health than housed PLWHA.9 Homeless PLWHA are also more likely to have high HIV viral load levels,9,10 which increase the likelihood of HIV transmission to others during risky sexual or drug use behaviors,11 yet minimal research has been conducted on substance use and sexual risk behaviors of homeless PLWHA.
Preventing HIV transmission by people who are known to be HIV-seropositive is a public health priority.12 This study used data from a large, multisite study of PLWHA to: (1) investigate differences between homeless and housed PLWHA on sociodemographic, sexual, drug, and alcohol risk factors associated with HIV transmission; and (2) examine the association between housing status and behavioral outcomes associated with HIV transmission risk.
Data were from the Centers for Disease Control and Prevention's Supplement to HIV/AIDS Surveillance (SHAS) project, a cross-sectional, multisite study of adults with HIV.9,13 Eligible persons were recruited from those recently reported to have HIV or AIDS to 19 health departments in the United States. Recruitment was either (1) facility-based at selected health care facilities (9 sites) or (2) population-based from defined geographical areas (10 sites). Potential participants had been reported to the local HIV/AIDS surveillance system within the previous 2 years. Recruitment occurred through printed materials, provider referral, and review of clinic lists and HIV/AIDS surveillance registries. Participant compensation was a site decision, and a majority offered participants up to $25 or a food/merchandise voucher.
Trained staff conducted individual interviews with people 18 years and older. Interviews lasted about 45 minutes and were conducted in either English or Spanish. Questionnaire modules included demographic information, alcohol and drug use, sex behaviors, reproductive history (for women), HIV testing and medical therapy, and health and social service use.
Data for the present study were collected between May 2000 and December 2003. During this period, 21,094 persons were eligible for interview; 8732 of these were unable to be located, were deceased, or did not have an attempted contact. Among 12,362 eligible persons offered enrollment, 8279 (67%) participated in an interview and 4083 (33%) refused. A total of 8075 persons had valid housing data (41% from population-based sites and 59% from facility-based sites) and were included in the current study.
Current Housing Status
People were categorized as homeless if they reported living in a shelter or on the streets at the time of the interview. Consistent with an earlier report,9 respondents were considered housed at the time of the interview if they were living in (1) a house or apartment alone, with spouse, partner, friends, or family; (2) a medical care facility; or (3) a correctional institution.
Sex, race/ethnicity, age, highest education level completed, marital status, employment status, annual household income, and primary source of income were assessed. HIV risk categories were based on lifetime behaviors.
Alcohol and Drug Use
Potential problem drinking was assessed using the 4-item CAGE scale (range 0-4),14,15 with scores of 2-4 indicating a potential lifetime drinking disorder. Participants were asked whether they had used any injected and any noninjected illicit drugs ever or in the past 12 months.
Participants were categorized as sexually active in the past 12 months if they had vaginal, oral, or anal sex. Variables representing sexual transmission risk (ie, unprotected vaginal or anal sex at last sex) by partner's HIV serostatus (positive, negative, or unknown HIV status) and type of partner (main, nonmain) were computed.
Bivariate analyses comparing homeless and housed respondents were conducted using χ2 tests (categorical variables) and t tests (continuous variables). The significance criterion for all models was set at P < 0.05.
To investigate the independent effects of homelessness, logistic regression analyses were conducted separately with alcohol use, drug use, and sexual behaviors as dependent variables. For each dependent variable, several separate logistic regressions were conducted. The first used housing status as a single predictor. The predictor variables for the second set of regression analyses included housing status and sociodemographic variables (HIV risk group, age, sex, race/ethnicity, marital status, education, annual household income, and employment status). If data were missing on 1 or more covariates, the respondent's data were excluded from the regression analyses (n = 674, 8%).
A third set of logistic regression analyses repeated the previous analyses for only the sexual behavior variables with the number of drugs used, IDU, and problem drinking (CAGE) added as covariates. The purpose of these analyses was to examine the effects of homelessness on risky sexual behavior, controlling for differences in alcohol and drug use across housing status group.
