Associations of WHO/UNICEF Joint Monitoring Program (JMP) Water, Sanitation and Hygiene (WASH) Service Ladder service levels and sociodemographic factors with diarrhoeal disease among children under 5 years in Bishoftu town, Ethiopia: a cross-sectional study

Abstract

Objective

To determine the associations of WHO/UNICEF Joint Monitoring Program Water, Sanitation and Hygiene (WASH) Service Ladder service levels and sociodemographic factors with diarrhoeal disease among children under 5 years in Bishoftu town, Ethiopia.

Design

A community-based cross-sectional study.

Setting

Bishoftu town, Ethiopia, January–February 2022.

Participants

A total of 1807 mothers with at least one child under 5 years were included. Sociodemographic and WASH variables were collected using a structured questionnaire. 378 drinking water samples were collected.

Outcome

The response variable was diarrhoeal disease among children under 5 years.

Results

The 2-week prevalence of diarrhoeal disease among children under 5 years was 14.8%. Illiteracy (adjusted OR 3.15; 95% CI 1.54 to 6.47), occupation (0.35; 0.20 to 0.62), mother’s age (1.63; 1.15 to 2.31), family size (2.38; 1.68 to 3.39), wealth index (5.91; 3.01 to 11.59), residence type (1.98; 1.35 to 2.90), sex of the child (1.62; 1.17 to 2.24), child’s age (3.52; 2.51 to 4.93), breastfeeding status (2.83; 1.74 to 4.59), food storage practice (3.49; 1.74 to 8.26), unimproved drinking water source (8.16; 1.69 to 39.46), limited drinking water service (4.68; 1.47 to 14.95), open defecation practice (5.17; 1.95 to 13.70), unimproved sanitation service (2.74; 1.60 to 4.67), limited sanitation service (1.71; 1.10 to 2.65), no hygiene service (3.43; 1.91 to 6.16) and limited hygiene service (2.13; 1.17 to 3.86) were significantly associated with diarrhoeal disease.

Conclusion

In this study, diarrhoea among children is a significant health issue. Child’s age, drinking water service, residence type and hygiene service were the largest contributors with respect to the prevalence of diarrhoeal disease. This investigation provides information that could help to inform interventions to reduce childhood diarrhoea. The findings suggest that state authorities should initiate robust WASH strategies to achieve the Sustainable Development Goal 3 agenda.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The study used an international study tool (Joint Monitoring Program core questions on water, sanitation and hygiene for household surveys) that other researchers used to allow for international comparisons.

• This study was conducted in only one town of Ethiopia, so the results may not be generalisable to other settings.

• The study’s cross-sectional, observational design precludes the researchers from attributing causality to the associations identified.

Introduction

Diarrhoea is the passage of three or more watery or loose stools per day.¹ It causes the body to lose a substantial amount of salt and water, which can either cause rapid and severe dehydration and death or predispose children to malnutrition and increase their risk of infection.² Diarrhoea is typically a sign of an infection in the gastrointestinal tract, which can be brought on by several bacterial, viral and parasitic species.³ It is spread through contaminated food or drinking water or from person to person due to poor hygiene.⁴

Diarrhoea among children under 5 years is a central public health issue worldwide.⁵ It causes more than 1.7 billion cases and kills 525 000 children under 5 years old each year, making it the world’s second leading cause of mortality among children under 5 years.⁶ Globally, diarrhoeal disease accounts for 15% of all causes of child mortality that occur each day.^{7 8} Despite simple and effective treatments available, more than 2195 children die from diarrhoea daily.⁷ In addition, children under 5 years get diarrhoea sickness on average three times each year in low-income and middle-income nations.⁹ Furthermore, diarrhoea causes more than 90% of deaths in children under 5 years old in low-income and lower middle-income countries.¹⁰ Similarly, more than 88% of all child deaths under the age of 5 years in sub-Saharan African and South Asian nations are attributable to diarrhoea.¹⁰ In particular, infants exposed to contaminated environments and weakened inborn immunity are more likely to die from diarrhoea in the second half of their life.¹¹ Evidence revealed that African children under 5 years have at least five episodes of diarrhoea yearly, and about 800 000 die from it.¹² Similarly, diarrhoeal disease is the most significant public health issue in sub-Saharan African nations. It is responsible for 80% of child fatalities and more than 50% of paediatric illnesses.¹³ Ethiopia is one of the sub-Saharan nations with the highest prevalence of diarrhoeal morbidity and mortality rate.¹⁴ According to the Ethiopian Demographic Health Survey (EDHS) 2016 report, 1 in every 15 children die before turning 5 years old in Ethiopia, and 12% of children under 5 years had diarrhoeal episodes in the 2 weeks before the survey.¹⁵ Each year, diarrhoeal disease kills half a million Ethiopian children under 5 years, accounting for 23% of Ethiopia’s child mortality causes.^15–17

