A previous review of 29 randomized control trials of alcohol withdrawal demonstrated that despite the poor quality of the trials benzodiazepines are probably the “drugs of choice” and have been since 1967. In addition, it was confirmed by 1967 that active drugs are more effective than placebo. To assess the impact of these studies on the practice of medicine, 364 physicians who treat withdrawal regularly were surveyed and 52 review articles recommending a drug or drug category were studied. Eighty-five percent of the 202 respondents use at least one benzodiazepine. Of these, 51% use only benzodiazepines whereas 44% also use other drugs. Chlordiazepoxide and diazepam are employed most frequently. Four percent of the physicians use no drugs. Factors deemed most useful for determining treatment were personal observation (80%) and several good clinical trials (62%). Considered least important were departmental guidelines (53%), psychiatric textbooks (50%), and residency or fellowship training (45%). Of the review articles 55% recommend benzodiazepines and 21% anticonvulsants; of those published after 1977, 77% recommend benzodiazepines. Overall, 63% of the review articles base their recommendations on clinical trials. Thus, it has taken a full 10 years for the review articles to “catch up” with the results of the randomized control trials. Similarly, at least 13% of the physicians queried still do not use benzodiazepines to treat alcohol withdrawal. Better treatment of alcohol withdrawal will require both better clinical trials and better dissemination of the results.