The search for indicators for monitoring progress toward safe motherhood has prompted research into population-based measures of obstetric morbidity. One possible such measure is based on women's reports of their past childbirth experiences. In this prospective study in three hospitals in South Kalimantan, Indonesia, the accuracy of women's reporting of severe birth-related complications was examined. The findings of this study suggest that poor agreement exists between the way women report their experience of childbirth and the way doctors diagnose obstetric problems, although the degree of agreement varies with the type of complication. Questionnaires relying on women's experience of childbirth will tend to overestimate the prevalence of medically diagnosed obstetric problems such as those associated with excessive vaginal bleeding or dysfunctional labor. Questions suggestive of eclampsia may be more promising, although the small number of eclamptic women in this study precludes firm conclusions.