The objective of this study is to determine whether the presence of a drug and alcohol unit and comprehensive medical education have made a difference to the detection and management of alcohol-related problems in a general teaching hospital. Data were obtained from hospital case notes before and after the introduction of the drug and alcohol unit and medical education. Samples of general admissions and admissions with an alcohol diagnosis were obtained. The outcome measures included alcohol history taking, quantification of alcohol consumption and management of alcohol dependence. It was found that for the general admission sample, interns were more likely to take an alcohol history, both in the accident and emergency (A&E) department and on the ward, in 1994 compared to 1988. The difference in the A&E department was statistically significant (P = 0.009). In the A&E department in 1994 there was a greater prevalence of alcohol history taking for male patients (odds ratio (OR) 10.09, 95% confidence interval (CI) 1.89 to 53.70, P = 0.007) compared with female patients (OR = 1.81, 95% CI 0.38 to 8.57, P = 0.045). There were no differences in alcohol history taking by interns in the samples of alcohol-related admissions. There were no statistically significant differences in the prevalence of documenting alcohol histories quantitatively in either sample. The use of alcohol withdrawal charts, ordering diazepam for alcohol withdrawal and ordering drugs which conformed to the hospital formulary all increased significantly by 1994. The prevalence of ordering thiamine and consulting the drug and alcohol unit both decreased slightly by 1994, but not significantly. Improvements have been found in the detection and management of alcohol use since the introduction of the Drug and Alcohol Unit and medical education, but there is still room for further improvement and particular areas where greater attention is needed are identified.