Triple return on investment: the cost and impact of 13 interventions that could prevent stillbirths and save the lives of mothers and babies in South Africa

Despite increasing global attention to maternal, newborn and child health, issues surrounding stillbirths remain largely neglected. Stillbirths were not included in tracking progress on the Millennium Development Goals (MDGs) and have received little recognition for intervention strategies and programs globally; stillbirth targets are absent from the Sustainable Development Goals. Stillbirths are however one of the most adverse outcomes of pregnancy, causing significant negative social and health implications for families and society [1].

In 2009, 2.6 million third-trimester stillbirths were estimated worldwide, of which 98% occurred in low and middle-income countries [2]. Over 55 stillbirths occur daily in South Africa and the 2013 institutional third trimester stillbirth rate (SBR) was 17.6 per 1 000 births [3]. This far exceeds the SBRs in other middle-income countries such as Brazil, Russia and China (all approximately 10 per 1 000 births) [2]. Unlike most low and middle-income countries (LMICs), South Africa does include stillbirths in the reporting of vital statistics. The Cause of Death Certification (DHA-1663) registers stillbirths and neonatal deaths occurring in both public and private sector facilities [4,5]. Data are registered with the Department of Home Affairs and analyzed by Statistics South Africa (Stats SA). Public sector data are collected by the District Health Information System (DHIS) and the Perinatal Problem Identification Program (PPIP). PPIP is an electronic audit of perinatal and maternal deaths implemented in 588 facilities in South Africa (representing 73% of institutional births recorded by the DHIS) [6]. Given that most South African women (approximately 90%) give birth in health facilities [7], the number of stillbirths occurring at home is likely small. Although these mechanisms are in place for capturing stillbirth statistics, records are often incomplete due to under-reporting and delayed or late registration. There is also evidence for misclassification of causes [8].

Within the ICD-10 classification, stillbirths are defined by a birthweight of 500 g or more, or if birthweight is unknown, gestational age of 22 completed weeks or more, or crown-heel length greater than 25 cm. For international comparison, WHO recommends a third trimester stillbirth definition where birthweight is above 1 000 g, gestational age is 28 complete weeks or more, or crown-heel length is greater than 35 cm [4]. In LMICs, stillbirths are often not weighed, making it difficult to differentiate between second and third trimester deaths [9,10].

The World Health Organization (WHO) has stressed the importance of ending these preventable deaths. The Every Newborn action plan launched in 2014 sets a target to reduce stillbirth rates globally to 10 per 1 000 births by 2035, with an interim goal of 12 per 1 000 births by 2030 [13]. While South Africa is a signatory to the plan and has set a national target that is more ambitious than the global one (reduce stillbirth rate to 10 per 1 000 births by 2016) [14], the country is not currently making progress towards meeting these goals.

Given the need for high quality, local data as a basis for priority setting and budgeting [15], our analysis (1) reviews available South African data to approximate trends in stillbirth rates between 2000 and 2013 and to investigate the national distribution of stillbirths; and (2) evaluates the costs of 13 priority interventions and their impact on stillbirths and maternal and newborn mortality in South Africa between 2015 and 2030. The paper predicts the potential numbers of deaths averted and estimates the resources required to scale up these interventions.

This is the first analysis in South Africa to assess the impact of scaling up interventions to avert stillbirths during pregnancy and childbirth. Our model shows that scaling up 13 essential interventions could prevent a significant percentage of stillbirths and save the lives of mothers and newborns. If full coverage (99%) is achieved in 2030, these interventions could reduce the SBR by 30% from 17.6 per 1 000 births in 2014 to 12.4 per 1 000 births, and could lead to an MMR of 132 per 100 000 live births and an NMR of 7 per 1 000 live births. This would lead to approximately 6 200 additional maternal and newborn lives saved and approximately 5 400 stillbirths averted in 2030. These estimates are made with the contraceptive prevalence rate for modern family planning methods set to 70% in 2030. When CPR is increased to 75% and 80%, the model predicts that 9 000 lives and 6 800 lives could be saved, respectively, due to the reduced number of pregnancies and deaths.

