This manuscript reviews strengths and weaknesses of common design options for managing non-study medications in randomized clinical trials of behavioral treatments for drug abusers. While the principal focus of these trials is on the relative efficacy of different behavioral interventions, drug abusers entering these trials are often either using or in need of prescribed psychotropic medications, which can have a major independent impact on treatment outcome. Options reviewed range from tight to minimal restriction on use of ancillary medications. Generally, major tradeoffs pertain to the balance between internal validity and generalizability. However, experimental rigor can be maintained while allowing subjects to use ancillary medications if steps are taken to monitor and minimize the variability in treatment delivery.