OBJECTIVE To estimate the prevalence of diabetes in the Maltese Islands from data regarding drug consumption, and to assess the prescribing habits and management attitudes of local doctors. METHODOLOGY The prescribed average daily dose of each type of anti-diabetic drug was calculated from random representative samples of patients (total 1,444) attending the Health Department's primary and secondary health centres as well as of patients seeing their private practitioners. This constituted circa 16% of the total estimated number of diabetics on drug therapy in Malta. The quantities of drugs consumed were obtained both from Government and private industry sources, whilst relevant data on therapy modalities, including proportions of patients on a combination of anti-diabetic preparations, were obtained from records held at the Dept. of Health. RESULTS Applying the appropriate formulae (as proposed by Eurodiab. subarea C 1989 programme) the global prevalence of diabetes worked out to be 5.22%, an underestimate of circa 12% of the figure extrapolated from the recent population based epidemiological studies. Overall 51% of cases were treated with 'diet only', whilst less than 10% of subjects were managed with insulin. Compared to patients attending the better staffed and equipped health centres, those seeing their private physicians were proportionately less frequently put on insulin (5.7% vs 11.7%), and were prescribed a slightly higher mean daily dose (1.7 vs. 1.4 tab) of glibenclamide suggesting the need of better education of the patient (and probably also of the physician). CONCLUSION This type of study proved to be relatively simple and inexpensive to perform. Although it cannot replace epidemiological field surveys it can still give a reasonably fair (albeit with limitations) estimate of diabetes prevalence in a population. Moreover it permits an important insight into local public health indicating ways to improve diabetes care.