Water and sanitation influence on child health in Namibia: using demographic and health surveys

Child survival and health are high on the global agenda for long-term development. However, children in underdeveloped countries such as Namibia continue to die from preventable causes before reaching their fifth birthday. According to the World Health Organization (WHO), the African region has the highest child mortality rates, with around 74 fatalities per 1,000 live births, nine times greater than cases reported in the European area (WHO 2021). Furthermore, WHO estimates that roughly 28% of all fatalities in children under five are preventable and attributed to environmental factors (WHO 2016).

The global under-five mortality (U5M) has reduced by 60% from 1990 to 2020, currently at 5.2 million, equating to a daily death rate of 14,000 children (UNICEF 2020). Reaching the targets set by the Millennium Development Goals (MDGs), which expired in 2015, remains a far-fetched dream (Brault et al. 2020). Goal 3 of the Sustainable Development Goals (SDGs), which replaced the MDGs, is to ‘reduce new-born mortality to at least 12 deaths per 1,000 live births and under-5 mortality to at least 25 per 1,000 live births’ (United Nations 2017). According to the SDGs and current statistics, countries and global efforts must be coordinated and accelerated to prevent the expected deaths of 48 million children between 2020 and 2030 before they reach their fifth birthday (UNICEF 2020).

Namibia is a sub-Saharan African (SSA) country with the world's highest child mortality rate (UNICEF 2020). Despite this, there has been significant progress in the African region to lower child mortality rates in recent years. With 76 deaths per 1,000 live births, SSA continues to have the world's highest U5M rate. In SSA, one out of every 13 children dies before reaching their fifth birthday due to preventable causes (UNICEF 2015, 2020).

The current burden of a high mortality rate among children under five in the SSA is due to a high neonatal mortality rate. The first 30 days of an infant's existence are critical for survival and adaptability. As a result, newborn children at this age are vulnerable to a variety of environmental stressors that, if not addressed, can result in premature death. Furthermore, exposure to environmental factors during infancy causes health conditions that continue into adulthood, leading to a lower life expectancy (UNICEF 2020), reducing the human capital required to achieve global SDGs. Improving environmental conditions could prevent 47% of U5M in the newborn period and increase longevity (WHO 2021).

In Namibia, infant mortality has decreased dramatically over time. The country had the highest U5M rates in Southern Africa in the early 1990s and the early 2000s (62 deaths per 1,000 live births). Furthermore, the country's infant mortality rate fell from 36.3 deaths per 1,000 live births in 2013 to 30.7 deaths per 1,000 live births in 2019 (O'Neill 2021). Between 2015 and 2018, there was also a significant reduction in the neonatal mortality rate. Improved availability of sanitation facilities and health services at the home level and improved household indices, and breastfeeding initiation contributed to lowering infant mortality in the early 1990s and the early 2000s (Ndirangu et al. 2018; Headey & Palloni 2019).

Although the child mortality rate has decreased over time, particularly in the late 1990s and the early 2000s, the infant mortality rate in Namibia remains higher than the 3.9 targets set by the SDGs. Reaching the SDG 3.9 target is also threatened by the rising in HIV infections in Namibia. Namibia had an HIV prevalence rate of 913 per 10,000 people in 2020, with children occupying 4% of the infected population (UNAIDS 2021).

Namibia is a middle-income country, although with social inequalities between ethnic groups and uneven distribution of wealth across regions (MOHSS et al. 2014). Namibia's population grew by 13% between 2001 and 2011 (National Planning Commission 2011); however, the country remains one of the least densely populated, with a population density of 2.6 people per square kilometres (MOHSS et al. 2014). Topography/terrain, population density, and poor rainfall patterns, on the other hand, make it challenging to distribute piped water from safe sources and enhance sanitation facilities (MOHSS et al. 2014). Despite the above knowledge, it's worth noting that no study investigated the impacts of the interactions of the HIV/AIDS epidemic, migration from rural to urban regions, population growth, and economic inequality in Namibia's U5M rate using the national demographic and health data.

