Utilization of drug sales data for the epidemiology of chronic diseases: the example of diabetes

An indirect method for estimating the prevalence rates of chronic diseases that are treated by specific drugs was proposed in 1988 to European countries in the framework of a European Community concerted action on diabetes epidemiology. Data on consumption of antidiabetic drugs were collected at the national level in nine countries and at a regional level in two. Using official drug sales data and recent demographic data, we estimated the diabetes prevalence rates in each country or region. The estimated diabetes prevalence in Europe varied from 1.6% in Northern Ireland to 4.7% in Malta. In four countries that already had diabetes prevalence data, the estimation through drug consumption provided figures 3-20% lower than those from field surveys. This study showed a large variety of prescribing habits for diabetic patients in Europe (for example, the proportion of insulin-treated patients varies from 13% to 36%) and underscores the need for a consensus on antidiabetic treatments based on valid clinical research. The proposed approach does not replace field surveys but provides an inexpensive and practical marker of disease frequency and therapeutic attitudes over space and time.