RESUMEN
This study evaluates the prevalence of diabetes mellitus (DM) in Pisa (Tuscany, Italy) using four independent data sources. The main source, represented by computerized prescriptions for anti-diabetic agents collected over a 4-month period, was validated using three secondary sources: (a) the list of diabetic patients who receive material of self-care from the National Health Service; (b) the clinical records of diabetic patients obtained from a random sample of family doctors; (c) the clinical records of diabetic patients attending our outpatient clinic. The main source provided 3806 patients, and 697 patients were added from the secondary sources, thus identifying a total number of 4503. The prevalence of known DM in the "Pisa area" exclusively reckoned by the main source, was 2.01%, and the prevalence corrected by the addition of the various sources resulted in 2.4%. The capture-recapture method showed a completeness of ascertainment of the survey of 90.1%, and thus an estimated prevalence of known diabetes of 2.64%. Of these, 141 patients had insulin-dependent diabetes mellitus (IDDM) corresponding to 3.2% of identified diabetic subjects (prevalence 0.07% inhabitants); 4362 patients had non-insulin-dependent diabetes mellitus (NIDDM), 96.8% of identified diabetic subjects (prevalence 2.36%). Of patients with NIDDM 10.5% was treated by diet, 65% with oral hypoglycaemic agents (OHA), 23% with insulin and 1.5% with insulin plus OHA. This study shows that the method used in this survey is suitable for epidemiological studies because it does not demand the cooperation of the diabetic patients, is addressed to the entire diabetic population without age discrimination and singles out the diabetic population in a very reliable way.