Abstract

In this paper, we investigate the relationship between household drinking water quality and irrigation and child nutrition using primary household survey data and microbiological water sample testings in two rural districts of Ethiopia. Anthropometric measures such as height-for-age z-scores (HAZ), weight-for-age z-scores (WAZ), and weight-for-height z-scores (WHZ) were used to measure stunting, underweight, and wasting, respectively. Our survey results show that 41% of the children are stunted, 26% underweight, and 8% wasted. More than 58% of household's stored drinking water samples were also contaminated with Escherichia coli bacteria. The multivariate regression results suggest that irrigation farming and on-premises water sources are significantly associated with lower HAZ, while uncontaminated household stored drinking water quality is correlated with higher WAZ. The results also reveal that dietary diversity score and the number of antenatal care visits by the primary caretaker are statistically significant predictors of child nutritional status. These findings, however, cast doubt on the hypothesis that irrigated agriculture exclusively has a positive effect on child nutrition outcomes.

HIGHLIGHTS

  • Using a primary household survey data, we discuss interlinkages between WATSAN and irrigation agriculture and their association with child nutrition in two rural districts of Ethiopia where multiple-use of water is common.

  • Drinking water quality is determined by testing the microbiological quality of stored household water samples.

  • Irrigation farming and on-premises drinking water sources are associated with higher odds of child stunting by 1.7 and 2.9, respectively. On the other hand, uncontaminated household's stored water quality is associated with lower odds of child underweight.

  • 58% of the household's stored drinking water samples were contaminated with E. coli.

  • Among irrigating households, more than half of them reported that they withdraw water from irrigation water sources for domestic purposes, such as drinking, cooking, and washing.

INTRODUCTION

According to the 2017 Global Food Policy Report, 155 million under-five children are stunted (a low height-for-age), 52 million children are wasted (a low weight-for-height), and 815 million people are chronically undernourished. Poor nutrition contributes to approximately 45% of all deaths of under-five children globally (IFPRI 2017). More than 50% of child undernutrition is due to both repeated diarrhea and intestinal infections because of a lack of improved water supply and sanitation (WATSAN) and insufficient hygiene practices (Prüss-Üstün et al. 2008). In Ethiopia, the problem of child malnutrition is among the highest in the world, for example 38% of under-five children are stunted (an indicator of chronic malnutrition), and 18% are severely stunted in 2016 (CSA & ICF International 2017), most of these children are living in rural areas (a child is considered as severely stunted if his/her height-for-age z-score is below –3 standard deviations below the mean on the WHO Child Growth Standards).

Child nutritional outcomes (hereafter child nutrition) are good indicators of a child's health. The first two years of children's life are essential for their optimal growth and development. Child nutrition is governed mainly by access to WATSAN, healthcare access and utilization, adequate diet, and care (UNICEF 1990; CSA & ICF International 2012; Ali 2019), all of which are inadequate in rural Ethiopia. Undernutrition can impair long-term cognitive development, school performance, and labor productivity (Humphrey 2009). Malnourished children also have lower resistance to diseases and are more susceptible to suffer from subsequent diarrheal and other infections (UNICEF 2013). For instance, a severely undernourished child (low weight-for-age) is almost ten times more likely to die from diarrhea compared to the average (Black et al. 2008).

Access to improved WATSAN is vital for improved child health and growth. Infections and repeated diarrheal diseases are one of the major causes of undernutrition (Prüss-Üstün et al. 2008; Dewey & Mayers 2011). Recent studies also suggest that environmental enteropathy, a disorder of the small intestine, significantly contributes to child undernutrition (Lin et al. 2013; Mara 2017). Environmental enteropathy is caused by a lack of adequate sanitation and proper hygiene practices and damages the wall of the small intestine; consequently, the capacity to absorb energy and nutrients is reduced, and they are diverted from growth to the immune system to fight the infection (Humphrey 2009; Mara 2017). So, preventing diarrheal and other infectious diseases through improving WATSAN can break the cycle and result in improved child nutrition.

WATSAN is closely linked with agriculture since domestic and irrigation water sources are often interchangeable in rural communities (Usman & Gerber 2019), with potential health risks. Irrigation can affect health and nutrition through a variety of pathways (see Figure 1). For example, poor irrigation practices could bring adverse impacts on the environment and human health through increasing water-related diseases and domestic water contamination (Usman & Gerber 2019). On the other hand, irrigation water may reduce the burden of water collection time (Usman & Gerber 2019). Further, irrigation allows households to grow more and diversified crops. A study of child nutrition in Kenya found evidence of higher energy intakes and lower chronic malnutrition in children where communities have access to irrigation as compared to communities without access (Kirogo et al. 2007). Despite its potential for improved food and nutrition security in sub-Saharan Africa, the overall effect of irrigation on health and nutrition is difficult to know a priori (see the review by Domenech & Ringler 2013).

