Synthetic estimation methods include the population projection method, which extrapolates from areas where prevalence is known to other areas, and the principal component method, which uses relationships observed among multiple indicators to obtain a single indicator of drug use. Small area estimations project risk estimates developed from small area surveys to larger areas, while capture-recapture methods provide reliable estimates only if the underlying assumptions are met. In making estimates of the number of hard-core drug users, researchers should understand data limitations and changes in drug use patterns. The challenge for future research is to refine each method and then to combine the strengths of each to produce the best estimates possible. 93 references and 1 table