The majority of stillbirths averted are during the intrapartum period (72%), despite higher rates of antepartum stillbirth at baseline. This occurs for several reasons. Causes of intrapartum stillbirth are better understood and documented, and have proven solutions. Unexplained intrauterine death is recorded for 48% and 16% of antepartum and intrapartum stillbirths, respectively [6]. These unidentified causes make it difficult to implement a comprehensive set of interventions during the antenatal period. As a result, the antenatal interventions which have been modelled are effective in small subsets of pregnant women and have effectiveness values that range from 10% for diabetes case management to 20% for management of pre-eclampsia, hypertensive diseases, and fetal growth restriction. While syphilis detection and treatment has an effect size of 82%, South Africa already has high coverage of syphilis detection and treatment in 2014 (92%), thereby reducing the incremental impact of scaling up this intervention.

Labor and delivery management has the most significant effect on stillbirth outcomes. Comprehensive emergency obstetric care for women delivering in facilities includes augmentation of labor, instrumental delivery and emergency caesarean section. Despite having high levels of facility deliveries and skilled birth attendance in South Africa, many women and babies die during pregnancy and childbirth due to delays in accessing services and poor quality of care once in health facilities [47]. This has not been represented in LiST, since the model relates the quality of an intervention to its level of coverage; it is assumed that labor and delivery management is delivered at high quality due to its high baseline coverage.

With full intervention coverage and family planning at 80%, the package to reduce stillbirths, maternal and newborn deaths is estimated to cost US$420 million (US$7.8 per capita) in 2030. This cost is less than the projected baseline cost in 2014. The intervention for labor and delivery management requires the most additional resources (US$20 million). The intervention costs increase with reduced family planning coverage. When family planning coverage is 70% in 2030, the overall cost increases to US$550 million (US$10.2 per capita), requiring an additional investment of US$84 million (US$1.6 per capita). The interventions are all highly cost-effective, given that their ICERs fall well below SA’s GDP per capita of US$7 500.

However, cost calculations do not include the costs of infrastructure development. Taking these into account would increase the total costs. The estimates provided can still guide provincial health department budgetary considerations with regard to providing a maternal and newborn health package.

The Lancet’s Stillbirth Series evaluated global rates and causes of stillbirths. It showed that globally, scaling up the initial 15 maternal and neonatal care interventions to full coverage could avert 1.1 million (45%) third-trimester stillbirths, 201 000 (54%) maternal deaths, and 1.4 million (43%) neonatal deaths annually, at an additional cost of US$10.9 billion or $2.32 per person in 2011 [31,32]. Our analysis has also shown a high reduction in mortality, yet the global and country level costs cannot be compared. Costing was additionally performed separately in The Lancet analysis and has since been incorporated in LiST [31].

The analysis models an ambitious goal of 99% intervention coverage, which may be difficult for South Africa to accomplish even within the 2030 timeframe. Other interventions such as PMTCT and immunizations have however shown that high coverage levels can be attained with concerted effort. The results provide evidence for the impact of focused attention on stillbirth prevention; South Africa could meet the WHO stillbirth target of 12 per 1 000 births in 2030. This should provide impetus for a renewed focus on stillbirths in the post-2015 development era. South Africa has already shown some commitment to this by setting a national stillbirth rate target [14], consistently monitoring and reporting stillbirths, as well as by endorsing the WHO Every Newborn action plan [13]. Stillbirth targets should also however be reflected in South Africa’s National Development Plan for 2030. The South African National Committee for the Confidential Inquiries into Maternal Deaths (NCCEMD) and the National Perinatal Mortality and Morbidity Committee (NaPeMMCo) have already recommended the interventions in this analysis [8,29]. Considerable effort is however required to ensure these interventions are effectively delivered to those who are currently not receiving them, often the most marginalized and at-risk women and babies. Since the largest proportion of stillbirths occur in district and regional hospitals (76%), efforts should be focused on these facilities where there is greatest potential to save lives.

The coverage of antenatal corticosteroids (ACS) is low at baseline (20%). Currently, ACS are administered primarily in district hospitals but coverage could be even greater if they were promoted at the community health center level. This would allow earlier treatment initiation, prior to transferring the mother to a district hospital. Concern has however been raised over the potentially harmful effects of administering ACS to women not at risk of premature labor and imminent delivery, especially in low income settings [48]. While the skills and resources at the district hospital level enable accurate gestational age estimation and premature labor diagnosis, this may be lacking at the CHC level. Promoting the use of ACS in CHCs should thus be approached with caution in South Africa.