Figure 1

The agriculture-WATSAN system: an integrated framework with health and nutrition links (Source: Gerber et al. 2019).

Figure 1

The agriculture-WATSAN system: an integrated framework with health and nutrition links (Source: Gerber et al. 2019).

Although previous studies show that poor WATSAN services are a cause for poor child health, empirical studies and knowledge of the nexus between WATSAN, irrigation, and child nutrition are lacking (Gerber et al. 2019). This study, therefore, investigates the relationship between household drinking water quality, irrigation, and child nutrition in rural areas of Ethiopia. The paper makes two contributions. First, since multiple-use of water is common in the study areas, we take irrigation practice into account in the multivariate regression analysis. Second, drinking water quality is determined by testing the microbiological quality of stored household water. Identifying the principal risk factors that influence child nutrition will help policymakers and practitioners to design the right intervention to improve child health and well-being. The main findings of the paper are that contrary to our expectations, having the drinking water source on-premises and engaging in irrigated agriculture are significantly associated with inferior child nutrition in the study areas.

MATERIALS AND METHODS

Household sample selection

For this analysis, a household survey was conducted in Fogera and Mecha districts of Ethiopia between February and March 2014. The survey covers 454 households with 562 under-five children. The sample households were selected using a stratified two-stage cluster sample design. In the first stage, 61 villages (the primary sampling units) were selected randomly from 20 identified kebeles (the lowest geographic administrative division in the country). In the second stage, 454 households were selected based on a systematic random sampling method. The selected districts are shaded, and the black dots represent the sampled households (see Figure 2).

Figure 2

Map of the study districts (Source: Usman & Gerber 2019).

Figure 2

Map of the study districts (Source: Usman & Gerber 2019).

Microbiological water quality tests

One of the most commonly used indicators for microbial water quality is the level of Escherichia coli (E. coli) bacteria, which only comes from human and animal feces and is associated with health risks, such as childhood diarrhea (Usman et al. 2019). In this study, E. coli concentrations are reported as colony-forming units per 100 mL (CFU/100 mL) of the water sample. The general WHO drinking water quality guideline suggests that the number of E. coli CFU/100 mL water sample should be zero for the water supply to be potable. Water samples were collected from the households' stored drinking water, and the samples were immediately placed into the portable test kit (a product of Wagtech WTD, UK) on-site and incubated for a maximum of 24 hours at 44 °C to detect the level of E. coli concentration.

Anthropometric measures

Information on the height and weight of all under-five children were collected to evaluate their nutritional status. Our measurements were carried out by the local health extension workers: following the DHS standards, children younger than 24 months were measured lying down on a board, and older children were measured while standing. The weight of young children was obtained by subtracting his/her mother's weight from their combined weight. A total of 562 under-five children were eligible to be weighed and measured; however, data are presented for 547 of these children for weight-for-age, and 479 children for height-for-age, and weight-for-height measurement indices. The sample for height-for-age is small because the height measurements for most children under six months old were not obtained. These children were afraid of lying down on the measuring board. However, we estimated the models for children aged 6–59 months only, but the results are the same.

The study protocol was approved by the Medical Ethics Commission of the Federal Republic of Germany (Arbeitskreis Medizinischer Ethikkommissionen in der Bundesrepublik Deutschland). In compliance with ethical considerations, selected households consented about their willingness to participate in the study, and a signed consent form was collected.

Nutritional outcomes

Anthropometric data on height and weight can be used as indicators of under-five child nutrition. Weight and height/length measurements were converted to weight-for-age z-scores (WAZ), height-for-age z-scores (HAZ), and weight-for-height z-scores (WHZ) for each child to form the standardized measures. The anthropometric indicators of children were calculated based on the new child growth standards released in 2006 by the WHO (2006). The z-scores are generated using the zscore06 Stata package (Leroy 2011). The z-scores indicate the number of standard deviations (SD) a child is below or above the median child. Child nutrition indicators such as height-for-age (stunting), weight-for-age (underweight), and weight-for-height (wasting) z-scores were computed. To be identified as stunted, underweight, or wasted, the child is at least –2 SD below the WHO child growth standards.

RESULTS AND DISCUSSION

Table 1 presents the descriptive statistics of the main variables for this study. The variable antenatal care (ANC) indicates the number of visits from a health professional during the pregnancy period, and the dietary diversity score measures the number of food groups (out of six food groups) consumed in the last 24 hours preceding the survey (the food groups include milk & milk byproducts, grains, vegetables, fruits, meat, and egg). According to Table 1, about 26% of children are underweight, 41% stunted, and 8% wasted. The anthropometric indicators against the WHO child growth standards are also shown in Figure A1 in the Supplementary material.

