Use of multiple data sources to improve maternal death ascertainment

ABSTRACT

This study was designed to demonstrate how multiple public health data sources can be used to improve the surveillance of our pregnant population from conception to one year postpartum. South Carolina's vital record registry was reviewed to obtain maternal deaths recorded form 1992 to 1996. Multiple data sources associated with pregnancy were then computer linked to death certificates of women between the ages of 10 to 50 years, for 1992 to 1996. These death certificates (over 6 000) were manually reviewed for written notation of pregnancy. Coroners' records are being reviewed to identify cases that have indications of pregnancy recorded. These provide us with cases of pregnancy associated (PA) deaths that will be analyzed to confirm cause of death and its relationship to pregnancy. Preliminary analysis of available data shows an 80 percent increase in our maternal mortality ratio for the period 1992 to 1996. Using our multiple data sources we were able to identify additional deaths in both prepartum and postpartum periods. Our surveillance identified deaths occurring from as early as 13 weeks of pregnancy to one year postpartum. No individual data source linkage was found to be more efficient than others. PREMMS is a complementary system where one data source complements the inadequacy of another. Using multiple public health data source enabled us to improve the surveillance of maternal deaths and enhance our ability to report mortality more accurately. Ultimately this will improve our ability to detect trends, identify risk factors, high risk individuals and avoidablity of death, and to develop intervention policies aimed at decreasing maternal mortality.(AU)