Family planning has been shown to have a significant impact in saving the lives of mothers and babies [49]. Fewer pregnancies in South Africa would also reduce the number of stillbirths. When modern contraceptive prevalence is increased to 80%, the total fertility rate is less than 2 (below the replacement rate), but this is still on par with countries such as Singapore and Brazil [50]. Family planning is crucial for health, development and women’s equality, particularly among high risk groups such as adolescents and HIV-positive women.

The heterogeneity in mortality rate and intervention coverage by province or district is not represented in this national analysis. There is a significant contrast between reported stillbirth rates in the Western Cape and the Free State, for example. Stillbirth reporting is inadequate in some provinces, hence the national SBR is likely an underestimate and the calculated provincial and national SBR reductions are uncertain. Furthermore, PPIP facility coverage has increased significantly between 2000 (27 sites) and 2013 (588 sites). While coverage included each province, the earlier data is less representative.

Barriers to stillbirth reporting are not entirely understood. The formal burial required for registered stillbirths presents a financial disincentive for reporting. Hospital staff may assist families in avoiding this cost by recording stillbirth as a miscarriage [51]. In rural Mpumalanga province, the average funeral cost for children younger than six years is US$116 (adjusted for inflation), which is 92% of the average monthly household income [52]. The definition of stillbirth also affects reporting. Difficulty differentiating between second and third trimester stillbirths, as well as antepartum and intrapartum stillbirths can present additional obstacles. Stats SA accounts for under-reporting by adjusting the reported stillbirth numbers by the same rate of under-reporting in infant deaths. However, it is not known whether the same improvement in infant death registration has been experienced for stillbirths. The decline in stillbirth rates reported by Stats SA may thus result from the adjustment factor [16]. Since the third trimester stillbirth definition was used for modelling and projections, the total number of stillbirths averted is not fully represented.

As the foremost health challenge facing South Africa, prevention and treatment of HIV remains critical at each level of care. The effect of early HIV treatment on maternal deaths has been assessed, yet the model does not consider the impact of HIV/AIDS on stillbirths or newborns. There is some evidence to suggest a relationship between HIV/AIDS and stillbirth outcomes [53-55], but the effects of an HIV/AIDS intervention in this sphere has not been firmly established. Including such effects in this analysis may have had an additional impact on stillbirths in South Africa.

Diabetes case management was the least cost-effective intervention, but the impact of diabetes prevention has not been assessed for stillbirths. With rising obesity rates in South Africa, especially amongst young women [56], weight control before pregnancy would potentially provide a high impact intervention. This has however not yet been documented.

Each year over 20 000 stillbirths are recorded in South Africa, yet these deaths receive insufficient attention and were not included as a target within the MDGs. South Africa has shown some commitment to reducing stillbirths, as one of the few countries in Africa to consistently record and report stillbirths and by setting a target stillbirth rate within the current national strategic plan [14]. There is however a need for more robust stillbirth data collection to ensure adequate progress tracking. The inter-ministerial committees on maternal, perinatal and child mortality have recognized the importance of focusing on stillbirths and family planning, and have made several recommendations for essential interventions. Family planning is included in the Sustainable Development Goals and can prevent unplanned pregnancies and the subsequent abortions, stillbirths and maternal and newborn deaths. In this analysis we show that focusing on stillbirths with full coverage of 13 essential interventions for mothers and newborns in 2030 could reduce the SBR to 12 per 1 000 births, and could result in an MMR of 132 per 100 000 live births and an NMR of 7 per 1 000 live births. If the coverage of family planning was simultaneously increased to 80% in 2030, the full package would cost US$7.8 per capita (less than the baseline cost in 2014). The costs appear to be substantial, but are within the capacity of the South African health budget. The cost-effectiveness values are also below the recommended willingness-to-pay threshold of the South African per capita GDP, implying that even at scale, these interventions could be affordable. In South Africa’s post-2015 agenda, investing in family planning and the 13 essential interventions will have a triple return in preventing stillbirths and saving the lives of mothers and newborns.