Comparison of timeline follow-back self-report and oral fluid testing to detect substance use in adult primary care patients

•Marijuana use had higher detection using Timeline follow-back self-report.•Cocaine, prescription opioids, and heroin had higher detection using oral fluid testing.•Triangulating multiple sources of information may improve detection of drug use. Timeline Follow-back (TLFB) interviews using self-report are often used to assess substance use. Oral fluid testing (OFT) offers an objective measure of substance use. There are limited data on the agreement between TLFB and OFT. In this secondary analysis from a multisite study in five primary care sites, self-reported TLFB and OFT data collected under confidential conditions were compared to assess concordance (N=1799). OFT samples were analyzed for marijuana, heroin, cocaine, and non-medical use of prescription opioids. Demographic differences in discordance relative to TLFB and OFT concordant results for marijuana, the only substance with an adequate sample size in this analysis, were examined using multinomial logistic regression. Overall concordance rates between TLFB and OFT were 94.9 % or higher for each substance, driven by large subgroups with no use. Among participants with discordant use, marijuana was the only substance with lower detection on OFT than self-report (27.6 % OFT-positive only vs 32.2 % TLFB-positive only), whereas cocaine (65.6 % vs 8.6 %), prescription opioids (90.4 % vs 6.0 %), and heroin (40.7 % vs 26.0 %) all had higher detection via OFT than TLFB. Participants who reported marijuana use but had a negative OFT were more likely to be younger, Hispanic, and White compared to those with TLFB and OFT concordant positive results. TLFB and OFT show disparate detection of different substances. Researchers should consider the implications of using either self-report or oral fluid testing in isolation, depending on the substance and collection setting. Triangulating multiple sources of information may improve detection of drug use.