Menu
Menu
Menu
Menu
Background: Improving child health remains one of the most significant health challenges in sub-Saharan Africa, a region that accounts for half of the global burden of under-five mortality despite having approximately 13% of the world population and 25% of births globally. Improving access to evidence-based community-level interventions has increasingly been advocated to contribute to reducing child mortality and, thus, help low-and middle-income countries (LMICs) achieve the child health related Sustainable Development Goal (SDG) target. Nevertheless, the coverage of community-level interventions remains suboptimal. In this study, we estimated the potential impact of scaling up various community-level interventions on child mortality in five East African Community (EAC) countries (i.e., Burundi, Kenya, Rwanda, Uganda and the United Republic of Tanzania). Methods: We identified ten preventive and curative community-level interventions that have been reported to reduce child mortality: Breastfeeding promotion, complementary feeding, vitamin A supplementation, Zinc for treatment of diarrhea, hand washing with soap, hygienic disposal of children’s stools, oral rehydration solution (ORS), oral antibiotics for treatment of pneumonia, treatment for moderate acute malnutrition (MAM), and prevention of malaria using insecticide-treated nets and indoor residual spraying (ITN/IRS). Using the Lives Saved Tool, we modeled the impact on child mortality of scaling up these 10 interventions from baseline coverage (2016) to ideal coverage (99%) by 2030 (ideal scale-up scenario) relative to business as usual (BAU) scenario (forecasted coverage based on prior coverage trends). Our outcome measures include number of child deaths prevented. Results: Compared to BAU scenario, ideal scale-up of the 10 interventions could prevent approximately 74,200 (sensitivity bounds 59,068–88,611) child deaths by 2030 including 10,100 (8210–11,870) deaths in Burundi, 10,300 (7831–12,619) deaths in Kenya, 4350 (3678–4958) deaths in Rwanda, 20,600 (16049–25,162) deaths in Uganda, and 28,900 (23300–34,002) deaths in the United Republic of Tanzania. The top four interventions (oral antibiotics for pneumonia, ORS, hand washing with soap, and treatment for MAM) account for over 75.0% of all deaths prevented in each EAC country: 78.4% in Burundi, 76.0% in Kenya, 81.8% in Rwanda, 91.0% in Uganda and 88.5% in the United Republic of Tanzania. Conclusions: Scaling up interventions that can be delivered at community level by community health workers could contribute to substantial reduction of child mortality in EAC and could help the EAC region achieve child health-related SDG target. Our findings suggest that the top four community-level interventions could account for more than three-quarters of all deaths prevented across EAC countries. Going forward, costs of scaling up each intervention will be estimated to guide policy decisions including health resource allocations in EAC countries.
Headquartered in Arusha, Tanzania, the East African Community (EAC) is a regional intergovernmental organization bringing together Kenya, Uganda, the United Republic of Tanzania (henceforth referred to Tanzania), Burundi and Rwanda for a wider and deeper cooperation among these countries and other regional economic communities for mutual economic, social and political benefit (https://au.int/en/recs/eac). In the health sector, Yamin et al. (2017) argue that achieving universal health coverage (UHC) in EAC would require EAC countries to put in place human rights-based approaches for ensuring the health needs and rights of the people are being met at the community level. This would alsofoster community ownership and legitimacy of health reforms [39]. Child mortality remains one of the primary public health challenges faced by the region and consequently programs related to the prevention and reduction of child mortality require a combined effort at all levels of goverment. Despite the remarkable progress made by three EAC countries (Rwanda, Uganda, Tanzania) to achieve the MDG 4 (Table 1), there is still a lot to be done in order to reduce preventable child mortality among these countries and across the EAC region as a whole. Table Table11 summarizes the EAC context including population size, economic and key health indicators. With a median age ranging from 15.9 years to 19.6 years, the EAC has one of the youngest populations globally (Table (Table1).1). Similarly, the region has one of the world highest birth rates (Table (Table11). Characteristics of the EAC countries included in our analysis GDP, gross domestic product; OOP, out of pocket; HDI, health development index; MDG, millennium development goal; US$, United States dollar; EAC, East African Community. Data presented in Table Table11 were abstracted from various publications [26, 40–45] Drawing on prior research [37, 38, 46, 47], we identified 10 preventive and curative CLIs that have been reported to reduce child mortality: Breastfeeding promotion, complementary feeding, vitamin A supplementation, Zinc for treatment of diarrhea, hand washing with soap, hygienic disposal of children’s stools, oral rehydration solution (ORS), oral antibiotics for treatment of pneumonia, treatment for moderate acute malnutrition (MAM) and prevention of malaria using insecticide-treated nets and indoor residual spraying (ITN/IRS). These interventions can be classified into three categories: Each of these interventions has impact on specific cause(s) of death and/or risk factors [37, 38, 46–51]. For example, vitamin A supplementation, Zinc for treatment of diarrhea, hand washing with soap, hygienic disposal of children’s stools, and ORS interventions reduce child mortality by decreasing diarrhea. Oral antibiotics for treatment of pneumonia intervention reduce child mortality by decreasing deaths due to pneumonia, while ITN/IRS prevent malaria and related deaths. Interventions that have impact on risk factors for disease (for example, breastfeeding and complementary feeding) affect multiple causes of child mortality by modifying the probability of death due to specific causes of death. For example, interventions that reduces stunting and wasting will also indirectly reduce the probability of dying of diarrhea, pneumonia and malaria. We focused on interventions that can be delivered at community level by CHWs. Nine of the 10 CLIs that we selected are delivered at community level at least 50% (Table 2). We retrieved data on the percent of each interventions per delivery channel from Lives Saved Tool (described below) and, our modeling exercise assumed that the delivery channel for each intervention would remain