Use of Rapid Ascertainment Process for Institutional Deaths (RAPID) to identify pregnancy-related deaths in tertiary-care obstetric hospitals in three departments in Haiti

BACKGROUND: Accurate assessment of maternal deaths is difficult in countries lacking standardized data sources for their review. As a first step to investigate suspected maternal deaths, WHO suggests surveillance of “pregnancy-related deaths”, defined as deaths of women while pregnant or within 42 days of termination of pregnancy, irrespective of cause. Rapid Ascertainment Process for Institutional Deaths (RAPID), a surveillance tool, retrospectively identifies pregnancy-related deaths occurring in health facilities that may be missed by routine surveillance to assess gaps in reporting these deaths. METHODS: We used RAPID to review pregnancy-related deaths in six tertiary obstetric care facilities in three departments in Haiti. We reviewed registers and medical dossiers of deaths among women of reproductive age occurring in 2014 and 2015 from all wards, along with any additional available dossiers of deaths not appearing in registers, to capture pregnancy status, suspected cause of death, and timing of death in relation to the pregnancy. We used capture-recapture analyses to estimate the true number of in-hospital pregnancy-related deaths in these facilities. RESULTS: Among 373 deaths of women of reproductive age, we found 111 pregnancy-related deaths, 25.2% more than were reported through routine surveillance, and 22.5% of which were misclassified as non-pregnancy-related. Hemorrhage (27.0%) and hypertensive disorders (18.0%) were the most common categories of suspected causes of death, and deaths after termination of pregnancy were statistically significantly more common than deaths during pregnancy or delivery. Data were missing at multiple levels: 210 deaths had an undetermined pregnancy status, 48.7% of pregnancy-related deaths lacked specific information about timing of death in relation to the pregnancy, and capture-recapture analyses in three hospitals suggested that approximately one-quarter of pregnancy-related deaths were not captured by RAPID or routine surveillance. CONCLUSIONS: Across six tertiary obstetric care facilities in Haiti, RAPID identified unreported pregnancy-related deaths, and showed that missing data was a widespread problem. RAPID is a useful tool to more completely identify facility-based pregnancy-related deaths, but its repeated use would require a concomitant effort to systematically improve documentation of clinical findings in medical records. Limitations of RAPID demonstrate the need to use it alongside other tools to more accurately measure and address maternal mortality.