PNS272 European Health Claims Databases in the Era of Real-World Evidence: Where Are WE in France and Germany?

Abstract

Based on the experience accrued since their availability for the conduct of pharmacoepidemiology and pharmacoeconomic studies, this work summarizes and compares data availability, challenges and opportunities of the French SNDS (‘Système National des Données de Santé’) and the German statutory health insurance GKV (‘Gesetzliche Krankenversicherung’). The SNDS links national hospital databases, health insurance and other nationwide databases. The GKV is administered by >100 different sickness funds, that can be further grouped by types of sickness funds, as large representative collectives of insured persons of the GKV. Both databases offer relative completeness in terms of diagnoses, prescriptions reimbursed, acts and procedures performed, as well as cost data in both inpatient and outpatient settings, yet with a few specific gaps. Population coverage is nearly total for SNDS (>99%) and about 4-10% according to the sickness funds considered for GKV; time coverage can be >10 years in some cases. After an overview of the SNDS and GKV databases, case-based comparisons are provided. Perspectives are highlighted based on on-going developments addressing some of the above-mentioned gaps. This work highlights strengths, weaknesses and opportunities to leverage two major European claims databases for pharmacoepidemiology and Health Economics and Outcomes Research (HEOR). Case studies for cost-effectiveness and clinical decision-making demonstrated high relevance of the two databases, for a range of study objectives. Sensitivity and specificity of patients identifications and outcomes definitions reach >90% in the cases explored for SNDS, which appears to be a powerful database to generate robust real-world evidence, especially when linked to a disease-specific registry. Disease severity and predisposing factors for some specific conditions were also accurately defined using GKV data. SNDS and GKV subsets are large, rich and reliable real-world databases that can offer a range of different potential applications to HEOR and clinical research, provided careful implementation and deep understanding of limitations.