Table 1

Summary statistics

VariablesDescriptionNMeanS.D.
Outcome variables 
Stunted 1 = yes; 0 otherwise 479 0.40  
Height-for-age z-scores  479 –1.71 1.20 
Underweight 1 = yes; 0 otherwise 547 0.26  
Weight-for-age z-scores  547 –1.32 1.00 
Wasted 1 = yes; 0 otherwise 479 0.08  
Weight-for-height z-scores  479 –0.62 0.99 
Water, sanitation, and hygiene 
Water quality (1 = no E. coli1 = uncontaminated stored water quality; 0 otherwise 454 0.42  
Drinking water source 1 = protected; 0 otherwise 454 0.50  
Water source location 1 = own yard; 0 otherwise 454 0.16  
Pit latrine 1 = yes; 0 otherwise 454 0.42  
Under-5 children characteristics 
Child age Age in months 562 29.02 16.30 
Child sex 1 = male; 0 = female 565 0.46  
Under-five years children Number of children under-five 454 1.24 0.45 
Dietary diversity score The mean food groups consumed 505 1.84 1.04 
Antenatal care (ANC) Number of antenatal-care visits 558 1.93 1.51 
Delivery with a health professional 1 = yes; 0 otherwise 561 0.16  
Household characteristics 
Irrigation practice 1 = yes; 0 otherwise 454 0.67  
Livestock holdings Livestock holding in Tropical Livestock Units 454 3.97 1.87 
Asset values Total assets value in 1,000 Birra 454 5.88 6.20 
Primary caretaker age Age of primary caretaker in years 453 30.29 6.73 
Primary caretaker literacy 1 = yes; 0 otherwise 453 0.09  
Household head education Household head completed years of education 454 1.02 1.96 
Household size Number of household members 454 5.98 1.77 
VariablesDescriptionNMeanS.D.
Outcome variables 
Stunted 1 = yes; 0 otherwise 479 0.40  
Height-for-age z-scores  479 –1.71 1.20 
Underweight 1 = yes; 0 otherwise 547 0.26  
Weight-for-age z-scores  547 –1.32 1.00 
Wasted 1 = yes; 0 otherwise 479 0.08  
Weight-for-height z-scores  479 –0.62 0.99 
Water, sanitation, and hygiene 
Water quality (1 = no E. coli1 = uncontaminated stored water quality; 0 otherwise 454 0.42  
Drinking water source 1 = protected; 0 otherwise 454 0.50  
Water source location 1 = own yard; 0 otherwise 454 0.16  
Pit latrine 1 = yes; 0 otherwise 454 0.42  
Under-5 children characteristics 
Child age Age in months 562 29.02 16.30 
Child sex 1 = male; 0 = female 565 0.46  
Under-five years children Number of children under-five 454 1.24 0.45 
Dietary diversity score The mean food groups consumed 505 1.84 1.04 
Antenatal care (ANC) Number of antenatal-care visits 558 1.93 1.51 
Delivery with a health professional 1 = yes; 0 otherwise 561 0.16  
Household characteristics 
Irrigation practice 1 = yes; 0 otherwise 454 0.67  
Livestock holdings Livestock holding in Tropical Livestock Units 454 3.97 1.87 
Asset values Total assets value in 1,000 Birra 454 5.88 6.20 
Primary caretaker age Age of primary caretaker in years 453 30.29 6.73 
Primary caretaker literacy 1 = yes; 0 otherwise 453 0.09  
Household head education Household head completed years of education 454 1.02 1.96 
Household size Number of household members 454 5.98 1.77 

Source: Authors' computation based on own survey data.

aThe exchange rate during the time of the survey was 1 Euro = 26.02 Ethiopian Birr. Livestock is excluded in computing the household's assets values.

Descriptive analysis

Reported in Table 2 is the prevalence of child malnutrition by WATSAN and child characteristics. It is interesting that the variation in child malnutrition based on access to improved water sources, on types of drinking water sources, and access to pit latrine is not significant. There is, however, a difference in the prevalence of child malnutrition by the household's drinking water quality. Children from households whose stored water is contaminated with E. coli are more likely to be wasted than their counterparts. Finally, children of irrigating households are more likely to be stunted or underweight than children of non-irrigating households.