In 2011, 80 and 51% of Namibian homes had access to safe drinking water and improved toilet facilities, respectively (National Planning Commission 2011). The statistics above suggest that between 20 and 49% of households lack safe drinking water and improved sanitary facilities. Children in those households are more likely to suffer or die from diarrheal diseases before reaching their fifth birthday or are more likely to die from environmental causes within their first month due to their underdeveloped physiological systems (US EPA 2011). Evidence from 172 countries shows that improved water and sanitation infrastructure lowers the odds of child diarrhoea by 7 and 13%, respectively (Gunther et al. 2010), suggesting improvement in children's health.

Children's health is a sensitive indicator of the country's overall health and economic situation (Rayhan & Khan 2006). To our knowledge, no research has looked into the impact of water and sanitation on the health of Namibian children.

Different studies have found a link between demographic, socio-economic, environmental, cultural, and health-related factors and child health in other nations (Rayhan & Khan 2006; Abir et al. 2015). A study conducted using demographic and health data from 28 SSA countries reported a correlation between enhanced immunization coverage and ethnic homogeneity with lower child death odds (Boco 2010). A positive association exists between rurality, residence region, mother's education, marital status, religion, and ethnicity towards vulnerability and a weak survival rate of children under five in Cameroon (Defo 1993). The findings from a Ghanaian study imply that numerous social and environmental factors influence children's health and survival, and this study will examine the Namibian community (Kanmiki et al. 2014). It is worth noting that both the Cameroon and Ghanaian studies used data from national demographic and health surveys.

This study attempted to determine the impact of demographic, socio-economic, and environmental determinants on child health in Namibia, using the Namibian demographic and health surveys (NDHS) from 2006 to 2013. The study used the logistic regression model to investigate the influence of water sources and sanitation facilities on U5M in Namibia. Our goal was to determine the impact of improved water sources and toilet facilities on U5M in Namibia, comparing variables from the NDHS, between 2006 and 2013. Secondly, the model compared the effects of improved toilet facilities and protected water sources on U5M in Namibia between 2006 and 2013 by region.

The NDHS of 2006 and 2013 were the sources of data in this study. The NDHS is a cross-sectional survey that questioned mothers in Namibia between 15 and 49 years old to provide policymakers with demographic, socio-economic, and health data. The household survey gathered data on various home amenities, such as the source of drinking water and the type of toilet facilities available, and data on family factors, maternal education, child health, and dietary variables.

The 2006 NDHS is Namibia's third comprehensive survey, with the interviews conducted between November 2006 and March 2007. Additionally, the 2013 NDHS is the fourth DHS survey, and it collected data from May to September of that year. The Ministry of Health and Social Services (MOHSS) was in charge of both surveys. The data are representative of Namibian households by area and residence (rural and urban). In this study, the participants are de jure household members or regular residents. The sample size was 4,821 in 2006 and 4,768 in 2013. In 2006, there were five missing variables for a mother's marital status, while in 2013, there were 36 missing ones. Finally, concerning the measurement of HIV results, 2,873 observations were observed in 2013, while no participants answered the HIV questions in the 2006 survey.

During the interviews, mothers indicated whether their child was still alive or dead as well as their age when they died. As a result, we created a binary variable for U5M, which takes one if the child died before the age of 5 and zero if the child is still alive. This study applied the model with the main determinants of U5M risk in two categories, according to the Mosley and Chen framework for child survival in underdeveloped countries (Mosley & Chen 1984).

Table 1 shows the categorization of the explanatory variables. The environmental factors covered in the analysis are the source of drinking water and the type of toilet facilities. The WHO/UNICEF Joint Monitoring Programme for Water Supply and Sanitation (WHO & UNICEF 2020) defines and classifies both variables. The source of drinking water is a good predictor of its fitness for use, which is essential in Namibia because waterborne infections like diarrhoea are common. Improved water supplies refer to water sources that are more likely to deliver adequate water. The options are a piped supply within the yard, the dwelling, or the plot; a tube well; protected well or protected spring; or bottled water.

