Intercommunity variations in antidiabetic drug utilization and in prevalence of diabetes mellitus

Abstract

In order to assess whether differences in the extent of antidiabetic drug utilization may be an indicator of variations in diabetes prevalence (Type 1 plus Type 2, including non-pharmacologically treated diabetics), intercommunity variations of antidiabetic drug utilization (insulin plus oral agents) in Sweden were examined through sales expressed in defined daily doses per 1000 inhabitants per day and compared with intercommunity prevalence variations. There were large (two-fold) and long-lasting (>10 years) between-county differences in antidiabetic drug sales. Low-sales counties were found both in the doctor-dense south and the doctor-sparse north of Sweden, which argues against the assumption that the sales variation simply was due to different access to medical care. Recent diabetes prevalence figures have been obtained for eight municipalities in different parts of Sweden. In these, antidiabetic drug sales and diabetes prevalence (including non-pharmacologically treated diabetics) showed a close correlation (r = 0.87; p < 0.01), supporting the assumption that variations in antidiabetic drug utilization may indicate variations in diabetes prevalence, at least in countries that, like Sweden, have uniform health care systems. The findings also suggest that, in Sweden, there is little intercommunity variation in the proportion of non-pharmacologically treated diabetics. On the other hand, it is likely that there are many undetected diabetics in areas where no diabetes screening has been performed. Hence, although differences in antidiabetic drug sales may help to indicate where to look for differences in diabetes prevalence, drug sales data are no substitutes for prevalence studies.