[The Piedmont Diabetes Register: opportunities and limits for epidemiological purposes].

Diabetes has a heavy impact on both population health and expenditure for health care. Epidemiological surveillance of the disease and monitoring of the health care provided to people with diabetes are priorities of the Italian Ministry of Health. A regional diabetes register (PDR) has been operating in Piedmont since 1989. The aim of this study is to evaluate its reliability for epidemiological purposes. We selected two populations of persons with diabetes and linked them with the PDR. The first population consisted of 123549 residents of Piedmont discharged with a diagnosis of diabetes (ICD-9 CM250**) between 01.01.01995 and 31.12.2002, and the second consisted of 122394 persons who had received at least two prescriptions of an antidiabetic drug (ATCA10**) in 2003. Of these two groups of patients, 66% of those with a hospital discharge and 73% of those with a drug prescription were found among the 183301 patients in the PDR. A logistic model was adapted to explore the independent roles of variables available in the two databases as determinants of registration in the PDR. The odds ratios were significantly higher for patients under 65 years of age, women, patients with a low educational level, Italians, married and divorced patients and those with conditions affecting the severity of the disease. There were wide geographic variations: registration in the PDR was more frequent for residents of metropolitan Turin than for residents of other parts of Piedmont. The PDR is therefore not representative of the entire population of persons with diabetes in Piedmont Region.