Unprotected springs, wells, and tanker trucks/carts with drums are all considered unprotected drinking water sources since they likely contain disease-causing substances. The type of toilet facilities was improved and non-improved. Improved toilet facilities refer to flushing toilets into a piped sewer system, septic tank, or pit latrine, and pit latrines with a slab in a household. When toilets are pit latrines with no slab and open pit, a household has an unimproved toilet facility. Both factors have binary variables, with one indicating improved toilet facilities and zero indicating no improvement. In the same way, the protected water supply takes one and zero if not protected.

The study also includes information on the household's socio-economic and demographic features, such as education, wealth, age, and residence. Maternal education is a significant driver of health outcomes. The mother's education level ranged from no education, primary, secondary, and higher education. We ranked the mother's wealth into three categories: poor, moderate, and rich. However, because the NDHS does not collect direct information on family income, the wealth index is calculated using the household's assets. The presence of power, a car, a motorcycle, a refrigerator, a radio, a television, and the type of flooring were all considered assets (Rutstein & Johnson 2004). The NDHS divided the wealth index into five quintiles, from low to high. The poor and poorest were combined in this study to form the contemporary poor. Again, in this study, the rich group comprised of wealthy and most affluent merged. Four categories for the mother's age were: under 20, 20–29, 30–39, and 40–49. Types of marital status were: never married, married, living with a partner, widowed, divorced, and no longer living together/separated.

The total number of household members is a continuous variable that ranged from 1 to 33 in 2006 and to 25 in 2013. In both years, the number of children aged 5 or younger ranged from zero to eight. Other explanatory variables for the Namibian setting were locations (urban, rural) and geographical regions (Caprivi, Erongo region) as dummy variables. Because of the country's severely unequal distribution of resources, location issues are essential in Namibia (MOHSS et al. 2014).

Biological, maternal, and dietary aspects were also taken into account, as shown in Table 1. In the previous 5 years, the number of births was in three categories: one, two, and three or more. First, second, third, or higher were birth order categories in this study. In addition, the gender of a child and breastfeeding status also accounted for child mortality. Finally, the study added the HIV variable in the 2013 model.

The acronym CS stands for child survival, and it refers to whether or not the child was alive at the time of the survey. The answer to this question is either yes or no. Mother's qualities as MC include her age, education, wealth, and marital status. Mother's age is in categories, and less than 20s is the reference group. The classes of mothers’ education: uneducated, primary education, secondary, and higher education, as the reference group.

The second model includes variables relevant to Namibia. The PR vector denotes the location of residence as well as geographical location, covering 14 political regions. The reference group used in this analysis is the Caprivi region. Additionally, it includes NH and NC vectors, which present the total number of household members and the total number of children under five, respectively, as continuous variables. Furthermore, in 2013, Model 2 added the mother's HIV results to determine the influence of a mother's HIV status on her child's health; the reference group is the negative results. Lastly, Models 1 and 2 are both estimated separately for 2006 and 2013.

We investigated the impact of drinking water sources, toilet facilities, mother's characteristics and socio-economic factors, and proximate factors on the health of children under the age of 5 in Namibia. Table 2 shows the descriptive statistics of the 2006 and 2013 surveys. The majority of residents had unimproved toilet facilities. In 2006, 65% of Namibian households had unimproved toilet facilities, and 61% in 2013. In both the 2006 and 2013 surveys, 87% of Namibian respondents had access to a protected water supply.