Infectious diseases linked with poor water, sanitation and hygiene (WASH) have been a substantial public health concern and have a negative impact on children’s health.¹⁸ The most common of these is diarrhoeal illness.¹⁹ However, diarrhoeal disease can be treated and prevented using easy and inexpensive measures.²⁰ Using safely managed drinking water, sanitation and hygiene services may be the best interventions to avoid diarrhoeal infection. Additionally, requiring health services to offer WASH services to all children is essential to reduce diarrhoeal disease, in saving many children lives.²¹ By 2030, the Sustainable Development Goal (SDG) 3 seeks to end the death of infants and children under 5 years old from preventable diseases, including diarrhoea. To do this, the SDGs call for the WHO and UNICEF Joint Monitoring Program (JMP) to set up the JMP ladders for WASH services, as effective WASH services are anticipated to reduce many infectious diseases, including diarrhoeal infection.^{22 23}

Factors for diarrhoeal disease occurrence are multifactorial, and it is frequently documented that diarrhoeal disease is linked to several sociodemographic, environmental and behavioural factors.²⁴ Notably, it is believed that type of occupation, household income level, age of the child and waste disposal systems are the main contributors to diarrhoeal disease among children.^{16 25 26} Particularly, poor access and lack of WASH services are responsible for 90% of diarrhoeal disease occurrences.^{27 28} However, no research demonstrates the link between the new JMP ladders for WASH services and child diarrhoea. Only a few studies have used representative municipal level from Ethiopia to assess the risk factors for diarrhoeal illness among children under 5 years. In addition, no study indicates the relative contribution of sociodemographic, WASH service and exposure to health information variables in the occurrence of diarrhoea among children under 5 years. Therefore, this study aimed to explore the associations of the JMP WASH Service Ladder service levels and sociodemographic factors with diarrhoeal disease.

Methods

Study design

A community-based cross-sectional study was conducted from January to February 2022 in Bishoftu town, Ethiopia.

Setting

This investigation was performed in Bishoftu town, Oromia National Regional State, Ethiopia. It is located at a distance of 44 km from the capital Addis Ababa.²⁹ The town is clustered into nine urban kebeles (administrative units) and five rural kebeles with a total population of 171 227.³⁰ The latitude and longitude of Bishoftu town are 8.734650 and 39.008533, respectively.³¹ The elevation of the city is 1920 m above sea level.³² The map of the study area is depicted in online supplemental file 1.

Inclusion criteria

Mothers with at least one child under 5 years were included in the sampling frame. Mothers with hearing problems or mental illnesses were excluded.

Sample size determination

For this investigation, the sample size was calculated using a single population proportion formula with 95% CI, 3% marginal error (e) and a 10% non-response rate. Based on a previous study, the prevalence of diarrhoeal disease among children under 5 years (P) was assumed to be 26%.³³

$${\displaystyle n=\frac{(Z^2\times P(1-P))}{e^2}}$$

$${\displaystyle n=\frac{(1.96)\times (1.96)\times (1-0.26)}{0.03\times 0.03}=821}$$

Hence, considering design effect 2 and 10% non-response rate, the total sample size for this study was 1807.

Sampling procedure

The sampling frame was stratified by gender and kebele administration, and the sample was selected in two stages. The Central Statistics Agency’s latest Population & Housing pre-census 2019 Enumeration Area (EA) sampling frame was used. In the first stage, 214 EAs were selected by the Ethiopian Central Statistics Agency. The EA size was the number of families listed in the EA. Each EA was demarcated using Q-field software. In the second stage, a new listing of households in which at least one child under 5 years resided was collected from each EA. Then, the new listing was divided by 8.5 (≈9) to obtain a sampling interval, and a constant of nine households was selected from each EA. At first, one household was determined using a lottery method and continuously by adding the sampling interval to acquire the required nine households per EA.

The JMP sampling technique for drinking water quality³⁴ stated that ‘within each cluster, 25 households are randomly selected for household interviews and 5 (20%) of these are randomly selected for water quality testing’. Similarly, within each EA, nine households were systematically selected for interview, and a fixed number of 1.5≈2 (>20%) of these was systematically selected for chemical and microbiological water quality tests in this study. In our design, three EAs are approximately equivalent to one JMP cluster, in terms of size. By taking two drinking water samples per EA, 378 drinking water samples were taken.

