What's the agreement between self-reported and biochemical verification of drug use? A look at permanent supportive housing residents

Abstract

Self-reported substance use is commonly used as an outcome measure in treatment research. We evaluated the validity of self-reported drug use in a sample of 334 adults with mental health problems who were residing in supportive housing programs. The primary analysis was the calculation of the positive predictive values (PPVs) of self-report compared to an oral fluid test taken at the same time. A sensitivity analysis compared the positive predictive values of two self-reported drug use histories: biological testing window (ranging between the past 96h to 30days depending on drug type) or the full past 90-day comparison window (maximum length recorded during interview). A multivariable logistic regression was used to predict discordance between self-report and the drug test for users. Self-reported drug use and oral fluid drug tests were compared to determine the positive predictive value for amphetamines/methamphetamines/PCP (47.1% agreement), cocaine (43.8% agreement), and marijuana (69.7% agreement) drug tests. Participants who misreported their drug use were more likely to be older, non-White, have no medical insurance, and not report any alcohol use. In general, amphetamine/methamphetamine/PCP and cocaine use was adequately captured by the biological test, while marijuana use was best captured by a combination of self-report and biological data. Using the full past 90day comparison window resulted in higher concordance with the oral fluid drug test, indicating that self-reported drug use in the past 90days may be a proxy for drug use within the biological testing window. Self-report has some disadvantages when used as the sole measure of drug use in this population.