Table 2

Prevalence of malnutrition by WATSAN and child characteristics

VariablesObs.Stunting (%)P-valuesWasting (%)P-valuesObs.Underweight (%)P-values
Age of child in months 
 6–11 47 19.15 0.002 8.51 0.018 55 23.64 0.829 
 12–23 98 34.69  16.33  100 31.00  
 24–35 123 49.59  4.88  130 26.15  
 36–47 120 42.50  5.83  122 30.33  
 48–59 91 43.96  6.59  92 29.35  
Gender of a child 
 Male 223 43.05 0.331 8.97 0.556 252 29.37 0.165 
 Girl 257 38.50  7.39  295 23.73  
Water source type based on WHO 
 Improved 235 37.45 0.147 8.94 0.516 267 27.34 0.535 
 Unimproved 245 43.67  7.35  280 25.36  
Drinking water source 
 Private-protected dug well 26 46.15 0.213 7.69 0.854 30 33.33 0.365 
 Shared-protected well/spring 209 36.45  9.09  237 26.58  
 Unprotected well/spring 191 41.88  7.85  218 27.06  
 Surface water 54 50.00  5.56  62 19.35  
Stored water quality 
 Contaminated 289 41.52 0.681 10.03 0.062 321 28.97 0.111 
 Uncontaminated 191 39.27  5.24  226 22.57  
Pit latrine 
 Yes 188 37.23 0.237 6.91 0.415 222 25.68 0.838 
 No 292 42.81  8.90  325 26.77  
Irrigated agriculture 
 Yes 323 45.20 0.004 9.29 0.188 368 30.42 0.002 
 No 157 31.21  5.73  179 17.88  
VariablesObs.Stunting (%)P-valuesWasting (%)P-valuesObs.Underweight (%)P-values
Age of child in months 
 6–11 47 19.15 0.002 8.51 0.018 55 23.64 0.829 
 12–23 98 34.69  16.33  100 31.00  
 24–35 123 49.59  4.88  130 26.15  
 36–47 120 42.50  5.83  122 30.33  
 48–59 91 43.96  6.59  92 29.35  
Gender of a child 
 Male 223 43.05 0.331 8.97 0.556 252 29.37 0.165 
 Girl 257 38.50  7.39  295 23.73  
Water source type based on WHO 
 Improved 235 37.45 0.147 8.94 0.516 267 27.34 0.535 
 Unimproved 245 43.67  7.35  280 25.36  
Drinking water source 
 Private-protected dug well 26 46.15 0.213 7.69 0.854 30 33.33 0.365 
 Shared-protected well/spring 209 36.45  9.09  237 26.58  
 Unprotected well/spring 191 41.88  7.85  218 27.06  
 Surface water 54 50.00  5.56  62 19.35  
Stored water quality 
 Contaminated 289 41.52 0.681 10.03 0.062 321 28.97 0.111 
 Uncontaminated 191 39.27  5.24  226 22.57  
Pit latrine 
 Yes 188 37.23 0.237 6.91 0.415 222 25.68 0.838 
 No 292 42.81  8.90  325 26.77  
Irrigated agriculture 
 Yes 323 45.20 0.004 9.29 0.188 368 30.42 0.002 
 No 157 31.21  5.73  179 17.88  

Source: Authors' estimates using own survey data. The p-values generated from the two-side t-test in mean difference.

Multivariate regression results

Reported in Table 3 are estimates of the odds of a child being stunted, underweight, and wasted, while the ordinary least squares (OLS) results based on the z-scores are summarized in Table 4. Robust standard errors adjusted for clustering at the village level are reported for each model. We include constant terms and district effects in all regressions. The explanatory power of the model ranges from 0.09 to 0.30 for nutritional outcomes.