A total of 8075 people were included in the analyses, 310 (3.8%) of whom were currently homeless. More than 70% were male (with homeless significantly more likely to be male), and more than half were black (Table 1). Mean age was 40 (SD = 9.2), and nearly two thirds were single (never married). Homeless respondents had lower education and were less likely to be employed. The homeless group earned less and was more likely to receive public assistance or have no identified income source. In the housed group, there was a higher percentage with the HIV transmission category of men who have sex with other men (MSM), whereas the homeless group had a higher percentage who reported HIV risk through IDU and engaging in both MSM and IDU behaviors (MSM/IDUs).
Homeless respondents were significantly more likely to have used any illicit drugs in the past 12 months, injected drugs in their lifetime and in the past 12 months, and have more potential alcohol abuse (Table 1). Homeless respondents also reported using a significantly greater number of different drugs in the past 12 months (mean = 1.33, SD = 1.73) than housed respondents (mean = 0.57, SD = 1.12, t = 7.5, P < 0.0001).
In the multivariable analyses that controlled for sociodemographic variables, housing status remained a significant predictor of all the substance use variables: possible alcohol abuse [adjusted odds ratio (AOR) = 1.63, 95% confidence interval (CI) = 1.26 to 2.10], drug use in past 12 months (AOR = 2.31, 95% CI = 1.77 to 3.00), IDU ever (AOR = 1.90, 95% CI = 1.46 to 2.47), and IDU in past 12 months (AOR = 2.75, 95% CI = 1.87 to 4.06).
Housed respondents were more likely to be sexually active in the past 12 months (Table 2). However, homeless respondents had more sex partners and were more likely to have unprotected sex with an unknown status partner at last sex (Table 2) and a negative or unknown status nonmain partner (not shown in table; odds ratio = 1.71, 95% CI = 1.12 to 2.63). There were no differences for unprotected sex with other sex partners.
In the multivariable analyses, homeless participants were more likely to have more sex partners during their lifetime and in the past 12 months (among those sexually active) and were more likely to exchange sex during their lifetime and in the past 12 months (Table 2). Homeless respondents were also more likely to have unprotected sex with an unknown status partner (Table 2) and a negative or unknown status nonmain partner (not shown in table; AOR = 1.73, 95% CI = 1.11 to 2.70).
When the alcohol and drug use variables were added to the multivariable analyses as additional predictors, housing status remained a significant predictor of number of sex partners in the past 12 months, sex exchange ever and in the past 12 months, and unprotected sex with an unknown status partner (Table 2).
Some of the respondents who were hospitalized (n = 157) or in a correctional facility (n = 15) at the time of interview may have been homeless before entering these facilities. Thus, all multivariate analyses were repeated without data from these respondents. The findings were the same as those for the entire sample.
This is one of the first large-scale studies to demonstrate that homeless PLWHA are at greater risk than housed PLWHA for substance abuse and potential transmission of HIV to others. This increased risk remained significant even after controlling for potentially confounding factors. Although housed PLWHA were more likely to be sexually active, homeless respondents who were sexually active reported a greater number of sexual partners in the prior 12 months, were more likely to exchange sex for money or drugs, and were nearly twice as likely to engage in unprotected sex. These results remained significant even when controlling for alcohol and drug use, indicating that they cannot be explained by differences in substance use. These results indicate that homeless PLWHA should be a special priority for programs that seek to reduce substance abuse, its harmful effects, and the further spread of HIV.