The proportion of uneducated mothers in Namibia was 12 and 8% in 2006 and 2013, respectively. The number of mothers who completed secondary education increased from 53% in 2006 to 63% in 2013. In both years, the percentage of mothers with a higher degree fell below 5%. In 2006, 40% of the population was impoverished compared with 44% in 2013. The wealthiest group accounted for 34% of all households in both years, while the rest were in the middle. Half of the mothers were between the ages of 20 and 29, with the lowest proportion between 20 and 40. The highest share of mothers in both surveys were never married, followed by married and living with a partner.

In terms of location, the vast majority of respondents lived in rural areas. In 2006, almost 63% of households lived in rural areas, and this percentage decreased to around 55% in 2013. The observed decrease in the percentage of rural households could be attributable to increased rural-urban migration observed in the country during the last two decades. Despite the observed increase in urbanization, Namibia is still mostly rural, with around four in ten residents living in urban regions (MOHSS et al. 2014).

Table 2 reveals that 60% of women had one child 5 years before the 2006 survey, whereas 34% had two children. Only 6% of mothers had three or more children in the last 5 years. In 2013, these percentages remained unchanged. Furthermore, a third of respondents had their first birth order at the time of the survey. In contrast, 44% of respondents have a birth order of three or higher. Males made up 51% of children under the age of 5 in 2006. In 2006, 41.5% of mothers were breastfeeding, compared with 38.8% in 2013. Finally, in 2006, households did not respond to the HIV status-related questions. In 2013, 2,873 participants responded to the inquiry about recent HIV results, with the majority of them having negative HIV results.

Table 3 illustrates the results of estimating model 1 for the years 2006 and 2013. In both years, the study discovered that the source of drinking water had no substantial impact on the child's health. The wide distribution of safe water supply and widespread usage of water treatment tablets in the country's households contributes to the insignificance of the drinking water source to child health (Kgabi et al. 2014).

Similar research in Egypt and 70 middle-income countries (Abou-ali 2003; Fink et al. 2011) corroborates our findings. According to the Egyptian study, lower childhood mortality did not statistically connect with access to municipal water, but it was significantly associated with lower infant mortality (Abou-ali 2003). Fink et al. (2011) confirmed the favourable effect associated with access to safe water supply on children aged 1–12 months while adjusting parental and household variables. As a result, access to safe water during the weaning process of an infant could be a crucial determinant of child survival and health, significantly if socio-economic, biological, and maternal characteristics are not improved (Abou-ali 2003).

Furthermore, this study found a significant association between the type of toilet facilities and child health. The results of the 2013 survey revealed that children in homes with unimproved toilet facilities were 80% more likely to die than those with improved toilet facilities (p=0.011). A study conducted using demographic and health data from 70 low and middle-income countries during 1986–2007 found that children who live in a home with improved toilet facilities were associated with a 23% lower risk of dying before reaching five (Fink et al. 2011). Namibia invested more resources into improving and expanding population access to safe drinking water than improved toilet facilities. By 2013, 61% of Namibian households did not have improved toilet facilities, and roughly 87% of Namibian households had access to safe water (MOHSS et al. 2014). These discrepancies could explain the results of this research.

In terms of a mother's socio-economic status, the study found that in 2013, a mother's education was crucial for her child's health. For example, compared with mothers with higher education, the probabilities of under-five children's deaths are 3.6 and 3.4 times higher for illiterate mothers and mothers with secondary education, respectively. Furthermore, mothers with primary education have a 4.4 times higher risk of having a child under the age of 5 die than mothers with a higher level of education. A study from Ghana supports our findings that women with a junior and secondary education were 45 and 76% less likely to experience child deaths, respectively (Kanmiki et al. 2014). In 2006, the mother's wealth was critical to the child's health. Rich mothers have a 45% lower risk of under-five child mortality than poor mothers, with a p-value of 0.023. Evidence from a study conducted in Namibia revealed that a unit increase in wealth was associated with a 0.052 and 0.301 increase in the mother's BMI in the 5th and 90th percentiles, respectively (Amugsi et al. 2016). Although this does not imply that educated mothers make better nutritional choices for their children, it suggests a greater chance of doing so. Finally, the mother's age and marital status are unimportant in the 2006 and 2013 analyses.

