[The validity of infant mortality statistics in Navarra].

In order to explore the validity of official infant mortality statistics, the number of infant deaths registered in the cohort of live births and resident in Navarra (Spain) from 1985 to 1989, according to data supplied by the Department of Statistics of Navarra, has been compared with that obtained in a retrospective search into the information of the hospitals. Among 24,383 live births of the cohort, 291 were certified as infant deaths. An over-register of 13 cases (4.5%) and an under-register of 26 cases (8.9%) was detected. By excluding those live births with less than six months of gestation, as done by the National Institute of Statistics, the validity is altered, because apart from the errors mentioned, one must add those resulting from the incorrect filling-in of the "number of weeks pregnant"= variable. When correcting these deficiencies, the final statistics (Infant Mortality Risk = 11.17 per 1000 live births) did not differ much from the official figures (IMR = 11.48 per 1000 live births).