Effects of quality improvement in health facilities and community mobilization through women's groups on maternal, neonatal and perinatal mortality in three districts of Malawi: MaiKhanda, a cluster randomized controlled effectiveness trial

Abstract

Background: Maternal, perinatal and neonatal mortality remains high in low-income countries. We evaluated community and facility-based interventions to reduce deaths in three districts of Malawi.

Methods: We evaluated a rural participatory women's group community intervention (CI) and a quality improvement intervention at health centres (FI) via a two-by-two factorial cluster randomized controlled trial. Consenting pregnant women were followed-up to 2 months after birth using key informants. Primary outcomes were maternal, perinatal and neonatal mortality. Clusters were health centre catchment areas assigned using stratified computer-generated randomization. Following exclusions, including non-birthing facilities, 61 clusters were analysed: control (17 clusters, 4912 births), FI (15, 5335), CI (15, 5080) and FI + CI (14, 5249). This trial was registered as International Standard Randomised Controlled Trial [ISRCTN18073903]. Outcomes for 14,576 and 20,576 births were recorded during baseline (June 2007-September 2008) and intervention (October 2008-December 2010) periods.

Results: For control, FI, CI and FI + CI clusters neonatal mortality rates were 34.0, 28.3, 29.9 and 27.0 neonatal deaths per 1000 live births and perinatal mortality rates were 56.2, 55.1, 48.0 and 48.4 per 1000 births, during the intervention period. Adjusting for clustering and stratification, the neonatal mortality rate was 22% lower in FI + CI than control clusters (OR = 0.78, 95% CI 0.60-1.01), and the perinatal mortality rate was 16% lower in CI clusters (OR = 0.84, 95% CI 0.72-0.97). We did not observe any intervention effects on maternal mortality.

Conclusions: Despite implementation problems, a combined community and facility approach using participatory women's groups and quality improvement at health centres reduced newborn mortality in rural Malawi.

Keywords: Malawi; Maternal mortality; Neonatal mortality; Perinatal mortality; Quality improvement; Women's groups.

Figure 1

Map showing the central region of Malawi with the three districts where the randomized controlled trial took place shown in darker grey: Kasungu (top), Lilongwe (bottom) and Salima (right, by Lake Malawi). Key to randomized intervention areas: green: control clusters; red: quality improvement at health centres (MaiKhanda facilities intervention); blue: community women’s groups (MaiKhanda community intervention); purple: facility and community interventions together.

Figure 2

MaiKhanda Women’s Group Action Cycle followed by the women’s groups.

Figure 3

Overview of the facilities quality improvement intervention.

Figure 4

Randomization design flowchart. CEOC: comprehensive emergency obstetric care; disp.: dispensary; HC: health centre; PH: private hospital. ^aSome of the rural clusters with dispensaries (HCs that do not do deliveries) that, according to the original trial design, were supposed to have women’s groups mistakenly never had women’s groups formed in them so there are actually 31 control clusters out of the 82 originally randomized to receive quality improvement and establishment of womens’ groups. ^bThe ‘extra’ facilities are three from Lilongwe and one from Kasungu that were left over after randomly allocating the others to each of the four intervention groups. To ensure the balance across interventions was maintained, they were allocated to each of the four intervention groups.

Figure 5

Trial profile flowchart (further details are provided in the technical report9). CI: community intervention; FI: facilities intervention. ^aDispensaries do not perform deliveries and were excluded as women’s groups were not set up in their catchment areas.

Figure 6

Internal consistency of randomized controlled trial (RCT) results. Percentages of observed PMR and NMR reductions due to reductions in stillbirth rate, early NMR, and late NMR by RCT comparison groups. CI: community intervention; FIL: facilities intervention; NMR: neonatal mortality rate; PMR: perinatal mortality rate.