[Quality of perinatal statistics from hospital discharge data: comparison with civil registration and the 2010 National Perinatal Survey].

OBJECTIVE:To compare hospital discharge data (PMSI) with data in the reference databases: vital statistics and National Perinatal Surveys (NPS) for the principal perinatal indicators. METHODS:Data concerning hospitalizations for delivery and childbirth were extracted from the PMSI 2010 database. The exhaustiveness was assessed by comparing discharge data with data from birth certificates. Indicators were compared with those in the 2010 NPS, which was based on a representative sample of births (n=15,000), using 95% confidence intervals. RESULTS:About 823,360 hospital abstracts with delivery and 829,351 hospital abstracts with live births were considered. The exhaustiveness of the PMSI was 99.6% for live births in Metropolitan France. The distribution of maternal age, mode of delivery, birth weight and gestational age in the PMSI and NPS were very similar. In Metropolitan France, the prematurity rate was 6.9% (PMSI) vs. 6.6% [6.2-7.0] (NPS) and the rate of caesarean was 20.6% vs. 20.4% [19.8-21.1]. There were marked differences for the percentage of birth weights<2500g and for maternal diseases. CONCLUSION:The routine use of the PMSI for some indicators for follow-up purposes is foreseeable. Validation studies are still necessary for maternal diseases, for which recording is less standardized.