Methods: Systematic screening selected 300 drug users and 100 non-users from a variety of public settings frequented by homeless people (shelters, day centers, and homeless rehabilitation programs) as well as from streets. The Composite International Diagnostic Interview/Substance Abuse Module (CIDI/SAM) provided detailed assessment of drug use and criterion symptoms for substance use disorders. Urine samples were collected at the time of interview and tested for alcohol, cannabis, cocaine, morphine (heroin), and amphetamine. Kappas and Yule’s Y tests examined concordance rates (corrected for chance agreement) and logistic regressions assessed factors associated with discordance between self-report and urine testing.
Results: Analyses demonstrated fair-to-good reliability between self-report methods of ascertaining use (screener vs. CIDI/SAM), but fair-to-poor reliability between self-report (screener) and urine testing. Approximately two-thirds of subjects with urine positive for drugs denied current drug use. Factors associated with denial of substance use despite a positive urine test were increased age, younger age at onset of homelessness, and greater number of years homeless. In detecting current drug users, self-report was only 48% sensitive, and urine screen was 84% sensitive. Adding urine testing to self-report doubled the yield of cases found; whereas, adding self-report to urine testing added little to identify users.
Implications for practice: These findings demonstrate the importance of urine screening in constructing the complete story of drug use for homeless individuals. Incomplete or misleading information by self-report alone on substance use can have grave implications for providing appropriate and effective services for this difficult-to-treat population.