Breastfeeding is a crucial factor in determining a child's survival and health. According to our findings, if a mother is not breastfeeding, she is seven and five times more likely to have a child die before the age of 5 in 2006 and 2013, respectively. Furthermore, the number of births in the previous 5 years is a significant determinant for U5M. In 2006 and 2013, mothers who had only one child in the last 5 years had 84 and 91% less probability than mothers with three or four children. A positive association exists between multiple gestations with child mortality in low- and middle-income nations, especially during infancy (Owais et al. 2013; Kayode et al. 2015). Multiple pregnancies are commonly associated with poverty, contributing to stillbirths and premature deliveries, risk factors of under-five child mortality (Kayode et al. 2015). The rate of newborn fatalities in Bangladesh is 29 per 1,000 live births, with the leading causes being asphyxia (35%), sepsis (28%), and preterm births (19%) (Owais et al. 2013). In addition, contributing factors to infant mortality worth mentioning are hypoglycaemia and hypothermia, which are difficult to manage in low-resource settings (Kibria et al. 2018).

This study also revealed that inappropriate birth spacing and order are associated with the risk of child mortality. In Ghana, inadequate birth spacing negatively relates to child survival; children with insufficient spacing had 2.5 times the chance of dying during the newborn stage (Kayode et al. 2015). In addition, primigravida contributes to an increased risk of infancy mortality rate, which decreases over the second to fourth years of life (Kibria et al. 2018). Incredibly, the probability of child mortality rises again from the fifth infant onwards; however, birth spacing and mother's education could significantly reduce these deaths (Kibria et al. 2018).

Table 4 shows the results of estimating model 2. Model 2 includes the household location, recent HIV results, the number of households, and children. The findings of model 2 reveal that the source of drinking water does not affect children's health. In 2013, the type of toilet facilities was still significant. The odds of U5M for women who do not have access to improved toilet facilities is 1.65, with a p-value of 0.041. The result predicts that children die more in households with unimproved toilet facilities than those with improved toilets.

In 2013, the impact of a mother's education on her child's health was significant. Compared with mothers with a higher level of education, the odds for uneducated and low-educated mothers is roughly four. A study from SSA confirms that children of mothers who received formal education had an 80% chance of surviving until the fifth birthday compared with those whose mothers have not been to school (Bado & Susuman 2016). This paper revealed that mothers between the ages of 20 and 29 have a 45% lower probability of under-five child death than mothers under 20. Furthermore, adolescent mothers (18) and mothers over 35 had a higher likelihood of U5M. However, there is no apparent danger associated with older mothers, especially with proper birth spacing patterns (Owais et al. 2013). Children born to older mothers with a 2-year gap between births were less likely to die (Kibria et al. 2018). Instead, children of grand multiparous mothers with poor health-seeking habits are more likely to die (Kayode et al. 2015). Throughout the study, the mother's marital status did not influence the child's health.

According to a 2006 model, male children have a 41% probability of dying before reaching 5 years compared with a female child (p=0.009). Despite this, there is no evidence of child deaths in 2013 based on gender disparity. According to the study conducted in Afghanistan from the 2015 DHS data, 52% of 266 children that died were boys. The same study predicts boys neonates were 0.5% more likely to die than girls (Kibria et al. 2018). Breastfeeding is vital to a child's health, as previously stated. In 2006 and 2013, mothers who do not breastfeed had 5.2 and 3.7 increased chances of having a child die under the age of 5 than mothers who do. In addition, there is an association between breastfeeding and a mother's nutrition status. A study in Bangladesh found that proper nutrition for mothers during pregnancy or breastfeeding correlated with a 0.4 reduction in newborn deaths (Owais et al. 2013).

