Validity of ACASI self-reports of drug use with on-site urine testing in the general population of Puerto Rico

The extent to which individuals are willing to disclose their use of illicit drugs in general population surveys is still unclear; this is particularly true both for minority populations such as Hispanics and in relation to potential under-reporting of stigmatized behaviors using ACASI. As part of a household survey of hepatitis C in the general population of Puerto Rico, self-reports of illicit drug use were compared with on-site urine toxicology testing. Information on recent use (last three days) and past year and lifetime use was collected using ACASI. On-site urine testing for drug metabolites to detect recent marihuana (last 10 days), cocaine (last 2-4 days) and opiate (last 2-3 days) use was performed prior to the interview. Sensitivity and specificity of self-reports, using the urine test as the standard, were estimated in a sub-sample of 520 adults aged 21-64 years. Prevalence of self-reported recent use was 2.9% for marihuana, 3.1% for cocaine, and 0.8% for heroin. Prevalence based on urine tests was 3.9% for marihuana and cocaine and 1.0% for heroin. Sensitivity for self-reports of recent use was 60% for marihuana, 70% for cocaine and 40% for heroin. Sensitivity increased with reports of use in more remote time periods. Specificity for self-reports of recent use for all drugs was 99.2% or higher, and it decreased with reports of use in more remote time periods. These results suggest that disclosure of marihuana and cocaine lifetime use using ACASI in the general population of Puerto Rico is higher than that for heroin.