Uganda ranks among the top 10 countries in the world for newborn and child mortality rates [1] and among the top 34 for burden of stunting [2–6]. In Uganda, the neonatal mortality rate (NMR) is at 27 per 1000 live births, the infant mortality rate (IMR) is at 43 per 1000 live births and the under-five mortality rate (U5MR) is at 64 per 1000 live births [3, 4]. One in every 25 Ugandan children dies before their first birthday, and one in every 15 children dies before their fifth birthday. The southwest region has the second highest IMR in Uganda, at 76 per 1000 live births, and a U5MR of 128 per 1000 live births [4]. Similarly, the prevalence of stunting in southwest region is the third highest in Uganda, at 42% [4].

Uganda did not meet its Millennium Development Goal (MDG) targets to reduce the U5MR to 56 deaths per 1000 live births and the IMR to 31 per 1000 live births in 2015, and is likely to fail to meet the national targets to reduce the U5MR from 90 to 51 per live births by 2020 [7]. This could be attributed to uneven sub-national progress on child mortality and stunting. The persistent high U5MR may also partly be attributed to the low coverage of preventive evidence-based interventions known to be cost-effective in reducing child mortality, for example the expanded programme on immunization (EPI), optimal infant and young child feeding (IYCF) practices and antenatal care [1, 3, 4, 8, 9].

There is evidence that providing vaccines to infants and children against tuberculosis, polio, whooping cough, diphtheria, tetanus, measles, hepatitis B and haemophilus influenzae reduces infant and under-five mortality rates [3, 4, 9]. Similarly, sub-optimal IYCF practices are a major risk factor for child mortality and stunting. Studies have shown that up to 19% of child deaths can be prevented through optimal IYCF practices [10, 11]. Therefore, optimal IYCF practices can have the single largest impact on child mortality of all preventive interventions [12–14].

Several studies have used the Lives Saved Tool (LiST) in developing countries to model the impact of evidence-based interventions and stunting on child survival at national level [15–20]. However, the impact of stunting on child survival has not been formally modelled at sub-national level particularly in resource-limited settings where the burden of child mortality is highest. Our study therefore used LiST to model the impact of stunting reduction on child mortality in the southwest region of Uganda where child mortality and stunted growth are of great public health concern [1, 3, 12]. We also assessed the coverage of programs (antenatal care, EPI and IYCF practices) that have been shown to have high impact on prevention of stunting and child survival [12, 15, 16, 20, 21].

This study was undertaken as part of a larger project evaluating the risk factors and associated cost of preventing childhood stunting in southwest Uganda. The results on prevalence and risk factors of stunting have been published elsewhere [22].

Focus group discussions (FGDs) and interviews with key informants

A total of 16 focus group discussions and 16 interviews with key informants on IYCF knowledge and practices, EPI and ANC services were conducted in 16 of the 32 villages that were selected for household interviews (Fig. 1). For each focus group discussion, 10–15 participants were purposively selected. These mainly included village leaders, health workers and mothers of children under 5 years of age, and each focus group discussion lasted about 1–2 h. Interviews with individual key informants were carried out to obtain specific information from individuals that were selected due to their knowledge of ANC services, EPI and IYCF practices.

The inclusion of stunting and child mortality targets in the Uganda vision 2040 aimed at reducing stunted growth from 31% to 0% and U5MR from 90 to 8 per 1000 live births by 2040 [7] creates an enabling environment and opportunity to sustain the momentum on improving child survival in Uganda. This creation of an enabling environment has progressively been translated into policy and financial support by the government to increase the coverage of cost-effective child survival interventions [7, 9, 11]. However, there is still inequitable coverage of these interventions particularly in southwest districts like Buhweju where stunting and child mortality are of great public health concern [4, 22]. Our study conducted in southwest region shows that IYCF knowledge and practices in rural Uganda remain very poor. Similarly, although EPI and ANC are known to be among the most cost-effective interventions [15, 16, 18], the coverage of these programs in the study population was very low.

The World Health Organization recommends initiation of breastfeeding within an hour of birth, exclusive breastfeeding of infants up to 6 months of age and continued breastfeeding until the children are 2 years or older [13, 30]. However, in the current study a large proportion of mothers neither understood the meaning of optimal breastfeeding nor practiced it. Majority of the mothers also reported use of pre-lacteals (mainly water) before the commencement of breastfeeding. The use of pre-lacteals denies these infants the protective effect of colostrum associated with early initiation of breastfeeding. This practice therefore exposes the infants to increased risk of infections, stunting and mortality as other studies have also reported [31, 32]. There is global evidence showing that initiation of breastfeeding within 24 h of birth is associated with 44–45% reductions in all-cause and infection-related neonatal mortality [11]. Pre-lacteal feeding also has many implications for exclusive breastfeeding and resulting consequences since by definition, these children are not exclusively breastfed. Similar to our study findings in which mothers reportedly used pre-lacteals to clean the throats of the newborns, a study that assessed the factors associated with pre-lacteal feeding in the rural population of northwest Ethiopia reported that the practice was largely associated with poor knowledge of IYCF [33].

When children are given complementary foods from the age of 6 months, the foods should be adequate to support growth (i.e., an adequate quantity of food for each meal and number of meals per day) and have the right consistency, nutrients and energy density. However, most mothers in the study area did not practice timely introduction of complementary foods and majority of the children did not receive a minimum acceptable diet. Sub-optimal IYCF practices are common in Uganda [3, 4, 34], which is consistent with the findings of our study in which majority of the children aged 0–23 months were classified as not receiving a good standard of IYCF. According to the UBOS [3, 4], although almost all children (98%) in the Uganda Demographic and Health Survey (UDHS) were breastfed at some point, only half (53%) were breastfed within an hour of birth and only 66% of children who were aged 0–5 months at the time of the survey were exclusively breastfed. By 4–5 months of age, the proportion of exclusively breastfed children dropped to only 43%. The proportion of 2-year-olds who were breastfed was only 50% [3, 4]. In addition, among children aged 6–23 months, 13% were fed the recommended minimum dietary diversity, 45% received the recommended minimum number of meals per day and only 14% received the minimum acceptable diet [3, 4].

