Almost all such data come from clinical observation and testing of cocaine abusers seeking treatment. These data are subject to several confounding factors that preclude drawing definitive conclusions about the pathophysiological mechanism for the observed abnormality, its prevalence and prognosis, and individual differences in susceptibility. Confounding factors include the unknown purity of cocaine used, uncertainty as to quantity and duration of cocaine exposure, effects of other substances frequently used by cocaine users, abnormalities associated with drug route of administration and the drug abuse lifestyle, and selection bias. Future studies should use more sophisticated research designs to address these issues. Such studies should include prospective, longitudinal followup of large samples of cocaine users more representative of age, ethnicity, gender, and variations in use patterns, including subjects not in medical or drug abuse treatment. Studies should also use case-control studies that compare cocaine users with a variety of appropriate comparison groups. Experimental subjects should use subjects whose cocaine exposure is known precisely, due to the cocaine's being administered under experimenter control. The latter methodology will be crucial in the exploration of mechanisms by which cocaine produces pathophysiological abnormalities and in the development of treatments for these. 52 references