Three basic types of measures are important for the estimation of the extent and consequences of drug abuse: the number of persons involved, the number and nature of episodes of use, and the volume of drugs consumed. In the case of numbers of persons, the types of users should also be estimated. Current data systems have been criticized on several grounds including coverage, consistency, and comparability. With limited resources available for improved measurement, priority should be given to the comparability of data across sources, time, and geographical areas. Comparability requires that definitions and inclusion criteria be consistent so that the extent of overlap between data sources can be estimated. The highest priority for any new data series should be the enhancement or augmentation of existing drug abuse indicators. The proposed nationwide minimum data set to be used for drug treatment admissions and discharges extends admission data collected and reported by some States to be more nationally representative.