Table 3

Multivariate regression results for stunting, underweight, and wasting

(1)(2)(3)
Explanatory variablesStunting: Yes = 1Underweight: Yes = 1Wasting: Yes = 1
Water quality (No E. coli, Yes = 1) 0.884 [0.23] 0.558** [0.16] 0.403* [0.19] 
Water source on-premises, Yes = 1 2.883** [1.21] 1.121 [0.37] 0.955 [0.55] 
Pit latrine, Yes = 1 1.024 [0.28] 1.816** [0.51] 1.509 [0.68] 
Irrigating household, Yes = 1 1.807** [0.46] 2.933*** [0.86] 2.583** [1.14] 
Livestock holding in TLUs 1.129 [0.13] 0.801** [0.08] 0.753* [0.11] 
Dietary diversity scores 0.214*** [0.04] 0.384*** [0.07] 0.410*** [0.12] 
Delivery with a health professional, Yes = 1 0.371*** [0.13] 0.260*** [0.10] 0.096** [0.11] 
Number of ANC visits 0.710*** [0.06] 0.685*** [0.05] 0.894 [0.12] 
Log (assets value) 0.879 [0.13] 0.660*** [0.09] 0.607 [0.19] 
Child age in months 1.124*** [0.04] 1.157*** [0.04] 0.944 [0.05] 
Child age square (/100) 0.859*** [0.05] 0.811*** [0.04] 1.037 [0.09] 
Child is male 1.312 [0.29] 1.576** [0.36] 1.572 [0.67] 
Primary caretaker age 0.953* [0.03] 0.975 [0.02] 1.011 [0.04] 
Household head education 0.899 [0.06] 1.219*** [0.07] 1.284** [0.13] 
Household size 1.183 [0.12] 1.052 [0.11] 1.107 [0.18] 
Observations 472 540 472 
Pseudo R-squared 0.30 0.24 0.21 
Model Chi2 126.70 151.0 48.99 
Model p-value 0.000 0.000 0.000 
(1)(2)(3)
Explanatory variablesStunting: Yes = 1Underweight: Yes = 1Wasting: Yes = 1
Water quality (No E. coli, Yes = 1) 0.884 [0.23] 0.558** [0.16] 0.403* [0.19] 
Water source on-premises, Yes = 1 2.883** [1.21] 1.121 [0.37] 0.955 [0.55] 
Pit latrine, Yes = 1 1.024 [0.28] 1.816** [0.51] 1.509 [0.68] 
Irrigating household, Yes = 1 1.807** [0.46] 2.933*** [0.86] 2.583** [1.14] 
Livestock holding in TLUs 1.129 [0.13] 0.801** [0.08] 0.753* [0.11] 
Dietary diversity scores 0.214*** [0.04] 0.384*** [0.07] 0.410*** [0.12] 
Delivery with a health professional, Yes = 1 0.371*** [0.13] 0.260*** [0.10] 0.096** [0.11] 
Number of ANC visits 0.710*** [0.06] 0.685*** [0.05] 0.894 [0.12] 
Log (assets value) 0.879 [0.13] 0.660*** [0.09] 0.607 [0.19] 
Child age in months 1.124*** [0.04] 1.157*** [0.04] 0.944 [0.05] 
Child age square (/100) 0.859*** [0.05] 0.811*** [0.04] 1.037 [0.09] 
Child is male 1.312 [0.29] 1.576** [0.36] 1.572 [0.67] 
Primary caretaker age 0.953* [0.03] 0.975 [0.02] 1.011 [0.04] 
Household head education 0.899 [0.06] 1.219*** [0.07] 1.284** [0.13] 
Household size 1.183 [0.12] 1.052 [0.11] 1.107 [0.18] 
Observations 472 540 472 
Pseudo R-squared 0.30 0.24 0.21 
Model Chi2 126.70 151.0 48.99 
Model p-value 0.000 0.000 0.000 

Notes: Robust standard errors adjusted for clustering at the village level are in square brackets.

Statistical significance denoted at *p <0.10, ** p < 0.05, *** p < 0.01.

Reported coefficients are odds ratios. We checked the robustness of these results to the exclusion of children below 6 months old due to length measurement issues, but the results did not change (Table A3 in the Supplementary material).

Table 4

Multivariate regression results for the standardized z-scores

(1)(2)(3)
Height-for-Age z-scoresWeight-for-Age z-scoresWeight-for-Height z-scores
Water quality (No E. coli, Yes = 1) 0.062 [0.11] 0.180** [0.09] 0.194* [0.10] 
Water source on-premises, Yes = 1) –0.486*** [0.14] –0.173 [0.12] 0.125 [0.16] 
Pit latrine, Yes = 1 0.013 [0.12] 0.019 [0.10] –0.004 [0.13] 
Irrigating household, Yes = 1 –0.283** [0.13] –0.293*** [0.10] –0.147 [0.11] 
Livestock holding in TLUs –0.038 [0.05] –0.009 [0.04] 0.052 [0.04] 
Dietary diversity score 0.418*** [0.06] 0.266*** [0.05] 0.159** [0.06] 
Delivery with a health professional, Yes = 1 0.307** [0.14] 0.104 [0.11] 0.053 [0.14] 
Number of ANC visits 0.146*** [0.04] 0.113*** [0.03] 0.040 [0.03] 
Log (assets value) –0.017 [0.07] 0.127** [0.05] 0.187*** [0.07] 
Child age in months –0.049*** [0.02] –0.060*** [0.01] 0.012 [0.01] 
Child age square (/100) 0.071*** [0.02] 0.078*** [0.02] –0.019 [0.02] 
Child is male −0.230** [0.10] –0.148* [0.08] –0.038 [0.08] 
Primary caretaker age 0.017 [0.01] 0.008 [0.01] 0.002 [0.01] 
Household head education 0.037 [0.03] –0.004 [0.03] –0.006 [0.03] 
Household size –0.031 [0.04] –0.004 [0.03] –0.026 [0.03] 
Observations 472 540 472 
R-squared 0.25 0.23 0.09 
Model F-stat 13.00 8.35 3.09 
Model p-value 0.000 0.000 0.001 
(1)(2)(3)
Height-for-Age z-scoresWeight-for-Age z-scoresWeight-for-Height z-scores
Water quality (No E. coli, Yes = 1) 0.062 [0.11] 0.180** [0.09] 0.194* [0.10] 
Water source on-premises, Yes = 1) –0.486*** [0.14] –0.173 [0.12] 0.125 [0.16] 
Pit latrine, Yes = 1 0.013 [0.12] 0.019 [0.10] –0.004 [0.13] 
Irrigating household, Yes = 1 –0.283** [0.13] –0.293*** [0.10] –0.147 [0.11] 
Livestock holding in TLUs –0.038 [0.05] –0.009 [0.04] 0.052 [0.04] 
Dietary diversity score 0.418*** [0.06] 0.266*** [0.05] 0.159** [0.06] 
Delivery with a health professional, Yes = 1 0.307** [0.14] 0.104 [0.11] 0.053 [0.14] 
Number of ANC visits 0.146*** [0.04] 0.113*** [0.03] 0.040 [0.03] 
Log (assets value) –0.017 [0.07] 0.127** [0.05] 0.187*** [0.07] 
Child age in months –0.049*** [0.02] –0.060*** [0.01] 0.012 [0.01] 
Child age square (/100) 0.071*** [0.02] 0.078*** [0.02] –0.019 [0.02] 
Child is male −0.230** [0.10] –0.148* [0.08] –0.038 [0.08] 
Primary caretaker age 0.017 [0.01] 0.008 [0.01] 0.002 [0.01] 
Household head education 0.037 [0.03] –0.004 [0.03] –0.006 [0.03] 
Household size –0.031 [0.04] –0.004 [0.03] –0.026 [0.03] 
Observations 472 540 472 
R-squared 0.25 0.23 0.09 
Model F-stat 13.00 8.35 3.09 
Model p-value 0.000 0.000 0.001 