The lower ICFI scores particularly among older children found in this study could be attributed to the lack of knowledge on optimal IYCF and adequate skilled support at health facility and community levels, poor complementary feeding practices including low meal frequencies (since most mothers were reportedly preoccupied with farming activities) and non-responsive feeding as reported during FGDs. Therefore, it is not surprising that only 12% of the children received the recommended minimum acceptable diet. A study in Cambodia also reported higher child feeding index (CFI) scores among young children compared with old ones [21].

Sub-optimal IYCF practices and low coverages of EPI and ANC are associated with increased levels of stunted growth and child mortality [2, 11, 35–37]. The consequences of sub-optimal IYCF practices and low coverage of EPI and ANC services are evident in the persistently high prevalence of stunting and child mortality rates in Uganda, as these practices compromise child growth and development. Therefore, it is not surprising that Uganda is among the developing countries with the largest proportions of stunted children and high burden of child mortality [1, 3, 4]. In Uganda, one out of every three (29%) children under 5 years are stunted and 167,000 children under 5 years die every year [1, 3, 4]. Children born in rural areas, with uneducated mothers or among households in the lowest wealth quintile are more likely to be stunted or die before their fifth birthday [1, 4]. The southwest region recorded the third highest proportion (42%) of children who were stunted and second highest U5MR of 128 per 1000 live births out of all the regions in Uganda [4]. The recent nutrition survey conducted in Buhweju district showed that very little progress has been made in reducing the prevalence of stunting which is currently at 51% [22]. Therefore, the findings of this study suggest that the high prevalence of stunting and child mortality in Buhweju district could be largely attributed to low coverage of child survival interventions (including sub-optimal IYCF practices and low coverage of EPI and ANC services).

In Uganda, although 97% of pregnant women attend their first antenatal visit, only 60% complete the World Health Organization (WHO) recommended minimum of four antenatal visits [1, 3, 4]. Much lower ANC attendance compared to the national and southwest region averages were registered in Buhweju district. This could be attributed to the low availability and access to antenatal care in the district. Most participants in FGDs and interviews with key informants also reported that the low quality of ANC services (understaffing, poor counseling services and poor client-provider relations) in facilities where it is provided has partly contributed to the poor attendances. It is well known that on average, less than one-quarter of facilities in Uganda have all essential equipment and supplies for basic ANC [1, 37]. In addition, although vaccination coverage in Uganda has improved over the last 10 years [1, 3, 4], majority of the children under 5 years in Buhweju district were not fully vaccinated. Low vaccination coverage exposes the children to an increased risk of dying from preventable diseases. Previous reviews of the main factors associated with low vaccination coverage in Uganda and other developing countries found these to be: maternal education, maternal age, exposure to media, maternal healthcare utilization, health staff attitudes and practices, reliability of health services, parents’ practical knowledge of vaccination, fear of side effects, conflicting priorities, parental beliefs and immunization plan [9, 38]. These risk factors for low vaccination coverage are similar to the findings of our study in which access to health services and ANC contributed to higher vaccination coverage.

In the study population, there were also many missed opportunities along the continuum of care (from gestation to 5 months of age) for the delivery of nutrition education regarding optimal IYCF and communication of information on EPI. For example, in our study a small proportion of the mothers reported receiving information on IYCF from health care providers. The low consumption of animal sources of protein (eggs, dairy products, meat and fish), as evidenced in the current study, is likely to be another major contributor to the high levels of stunted growth and mortality among children in Buhweju district. A study in Malawi that analyzed blood samples from 313 children aged 12–59 months (with and without stunting) found lower serum concentrations of all nine essential amino acids in the stunted children compared with non-stunted children [39].

In modelling the impact of stunting on child survival in a rural setting, our study demonstrates that reducing the prevalence of stunting to zero would result in the highest impact on child mortality – saving more lives and averting more cases of stunted children. It is therefore crucial to inform decision-makers on the need to support the prioritization of interventions aimed at reduction of stunting and advocacy for increased funding of maternal and child survival interventions. There is already evidence that LiST identifies priority areas for child health investment based on modelling the impact of evidence-based interventions at varying levels of coverage [15–18, 20]. Although LiST has been mainly used to model the impact of scaling up maternal, newborn and child health intervention on child mortality [15–20], in our study, we presented the impact of reducing the prevalence of stunting on child survival in a rural setting. It is therefore important to note that the approach used in our study, models the impact of reducing stunted growth on child mortality but does not provide details on how stunting reduction can be achieved.

The study concludes that IYCF practices and coverage of EPI and ANC in the study population are sub-optimal thus the high prevalence of stunting and child mortality in the region. LiST demonstrated that prevention of stunting would reduce child mortality in rural Uganda. Therefore, increased investment in cost-effective child survival interventions targeting rural areas of Uganda would have a significant impact on stunting and child mortality. The translation of messages regarding nutrition education into actual practice focusing on increasing exclusive breastfeeding and consumption of animal sources of protein would boost linear growth in children under 2 years of age and reduce mortality. Social and behavior change communication (SBCC) designed to promote optimal IYCF, EPI and ANC should target multiple delivery channels and also address reported barriers.