Objective: To establish the actual number of maternal deaths in The Netherlands by determining the degree of underreporting.
Methods: We conducted a nationwide, retrospective cross-check of the three available maternal mortality registration systems and issued a questionnaire to senior obstetricians in all hospitals during the years 1983-1992.
Results: The officially reported maternal mortality rate during the study period was 7.1 per 100,000 live births (133 maternal deaths per 1,862,985 live births). After completion of the study, our data indicate that the rate should be at least 9.7 per 100,000 live births (180 maternal deaths). Early pregnancy and indirect deaths were more likely to be underreported than direct deaths during labor and the puerperium. Failure to register the recent pregnancy on the death certificate was a frequent problem. Misclassification was particularly evident for cerebrovascular disorders, cardiovascular disorders, and eclampsia.
Conclusion: The level of underreporting of maternal mortality in The Netherlands was estimated at 26%. The pregnancy status of women should be registered on death certificates. Officially reported maternal mortality rates are unreliable and international comparisons using these data thus are less meaningful.