155:oral Monitoring the impact of health system strengthening for maternal and child health in Guinea-Bissau: focus on universal health coverage removes focus from stagnating perinatal mortality

ObjectiveTo investigate coverage of antenatal care (ANC) and facility births and perinatal mortality before and during the stepwise implementation of the ‘Integrated Programme for the Reduction of Maternal and Child Mortality’ (PIMI), a health system strengthening programme which included free care, health worker training and infrastructure rehabilitation in Guinea-Bissau.MethodsWe used data from Bandim Health Project’s rural health and demographic surveillance system from three 24-months birth cohorts: pre-PIMI (2011-13), during PIMI’s pilot phase (2014-16) and its nation-wide full-scale implementation (2017-19); and two areas: pilot regions (PIMI since 2013) and scale-up regions (PIMI since 2017). Using generalized estimating equations, we compared service coverage (first/fourth ANC consultation (ANC1/4) and facility births) and perinatal mortality over time and across areas. We also assessed associations between perinatal mortality and cluster-level ANC4 and facility birth coverage.ResultsAcross the three cohorts, 23,828 births were included. Pre-PIMI, approx. 1/3 women obtained ANC4 and facility birth in both areas. ANC4 and facility birth coverage increased to approx. 1/2 in both areas. Relative increases were largest in the scale-up area for ANC4 (p=0.007 for same development), and comparable across areas for facility births (p=0.16). Perinatal mortality was around 8% pre-PIMI and did not decline over time. Higher cluster-level ANC4 (both areas) and facility birth coverage (pilot area) were associated with a tendency towards lower perinatal mortality pre-PIMI, but this association disappeared over time.ConclusionWhile universal access to quality maternal and child health services is considered essential to improve maternal-perinatal survival, increases in ANC and facility birth coverage did not translate into reduced perinatal mortality. Hence, measures of health outcomes cannot be replaced by measures of service coverage. Rigorous evaluation designs are required to assess the real-life effects of policies aiming to improve survival and identify the potential causes of (absence of) effects.