constant over the study horizon. Similarly, it is assumed that variations in intervention coverage drive mortality changes, and the impacts on mortality of distal factors (for example, socioeconomic status) are mediated by changes in intervention coverage [49–52]. Percent of each intervention delivered at each level of healthcare delivery channels across EAC aIncluded interventions that are offered at community at 40% or more Breastfeeding promotion (exclusive breastfeeding 1-5 months). ITN/IRS insecticide-treated bed nets (ITNs) and indoor residual spraying (IRS); EAC East African Community. Source: Lives Saved Tool We used the Lives Saved Tool (LiST) [53, 54] – one of the modules in the Spectrum software package – to model the number of deaths among children younger than five years that could be prevented across EAC as a result of expanding proven effective CLIs (change in coverage), while accounting for EAC country specific health status (Table (Table1)1) and distribution of cause-specific mortality (Figs. 1 and and2).2). LiST has been used widely in lower- and middle-income countries (LMICs) to estimate the potential impact and cost of expanding maternal, newborn and child health interventions across the continuum of care [37, 38, 55–57]. Percent of neonatal deaths by proximate causes across East African Community (2014/2015). Source: Lives Saved Tool Percent of child death-post neonatal by proximate causes across East African Community (2014/2015). Source: Lives Saved Tool. While the details for ‘Other’ in the Fig. 2 was not provided in LiST, drawing on existing literature of global burden of diseases, injuries and risk factors, we believe that this section would include malnutrition, congenital anomalies, drowning, and foreign bodies [58] To make the projections, LiST employs a linear deterministic model and links with other modules (e.g., Family Planning module, AIDS Impact module and Demographic Projections module) available in the Spectrum package [53]. Our LiST model input include estimates of intervention effects and intervention coverage – defined as “the proportion of women and children in need of life–saving intervention who actually receive it” [37]. The model output was the number of deaths prevented disaggregated by each CLI. Estimates of the effects of interventions on cause specific child mortality were generated using the Child Health Epidemiology Reference Group intervention review process that draws on Cochrane Collaboration and the Working Group for Grading of Recommendations Assessment, Development and Evaluation (GRADE) [59]. The baseline population level coverage data for each intervention were derived from the most recent nationally representative surveys including demographic and health survey (DHS) and world population prospects (WPP) [37, 53]. Using LiST, we modeled the impact on under-five child mortality of scaling up the 10 CLIs from baseline coverage (2016) to ideal coverage (99%) by 2030 (Table 3). To estimate the impact under the ideal scale up scenario, we increased the coverage only for the 10 interventions that can be delivered by CHWs at the community level (Table (Table3),3), while holding all baseline population level coverage for other interventions in LiST module constant. We increased the coverage of our target interventions gradually using linear interpolation from 2016 to 2030 (i.e., study time horizon) (Table (Table3).3). We selected the study time horizon to cover the period post MDG era through the end of SGD era. To estimate the counterfactual (what would happen under business as usual (BAU) scenario), we forecasted coverage of the 10 interventions from 2016 to 2030 based upon existing trends in coverage for these interventions from 2010 to 2016 (7 years) using exponential smoothing methods and adjusted for seasonality as appropriate. We then calculated (and report in the results) number of deaths that could be prevented by ideal scale up of the 10 CLIs relative to scale up under business as usual scenario (Table 4). Baseline coverage and percent scale-up for community level interventions across EAC aExcluding breastfeeding; bSupplementary feeding and education; CLIs community-level interventions, EAC East African Community, ORS oral rehydration solution, ITN/IRS insecticide-treated bed nets (ITNs) and indoor residual spraying (IRS) Number of deaths averted by target year (2030) by intervention under ideal coverage scenario relative to BAU scenario ITN/IRS insecticide-treated bed nets (ITNs) and indoor residual spraying (IRS), BAU business as usual. *Sensitivity bounds were derived from sensitivity analyses that estimated effects of interventions based upon the highest level of effectiveness reported for all interventions (upper bound) relative to the lowest levels of effectiveness (lower bound). An em dash (─) indicates that the item is not applicable, or the value is zero, because the coverage under BAU scenario reached 99% by 2030, which is equivalent to the coverage under the ideal scale up scenario For intervention coverage where the existing trends were decreasing in the period of 2010–2016, forecasting the coverage from 2016 to 2030 would have led to considerably lower coverage by 2030 under BAU scenario, thus overestimating the number of deaths prevented under ideal scale up scenario relative to BAU scenario. Given ongoing emphasis on increasing coverage community level interventions to help LMICs achieve universal health coverage by 2030, it is unlikely that the decreasing trend in coverage reported for some interventions (from 2010 to 2016) would continue to 2030. As such, we used a more conservative approach by using mean coverage from the existing trends over 7 years (2010–2016) instead of the decreasing forecasted values. We assumed the percent delivery of each CLI at various delivery channels constant throughout the time horizon (Table (Table3).3). Using autoregressive integrated moving average (ARIMA) time series approach and reported under-five mortality from 2000 to 2017, we forecasted under-five mortality trends in EAC up to 2030 (Fig. 3). We used Spectrum software v5.753 (https://www.livessavedtool.org/listspectrum) and R software 3.4.4 for all analyses [60]. Reported and forecasted trends in under-five mortality across EAC (UNICEF reported estimates, 2000–2017, and forecasted estimates, 2018–2030). We forecasted under-five mortality trends in EAC from 2018 to 2030 using UNICEF reported under-five mortality from 2000 to 2017 and autoregressive integrated moving average time series approach. Based on our forecasted estimates, Rwanda and Uganda would meet the SDG target for under-five mortality of at least as low as 25 per 1000 live births
N/A
