Effect of a low-dose, high-frequency training approach on stillbirths and early neonatal deaths: a before-and-after study in 12 districts of Uganda

Abstract

BackgroundPost-partum haemorrhage and neonatal asphyxia are the leading causes of maternal and newborn deaths, respectively, in Uganda. However, proven interventions that can save the lives of women and children are not being widely practised. Here, we assess the effects of two onsite training programmes for interventions to reduce deaths from post-partum haemorrhage and neonatal asphyxia, plus three different levels of performance support after the training in health-care facilities in 12 districts of Uganda.MethodsWe invited all public and some private non-profit health facilities that deliver babies in western and eastern regions of Uganda to participate. At all study sites, training in the Helping Mothers Survive and Helping Babies Breathe interventions was given to all health-care providers who attend births—medical doctors, clinical officers, midwives, nurses, and nurse assistants. The training consisted of low-dose, high-frequency onsite training followed by 20 weeks of provider-led practise with low-cost simulators. At all intervention sites, a midwife was designated as a clinical mentor or peer leader of simulation-based practice. We used a cluster randomised design to match districts for volume of births and presence of an operating theatre, and then randomly allocated districts to one of three training follow-up groups. A full-support group had onsite training supported with a peer practice coordinator and weekly practice sessions, supplemented with mobile messages to remind providers to practise. A partial-support group had an onsite training programme supported with a peer practice coordinator; and a control group received onsite training only. We collected routine monthly service-delivery statistics from the government Integrated Maternity Register (IMR) and introduced a supplemental maternity register to capture indicators not present in the IMR. Outcomes of interest were fresh stillbirths and neonatal deaths within 24 h at baseline and 6 months after the intervention. We had also aimed to study post-partum haemorrhage retained placenta, but no baseline data were available.FindingsBetween May, 2013, and May, 2016, training was delivered at 125 participating health facilities: 11 hospitals, 21 level IV health centres, 79 level III health centres, and 14 level II health centres. The control and partial-support groups had three hospitals each, but the full group had five hospitals, including the only regional referral hospital. The partial-support group had more level III and II (n=35) hospitals than did the full-support and control groups (n=29 each). The proportion of adverse birth outcomes (combined fresh stillbirths and neonatal deaths within 24 h), appeared to decrease significantly in all the study groups from baseline to 6 months post intervention: from 26·5 to 9·6 per 1000 livebirths (p<0·001) in the full support group, from 24·2 to 7·8 per 1000 livebirths (p<0·001) in the partial-support group, and from 17·2 to 9·5 per 1000 livebirths in the control group (p=0·008). We also noted a 47% reduction in retained placenta and a 17% reduction in post-partum haemorrhage between the intervention period and 6–9 months after the intervention.InterpretationFacility-based simulation training is feasible and can be effective in reducing adverse birth outcomes.FundingSaving Lives at Birth: a partnership of USAID, Norwegian Ministry of Foreign Affairs, Bill & Melinda Gates Foundation, Grand Challenges Canada and UK Aid.