Background: Problems with self-reported drug use include difficulties with recall and recognition as well as the desire to respond to questions in a socially desirable manner. Various methods have been developed to improve and/or validate estimates based on direct questioning of individuals regarding their substance use. For this study, we were interested in validating self-reported use of: 1) tobacco, 2) marijuana, and 3) other substances (i.e., heroin, cocaine, opiates, oxycodone, benzodiazepines, methamphetamine, phencyclidine, and barbiturates) employing urinalysis among inmates who participated in a randomized controlled trial of a smoking abstinence intervention in a tobacco-free prison located in the northeastern United States.

Methods: Two-hundred and seven men and women with a mean age of 34.9 (standard deviation = 9.0) completed questions regarding their substance use on a 7-day Timeline Follow-Back and provided urine specimens three weeks following prison release.

Results: Self-reported tobacco and marijuana use were highly consistent with urine drug testing in terms of overall agreement and Kappa (93.7% and.804 for tobacco, respectively; and 90.3% and.804 for marijuana, respectively); however, consistency was much lower for other drug use grouped together (62.7% and.270).

Discussion: Although some former inmates may not accurately report substance use, our findings indicate that they are in the minority, suggesting that self-report is valid for tobacco and marijuana use but much less so for other drugs grouped together. Future research should be conducted with a larger and more diverse sample of former inmates to establish the generalizability of our findings from this study.

Jacob J. van den Berg, Samuel Adeyemo, Mary B. Roberts, Beth C. Bock, L.A.R. Stein, Rosemarie A. Martin, Donna R. Parker & Jennifer G. Clarke
Substance Use & Misuse, 08 Feb 2018