ABSTRACT
INTRODUCTION: In France, the prevalence of multiple sclerosis is estimated between 65 and 125 patients per 100,000 inhabitants with a South-West towards North-East gradient. Nevertheless, the epidemiology of multiple sclerosis remains still imperfectly known, the recent studies being realized, either in a region of France, or from a single data source and thus suscepted not to be exhaustive. OBJECTIVE: Assessing the prevalence of the multiple sclerosis in 2005 in Haute-Garonne by matching several data sources completed by a capture-recapture method; estimating the exhaustivity of each of the sources. METHODS: The data sources were hospital data (DRG for the hospitalization, data of consultation), data of public health insurance system (main health insurance, agricultural health insurance, social welfare for self employed), and data from the MIPSEP network. The linkage was based on name, maiden name, first name, date of birth and sex and allowed a first estimation of the number of cases. Models of loglinear regression allowed estimating the total number of case and the sensitivity of each source. RESULTS: The total number of cases obtained by matching several sources of information amounted to 1549. The use of several data sources increased by 25.6 % the maximum number of patients identified with a single source of information (national health insurance, any insurance). According to the model used, the method of capture-recapture estimated the number of cases up to 1722. Therefore, this study estimated a prevalence of multiple sclerosis between 110 and 149 cases per 100,000 inhabitants in Haute-Garonne. CONCLUSION: The prevalence of the multiple sclerosis is largely underestimated in Haute-Garonne and questions the magnitude over the so-called gradient. Matching several sources of information is indispensable to improve collection of the total number of cases.
