Between 1990 and 1999, data collection was conducted in Houston for 14 consecutive days on a quarterly basis. During each day of data collection, field staff obtained a list of arrestees who had been in custody for no more than 48 hours. For those drugs the arrestees reported having ever tried, they were asked to indicate age of first use, whether they had used the drug within the past 12 months, the number of times used within the past 30 days, and whether they had used the drug within the past 3 days. Participants who admitted to drug use were also asked whether or not they considered themselves drug-dependent, as well as whether they were under the influence or in need of drugs at the time of arrest. Several questions also focused on treatment. In addition to the self-report data, a urine sample was obtained. The Enzyme Multiplied Immunoassay Test (EMIT) screened for 10 drugs. Data analysis was performed in two phases for each year between 1990 and 1999. First, descriptive statistics were calculated. Then kappa statistics were computed to explore the relationship between urinalysis results for cocaine and opiates and self-reported 3-day cocaine and opiate use. Kappa statistics indicated that the correlations between self-reported 3-day cocaine and opiate use and their urinalysis counterparts were consistent over this 1-year period. The range of kappa statistics between 1990 and 1999 was only .16 for cocaine and .12 for opiates. These data suggest that urinalysis may be unnecessary, because the correlations between the two drug-use measures were remarkably similar over time. This would result in a significant cost savings without compromising the accuracy of the data obtained from self-reports. 2 tables and 16 references