The prevalence of drug use was found to be very high among homeless individuals in Toronto compared to rates previously reported for the general population. Lifetime drug use among participants was 60%; in contrast, a study by Vega and colleagues (2002) on the prevalence of drug use at 7 international sites found rates of 32% for men and 20% for women in the province of Ontario . A similar pattern was observed for the use of specific drugs. The prevalence of cocaine use within the past 2 years among study participants (27%) was 27-fold higher than the corresponding figure of 1% in the general Canadian population, as reported by the Toronto Drug Strategy Advisory Committee (2005) . Regular marijuana use (40%) was almost three times higher than that found in the general population (14%) in the 2006 Canadian Addiction Survey .
High rates of drug use among homeless people result from a number of processes. Drug use is a risk factor for becoming homeless . Drug use is also a risk factor for prolonged homelessness , which is confirmed by our data. As a result of this association, drug users are over-represented in cross-sectional surveys of homeless populations . Homelessness may increase the likelihood that an individual will use drugs.
However, only about one-quarter of participants with current drug problems identified drug and/or alcohol use as an impediment to acquiring stable housing, which is consistent with a recent survey of 368 homeless individuals in Toronto in which only 23% of participants identified their drug or alcohol use as the main reason for becoming homeless . Some homeless individuals may be in a state of denial or lack insight regarding the impact of their substance use. Alternatively, this finding may reflect the multifactorial causes of homelessness. Either of these possibilities has practical implications for assisting homeless individuals with drug problems. There is controversy regarding the effectiveness of adopting a "housing first" approach versus interventions that require substance abuse treatment and/or abstinence from drug use as a pathway to obtaining stable housing .
As expected, drug problems were associated with significantly poorer mental health status. The lack of association between drug problems and physical health status is surprising given the many physical health complications related to drug use [4, 5]. This finding may be explained by a long latency period between the initiation of drug use and deterioration in physical health, survival bias due to the death of drug users with poor physical health, or selection bias due to the exclusion of hospitalized individuals. Nonetheless, our findings suggest that service providers should recognize that mental health, rather than physical health, may represent the greatest challenge for homeless individuals with drug problems.
This study has a number of strengths. A large stratified random sample of homeless men and women, including both shelter and non-shelter users, were recruited across numerous community sites in a major North American city. The recruitment rate of eligible individuals was very high at 73%. Also, the ASI instrument utilized in this study has been previously validated among homeless people [12–14].
This study has a few limitations. There was no control group of non-homeless individuals. The exclusion of individuals who were unable to give informed consent may have resulted in the elimination of people who were under the influence of drugs. The requirement that study participants have a health insurance number resulted in the exclusion of refugees and refugee claimants. Lastly, drug use was assessed on the basis of self-report. While this method is subject to recall errors and social desirability bias, it has been validated as a fairly accurate measure among homeless individuals .