Reliability and validity of drug users' self-reports of amphetamine use

Background: Methamphetamine use has increased in recent years and is associated with HIV and STD infection. Little research has addressed the reliability and validity of drug users' self-reports of amphetamine use.

Method: Participants were drug users not currently in treatment who completed items assessing amphetamine use on the Risk Behavior Assessment. Participants reported lifetime use of amphetamines (ever/never), number of days and times used in the past 30 days, and use in the past 48 hours. To evaluate validity, 4037 participants provided urine samples that were compared to self-reports of amphetamine use in the past 48 hours. To evaluate reliability, 218 participants completed the RBA at two time points, 48 hours apart.

Results: 2258/4037=55.9% reported having ever used amphetamines and 254/4037=6.3% having used it in the past 48 hours. Self-reports of use in the past 48 hours demonstrated moderately high validity, with an 95% accuracy rate (kappa =.56). Test-retest data indicated good reliability for self-reports of ever having used amphetamines (kappa =.79), number of days having used in the last 30 days (r=.88, p<.001), number of days having injected amphetamines in the last 30 days (r=.91, p<.001), times injected amphetamines in the last 30 days (r=.90, p<.001), and number of days having used without injecting in the last 30 days (r=.75, p<.001).

Discussion: Drug users provided reliable and accurate self-reports of amphetamine use. Reliable and valid measures are essential for describing and predicting trends in amphetamine use, evaluating the effectiveness of interventions targeting amphetamine use, and developing policies and programs.