Table 4 shows a similar odds ratio of births in the previous 5 years in 2006 and 2013. Compared with mothers who had three or more babies in the last 5 years before the study, mothers who had only one birth had a 96% lower likelihood of experiencing U5M. At the same time, mothers have an 80% lower chance of U5M if they had two births in the last 5 years. Finally, in 2006 and 2013, the birth order had no significance.

Model 2 investigated the mother's region and the type of location (urban/rural) to child health. In Namibia, the mother's location (urban versus rural) was significantly (p=0.04) linked to U5M rates. First, according to the 2013 model, mothers who reside in cities are 32% less likely to experience U5M than mothers who live in villages. Second, according to the 2006 model, mothers from the Khomas, Oshana, and Oshikoto regions have a lower risk of U5M than mothers from the Caprivi region (Zambezi Region). In addition, according to the 2013 model, mothers from the Hardap, Omusati, and Oshana regions were 50% less likely to experience U5M (p=0.05) than mothers from the Zambezi region. Furthermore, with a p-value of 0.038, mothers in the Otjozondjupa region have a 58% lower risk of U5M.

Socio-economic inequalities between the Zambezi and Otjozondjupa, Hardap, Omusati, Oshana, Khomas, and Oshikoto regions could explain the observed differences in the risk of U5M rate. The Zambezi region is among the least developed in the country, with a low literacy rate. Seasonal flooding and high teenage pregnancies are also common in the region (Maemeko et al. 2018). Furthermore, a considerable proportion of the population in the Zambezi region has no access to protected water sources (26.8%) and improved toilet facilities (73.5%) (National Planning Commission 2011). In addition, the Zambezi region has one of the lowest poverty inequality indexes in the world (0.493) (Knoema 2011) and an HIV/AIDS prevalence rate of 237 per 1,000 people (MOHSS et al. 2014). The region's high-risk ratio of U5M is attributable to the interplay of factors mentioned above. Therefore, improving access to safe water sources and sanitation is necessary to reduce the risk factors of U5M in the Zambezi region.

Furthermore, the odds of children aged 5 and under 5 in homes show that every one-child increase in the household reduces the likelihood of U5M roughly by 64 and 61% in 2006 and 2013, respectively. Furthermore, the odds of under-five death increases with every one-member increase in the household by 14 and 13% in 2006 and 2013, respectively.

The result of the last column of the 2013 model demonstrate that HIV-positive mothers experience a 2.3 times higher probability of U5M than HIV-negative mothers. After adding HIV to the model, a mother's education is no longer significant, and there is no evidence of the impact of a mother's educational level on child health. However, other variables such as breastfeeding, number of children in the last 5 years, number of under-fives, and members in the household remained significant.

In summary, there is a need for health programmes that target health determinants connected to child health and survival in Namibia. As revealed by both models, increased access to improved toilet facilities is a crucial determinant of child survival in Namibia. In particular, to Namibia, access to safe drinking water has no impact on children's health. The Namibian government's efforts in enhancing access to safe water sources are praiseworthy, as the findings of this study illustrate. Generally, access to safe water is significant to child health. To accomplish the SDG's target of reducing under five-related death rates, the government must address health inequalities among the different regions in Namibia. Furthermore, we recommend strengthening health promotion and education programmes on breastfeeding, a suitable child spacing pattern, maternal age, and mother's education. Finally, HIV-positive, multiparous, adolescent, and older mothers should be encouraged to regularly seek health advice during the gestational and neonatal period to reduce the risk of under-five child mortality.

The authors express their gratitude to the MOHSS in Namibia for this study's demographic and health data.

M.A. conceptualized the idea, constructed the work's methodology and data analysis. The introduction, discussion, and conclusion of this paper were the work of A.A. and N.H. To finalize the manuscript, all authors contributed to the final manuscript's editing and revision. All authors read and approved the current version of the work.

The authors state that they have no conflicts of interest. This work received no financial support from any institution.

Data cannot be made publicly available; readers should contact the corresponding author for details.