There are some limitations of this study.9 Causality cannot be reliably determined from these cross-sectional data. It is possible that those who reported being homeless were more likely to engage in risk behavior when they became homeless or that factors associated with substance use, exchange of sex for money or drugs, or other unmeasured factors (such as mental illness) contributed to homelessness. SHAS participants are not representative of all PLWHA in the United States, and it is unknown how inability to locate persons or refusal to participate might have affected the results (eg, due to mental illness or substance use). Data were self-reported and are subject to socially desirable responding and recall biases. The number of PLWHA who recently experienced homelessness is likely to be underestimated, and no information was collected regarding prior housing situations. SHAS did not specifically target homeless persons, who are typically more difficult to locate and may be less engaged in medical care. Thus, those most severely affected by homelessness (street homeless and those unconnected to services) are likely underrepresented, and these findings are likely a conservative estimate of the association between housing status and risk behaviors. Even using what is likely to be a conservative estimate from this study (4%), tens of thousands of the estimated 1 million plus people in the United States who are HIV positive16 may be homeless at any given time.
The increased rates of substance use and HIV transmission risk among homeless PLWHA indicate the importance of screening for housing status, delivering behavioral interventions that have been shown to reduce risk behavior among PLWHA (eg, condom use, serostatus disclosure to partners), and the potential role of placing and maintaining homeless PLWHA in stable housing.3,9,17 Health departments and individual providers of medical and social services all have important roles to play in better addressing the housing, behavioral, and health-related needs of homeless PLWHA.
1. Aidala A, Cross JE, Stall R, et al. Housing status and HIV risk behaviors: implications for prevention and policy. AIDS Behav
2. Culhane DP, Gollub E, Kuhn R, et al. The co-occurrence of AIDS and homelessness: results from the integration of administrative databases for AIDS surveillance and public shelter utilisation in Philadelphia. J Epidemiol Community Health
3. Wolitski RJ, Kidder DP, Fenton KA. HIV, Homelessness, and public health: critical issues and a call for increased action. AIDS Behav
4. Aidala A, Sumartojo E. Why housing? AIDS Behav
5. Kidder DP, Wolitski RJ, Royal S, et al, for the Housing and Health Study Team. Access to housing as a structural intervention for homeless and unstably housed people living with HIV: rationale, methods, and implementation of the Housing and Health Study. AIDS Behav
6. Kipke MD, Weiss G, Wong CF. Residential status as a risk factor for drug use and HIV risk among young men who have sex with men. AIDS Behav
7. Wenzel SL, Tucker JS, Elliott MN, et al. Sexual risk among impoverished women: understanding the role of housing status. AIDS Behav
8. Estebanez PE, Russell NK, Aguilar MD, et al. Women, drugs and HIV/AIDS: results of a multicentre European study. Int J Epidemiol
9. Kidder DP, Wolitski RJ, Campsmith ML, et al. Health status, health care use, medication use, and medication adherence in homeless people living with HIV/AIDS. Am J Public Health
10. Knowlton A, Arnsten J, Eldred L, et al, and the INSPIRE Team. Individual, interpersonal, and structural correlates of effective HAART use among urban active injection drug users. J Acquir Immune Defic Syndr
11. Montaner JSG, Hogg R, Wood E, et al. The case for expanding access to highly active antiretroviral therapy to curb the growth of the HIV epidemic. Lancet
12. Janssen RS, Holtgrave DR, Valdiserri RO, et al. The serostatus approach to fighting the HIV epidemic: prevention strategies for infected individuals. Am J Public Health
13. Buehler JW, Diaz T, Hersh BS, et al. The supplement to HIV-AIDS surveillance project: an approach for monitoring HIV risk behaviors. Public Health Rep
14. Buchsbaum DG, Buchanan RG, Centor RM, et al. Screening for alcohol abuse using CAGE scores and likelihood ratios. Ann Intern Med
15. Ewing JA. Detecting alcoholism: the CAGE questionnaire. JAMA
16. Glynn M, Rhodes P. Estimated HIV prevalence in the United States at the end of 2003. Presented at: National HIV Prevention Conference; 2005; Atlanta, GA.
17. Shubert V, Bernstein N. Moving from fact to policy: housing is HIV prevention and health care. AIDS Behav