I've seen similar large errors in statistics on infant mortality. Different countries define it very differently, for example a different line between a miscarriage and infant mortality.