Notes: Robust standard errors adjusted for clustering at the village level are in square brackets.

Statistical significance denoted at *p < 0.1, **p < 0.05, ***p < 0.01.

The regression results reported in Tables 3 and 4 indicate that household drinking water quality is positively correlated with child nutrition. For example, the weight of children living in households with uncontaminated stored drinking water quality is, on average, 0.18 SD higher compared to their counterparts (column 2, Table 4). Water source location is positively and significantly associated with child stunting (HAZ): children living in households with their own drinking water sources are 2.9 times more likely to be stunted than children living without their own water source (column 1, Table 3).

According to Table 3, in contrast to our expectation, children from irrigating households are 1.7, 3.0, and 2.4 times more likely to be stunted, underweight, and wasted, respectively, than children from non-irrigating households. Similarly, the height (weight) of children from irrigating households, on average, is 0.28 SD lower compared to children living in non-irrigating households (columns 1 and 2 of Table 4). On the other hand, livestock ownership is marginally associated with a lower risk of childhood underweight and wasting (columns 2 and 3 of Table 3). Livestock improves child nutrition by supplying richer animal proteins, such as cow's milk, which may be the only essential source of animal protein and micronutrients for children in rural communities (Hoddinott et al. 2015). Livestock holding, however, can reduce household water quality in rural areas (Usman et al. 2018), which could partly offset its positive impact on child nutrition. As expected, the dietary diversity score is significantly associated with improved child nutrition in both model specifications.

The associations between child nutrition and access to health facility indicator variables (the number of ANC visits and delivery with health professionals) are also statistically significant and positively associated with child nutrition (Tables 3 and 4). The effects of ANC visits appear large and robust, and this could mean that it captures the influence of other variables such as health knowledge of the primary caretaker or access to health services. Even though access to or utilization of health facilities is inadequate in the country, a study by Headey (2014) reported similar results that four or more ANC visits are significantly associated with improved HAZ. Children delivered with the help of health professionals are also 63, 74, and 90% less likely to be stunted, underweight, and wasted, respectively (Table 3). Of note, only 16% of children were delivered with the help of health professionals, which may drive the results.

Regarding the other control variables, household assets have a positive association with better nutrition outcomes. Male children are slightly more likely to be underweight than female children. For instance, the weight of male children, on average, is expected to be 0.23 SD lower than female children (Table 4). This finding is consistent with previous studies showing that girls have relatively better nutritional status than boys (Kabubo-Mariara et al. 2009; CSA & ICF International 2012). It is worth pointing that child nutrition is a result of the complex interaction of different variables, and WATSAN is one of the many ingredients for improved child growth outcomes.

Our results suggest that while uncontaminated drinking water quality is significantly associated with higher weight-for-age z-scores, having a drinking water source on-premises is consistently and significantly associated with stunting (HAZ). The latter is in contrast with an earlier finding that water collection time is negatively associated with improved child nutrition (Pickering & Davis 2012). It is assumed that the time saving from a water collection trip is used for other activities (e.g. homemaking, childcaring) that could improve child nutrition (e.g. Sorenson et al. 2011; van Houweling et al. 2012). The paradox may be partially attributable to the poor water quality of the source, for example households with a pit latrine and having water sources close to their dwelling have poor stored water quality. About 16% of the sampled households have a water source on-premises, of which 67% of the water sources are open wells. In this case, the result may not be surprising except that the effect size is large.

