Cannabis use: consistency and validity of self-report, on-site urine testing and laboratory testing.

Aims: To evaluate the agreement between adolescent self-reported cannabis use, "on-site" qualitative urine screening, and quantitative laboratory testing.

Design: A cross-sectional study of intake and follow-up data from 248 adolescents entering substance abuse treatment for cannabis use disorders (abuse or dependence). This is part of the multi-site cooperative agreement Cannabis Youth Treatment study.

Setting: Data collected from adolescents randomly assigned to one of five outpatient treatments at four sites: Operation PAR, Inc., Florida; Chestnut Health Systems, Illinois; University of Connecticut Health Center, Connecticut; and Children's Hospital of Philadelphia, Pennsylvania.

Participants: The data represent 248 unique individuals from a sample of 297 adolescents ranging in age from 12 to 18 years.

Measurements: Prevalence, agreement, kappa, sensitivity, specificity, positive and negative predictive value.

Findings: The self-report rates were higher at intake than either urine test (82.4% vs. 77.0% vs. 52.7%), but both lower and higher at the 3-month follow-up (55.5% vs. 70.0% vs. 47.3%) and 6-month follow-up (60.2% vs. 73.5% vs. 55.8%). The disagreements went in both directions and the kappa coefficients were only in the moderate range (0.4). Over two-thirds of these frequent cannabis users tested positive when they said they had not used in 1 week and one-third tested positive even though they said it had been more than 4 weeks since last use.

Conclusions: The findings suggest both the advantages of multiple sources of information and the need for further work on the latency of cannabis metabolites in clinical populations.