Abstract: Problems in the use of urinalysis results in metanalysis of RCTs on drug Dependence.The most common outcomes reported in drug treatment studies are self-reported drug use and use of drugs based on urine analysis for drug metabolites. In these studies, urine samples are usually collected regularly over time for each patient, and are best understood as multiple binary (dichotomous) outcomes. Most of these studies summarise this information by reporting for each treatment group the percentage of positive urine tests, obtained by dividing the number of positive urines by the number of patients. This information not taking into account the fact that some patients can have different number of samples collected can lead to misleading conclusions if treated an independent observations. Additionally, the comparison of rates of recurrent events requires consideration of the tendency for some individualsto have greater propensities for recurrent events than others. The aim of this presentation is to discuss possible approaches to the use and treatment of urinalysis results. The possible options seem to be the followings: a. consider urinalysis as dichotomous variable (i.e. having at least two positive tests); b. consider urinalysis as continuous variable (calculate mean number of positive tests and standard deviation for each group); and c. have new software that can allow for the meta-analysis of proportional data. Such information is usually not present in the published paper. Access to original data is probably needed to address this issue properly in the analysis. The feasibility of this approach is subject to ongoing consideration. The minimum result would be indication for proper reporting of urinalysis results in future publications, and it may prove necessary to have software written in RevMan to deal with such multiple binary outcomes.