The impact of irrigated agriculture on child nutrition is counter-intuitive. Irrigated agriculture is likely to improve household food availability and nutrition since irrigation helps farmers to grow micronutrients and fresh vegetables (Kirogo et al. 2007). Irrigation also allows farmers to produce and sell high-value crops throughout the year and increases their income, which can enhance household nutrition (von Braun et al. 1989). Indeed, our survey shows that irrigating households (mainly involved in the production of rice, onions and tomatoes) reported higher income from agriculture than non-irrigating households, but the dietary diversity score was almost equal in both groups. In line with our findings, a recent study in Ethiopia shows that although irrigation increases production diversity, it does not translate into the consumption of higher dietary diversity (Passarelli et al. 2018). However, poor child nutrition among irrigating households could be partly explained by less time for childcare due to the high labor demand for irrigated agriculture (Usman & Gerber 2019). Typically, irrigated agriculture is more labor-intensive and demands more family labor than rain-fed agriculture. Alternatively, irrigation may create new sources of drinking water contamination that leads to poor child health (Gerber et al. 2019). Indeed, about 51% of the irrigating households use irrigation water for domestic purposes, such as drinking, cooking, and washing. So, the effects of irrigation on child nutrition seem inconclusive, highly complex, and context-specific.

So far, irrigation was measured as a dummy variable (whether a household practices irrigation farming) that did not capture the extent to which a household is engaged in irrigated agriculture. We then measured the variable as the proportion of irrigated land (Table A1). Again, irrigation is significantly and positively associated with stunting and underweight, but no longer with the z-scores. If irrigation affects child nutrition through water quality, we should find a higher proportion of undernourished children living in irrigating households who report using irrigation water for domestic purposes. To test this, we estimated the models for irrigating households of the sample (Table A2); however, the domestic use of irrigation water has only a limited effect on stunting. This suggests that the impact of irrigation infrastructures that influence child nutrition may not be captured fully at the household level.

Finally, this analysis would have benefited if the mother's nutrition and child's birth-weight information was available. For instance, children who are small at birth are also more likely to be malnourished than children who are average/larger at birth (CSA & ICF International 2012). Birth weights, however, were not known for many babies as most of them were delivered at home and not weighted at birth.

CONCLUSIONS

Although Ethiopia has made remarkable progress in reducing child malnutrition, a substantial portion of the population is still undernourished. The country also lags the rest of the world in terms of access to improved WATSAN services. Using primary household survey data, we explore the degree of association between WATSAN services, irrigation, and child nutrition in two rural districts of Ethiopia. Our findings reveal that preschool children in rural Ethiopia have a lag in HAZ and WAZ relative to the WHO child growth standards. The results explain a small portion of the variations in child nutrition, which is influenced by a range of interconnected variables. The study highlights the need to promote water treatment at the point-of-use and the protection of drinking water sources from external contamination. Moreover, providing diverse food groups that supply all the vital micronutrients to meet the requirements for optimal child growth is crucial for young children. Generally, solving the problem of child malnutrition requires a mix of instruments and working across various sectors (e.g. WATSAN, health, agriculture). For instance, providing improved drinking water without proper water treatment and fundamental behavioral changes will not work. Likewise, providing adequate nutrition without a safe WATSAN environment, then the latter will undo much of the gains from improved diets.

