Table 4.

Previous studies on MS prevalence in France

Author	Year of the study	Data sources and methodology	MS prevalence (per 100,000 inhabitants) with its 95% confidence interval, when available	Strengths and limits
Vukusic [4]	2003	Identification of MS patients using the LTD diagnosis in the French agricultural workers health insurance system database. Prevalence standardized using the age structure of the French population	Age-standardized prevalence: 65.0 (62.5–67.5)	Strengths: nationwide estimation Limits: lack of representativeness. Only 7% of the French population is covered by the agricultural workers health insurance system. LTD status for MS may not be sufficiently sensitive to identify all MS cases
Fromont [3]	2004	Identification of MS patients using the LTD diagnosis in the General Scheme health insurance system database. Prevalence standardized using the age structure of the French population	Age-standardized prevalence: 94.7 (94.3–95.1) Higher rates in North-Eastern versus South-Western regions	Strengths: nationwide estimation, population more representative of the French population than in Vukusic paper Limits: LTD status for MS may not be sufficiently sensitive to identify all MS cases
Sagnes-Raffy [6]	2005	Identification of MS patients by matching several data sources by a capture-recapture method in Haute Garonne, a department in the South West of France. The data were hospital data, LTD status and reimbursement of specific treatments of MS using local health insurance data, and data from a MS health network	Modelled prevalence using the capture-recapture method: 138–149	Strengths: multisource and independent data collection Limits: local estimation
El Adssi [5]	2008	Identification of MS patients by matching several data sources by a capture-recapture method in Lorraine, a region in the North East of France. The data were hospital data, LTD status and reimbursement of specific treatments of MS using local health insurance data, and data from the Lorraine registry of MS	Crude prevalence: 170.9 (165.7–176.3) Modelled prevalence using the capture-recapture method: 188.2 (182.7–193.8)	Strengths: multisource and independent data collection Limits: local estimation