DATA AVAILABILITY STATEMENT

REFERENCES

Ali
M.
2019
Multisectoral approach to nutrition in Pakistan
.
J. Water Sanit. Hyg. Dev.
9
(
1
),
152
163
.
doi:10.2166/washdev.2018.122
.
Black
R. E.
Allen
L. H.
Bhutta
Z. A.
Caulfield
L. E.
de Onis
M.
Ezzati
M.
Child Undernutrition Study
G.
2008
Maternal and child undernutrition: global and regional exposures and health consequences
.
Lancet
371
(
9608
),
243
260
.
doi:10.1016/S0140-6736(07)61690-0
.
CSA & ICF International
2012
Ethiopia Demographic and Health Survey 2011
.
Central Statistical Agency (CSA) & ICF International
,
Addis Ababa
,
Ethiopia
and Calverton, Maryland, USA. Retrieved from: www.measuredhs.com/pubs/pdf/FR255/FR255.pdf.
CSA & ICF International
2017
Ethiopia Demographic and Health Survey 2016
.
Central Statistical Agency (CSA) & ICF International
,
Addis Ababa, Ethiopia
,
and Rockville, Maryland, USA. Retrieved from: https://dhsprogram.com/pubs/pdf/FR328/FR328.pdf.
Dewey
K. G.
Mayers
D. R.
2011
Early child growth: how do nutrition and infection interact?
Matern. Child Nutr.
7
,
129
142
.
doi:10.1111/j.1740-8709.2011.00357.x
.
Domenech
L.
Ringler
C.
2013
The Impact of Irrigation on Nutrition, Health, and Gender: A Review Paper with Insights for Africa South of the Sahara (IFPRI Discussion Paper No. 01259)
.
Gerber
N.
von Braun
J.
Usman
M. A.
Hasen
M. M.
Okyere
C. Y.
Vangani
R.
Wiesmann
D.
2019
Water, Sanitation and Agriculture Linkages with Health and Nutrition Improvement. (ZEF-Discussion Papers on Development Policy No. 282)
. .
Headey
D.
2014
An Analysis of Trends and Determinants of Child Undernutrition in Ethiopia, 2000–2011
.
International Food Policy Research Institute (IFPRI)
,
Washington, DC
Hoddinott
J.
Headey
D.
Dereje
M.
2015
Cows, missing milk markets, and nutrition in rural Ethiopia
.
J. Dev. Stud.
51
(
8
),
958
975
.
Humphrey
J. H.
2009
Child undernutrition, tropical enteropathy, toilets, and handwashing
.
Lancet
374
(
9694
),
1032
1035
.
https://doi.org/10.1016/S0140-6736(09)60950-8
.
IFPRI 2017
2017
Global Food Policy Report
.
International Food Policy Research Institute
,
Washington, DC
. .
Kirogo
V.
Wambui
K.
Muroki
N.
2007
The role of irrigation on improvement of nutritional status of young children in central Kenya
.
Afr. J. Food Agric. Nutr. Dev.
7
(
2
),
1
16
.
Leroy
J. L.
2011
zscore06: Stata Command for the Calculation of Anthropometric Z-Scores Using the 2006 WHO Child Growth Standards
.
Lin
A.
Arnold
B. F.
Afreen
S.
Goto
R.
Huda
T. M. N.
Haque
R.
Luby
S.
2013
Household environmental conditions are associated with enteropathy and impaired growth in rural Bangladesh
.
Am. J. Trop. Med. Hyg.
89
(
1
),
130
137
.
doi:10.4269/ajtmh.12-0629
.
Pickering
A. J.
Davis
J.
2012
Freshwater availability and water fetching distance affect child health in sub-Saharan Africa
.
Environ. Sci. Technol.
46
(
4
),
2391
2397
.
doi:10.1021/es203177v
.
Prüss-Üstün
A.
Bos
R.
Gore
F.
Bartram
J.
2008
Safer Water, Better Health: Costs, Benefits and Sustainability of Interventions to Protect and Promote Health
.
World Health Organization
,
Geneva
.
Sorenson
S. B.
Morssink
C.
Campos
P. A.
2011
Safe access to safe water in low income countries: water fetching in current times
.
Soc. Sci. Med.
72
(
9
),
1522
1526
.
http://dx.doi.org/10.1016/j.socscimed.2011.03.010
.
UNICEF
1990
Strategy for Improved Nutrition of Children and Women in Developing Countries
.
A UNICEF Policy Review
.
E/ICEF/1990/L.6
,
New York, NY
.
UNICEF
2013
Improving Child Nutrition: The Achievable Imperatives for Global Progress
.
United Nations Children's Fund (UNICEF)
,
New York
.
Usman
M. A.
Gerber
N.
2019
Assessing the effect of irrigation on household water quality and health: a case study in rural Ethiopia
.
Int. J. Environ. Health Res.
doi:10.1080/09603123.2019.1668544
.
Usman
M. A.
Gerber
N.
von Braun
J.
2019
The impact of drinking water quality and sanitation on child health: evidence from rural Ethiopia
.
J. Dev. Stud.
55
(
10
),
2193
2211
.
doi:10.1080/00220388.2018.1493193
.
van Houweling
E.
Hall
R. P.
Diop
A. S.
Davis
J.
Seiss
M.
2012
The role of productive water use in women's livelihoods: evidence from rural Senegal
.
Water Alternat.
5
(
3
),
658
677
.
von Braun
J.
Puetz
D.
Webb
P.
1989
Irrigation Technology and Commercialization of Rice in the Gambia: Effects on Income and Nutrition (Research Report 75)
.
WHO
2006
WHO Child Growth Standards: Length/Height-for-age, Weight-for-age, Weight-for-Length, Weight-for-Height and Body Mass Index-for-age: Methods and Development
.
World Health Organization
,
Geneva
.

Supplementary data