Data collection

Data collectors were selected based on technical experience in collecting the required data using Open Data Kit software. Forty-eight data collectors with a Bachelor of Science were employed to conduct the survey. In addition, eight master of public health degree holders served as team supervisors. Five days of training were provided to data collectors and supervisors. Written consent was acquired from each study participant. Data were collected using a structured questionnaire. Face-to-face interviews were conducted with mothers of children under 5 years. The survey questionnaire included questions on sociodemographic indicators household-level food safety and waste management practices, prevalence of environmental health-related diseases, exposure to hygiene and environmental health information, and child health status as well as JMP core questions on WASH (online supplemental file 2). The immunisation status of the children was validated by reviewing the child’s vaccination card.

In addition, data collectors completed observational modules of the questionnaire. These were related to WASH facilities. Drinking water samples were collected directly from households’ water storage. Plastic bottles with a capacity of 1000 mL for chemical parameters and 250 mL for microbiological characteristics were used to collect the drinking water samples. The sampling method was adapted from the WHO guidelines for drinking water quality.³⁵ The samples were conveyed to a laboratory within 6 hours and kept in a refrigerator at 4°C until analysis.

Data analysis

Data were checked for completeness and imported into Stata V.16 for analysis. Variables were constructed according to the JMP ladders for WASH service definitions. Principal component analysis was used to construct household wealth quintiles. Binary and multivariate logistic regression analyses were performed to identify factors affecting diarrhoeal disease, controlling for the effect of confounders. In the binary logistic regression analysis, all variables having a p value less than 0.05 were included in the multivariate logistic regression analysis. Additionally, in the multivariate logistic regression analysis, a p value less than 0.05 was considered statistically significant. To determine the relative contribution of explanatory variables, Shapley decomposition analysis was conducted. We tested the four logistic regression analysis assumptions: independence of observations, adequate sample size, no extreme influential outliers^{36 37} and no multicollinearity.³⁸ A Pearson goodness-of-fit test was used to assess and validate model adequacy.³⁹ Cronbach’s alpha (α) was used to assess the tool’s internal consistency and validity, and the value of ‘α’ was 0.72. This proved that the questionnaire was good, valid and trustworthy.⁴⁰ According to the US Food and Drug Administration guideline,⁴¹ the membrane filtration method was used for microbiological analysis and enumeration of Escherichia coli, total coliform and faecal coliform bacteria. Additionally, according to the US Environmental Protection Authority guideline, ion-selective electrode was used to assess fluoride levels.⁴² The level of nitrate concentration was determined by the ultraviolet spectrophotometer screening method according to the American Public Health Association/American Water Works Association/Water Environment Federation (1998) standard methods for the examination of water and wastewater.⁴³

Applied international definitions

Drinking water quality

According to the WHO drinking water guideline 2017, water samples with <1 colony-forming unit (CFU)/100 mL for indicator bacteria were considered pure, and samples with ≥1 CFU/100 mL were polluted.⁴⁴ Furthermore, drinking water samples with ≤1.5 mg/L and ≤50 mg/L concentration of fluoride and nitrate values were considered to be free from contamination, and samples exceeding these concentration values were considered contaminated.

JMP ladders for WASH services in households

Table 1 lists the JMP ladders for WASH services in households.

Table 1.

JMP ladders for water, sanitation and hygiene services in households

Service level	Definition
Drinking water service ladders
Safely managed	Drinking water from an improved source accessible on premises, available when needed, and free from faecal and priority chemical contamination (fluoride and nitrate)
Basic	Drinking water from an improved source, provided collection time is not more than 30 min for a round trip, including queuing
Limited	Drinking water from an improved source, collection time exceeds 30 min for a round trip, including queuing
Unimproved	Drinking water from an unprotected dug well or unprotected spring
Surface water	Drinking water directly from a river, dam, lake, pond, stream or irrigation canal
Sanitation service ladders
Safely managed	Use improved facilities that are not shared with other households and where excreta are safely disposed of in situ or removed and treated off-site
Basic	Use of improved facilities that are not shared with other households
Limited	Use of improved facilities that are shared with other households
Unimproved	Use of pit latrines without a slab or platform, hanging toilets or bucket latrines
Open defecation	Defecation in fields, forests, bushes, bodies of water or other open spaces
Hygiene service ladders
Basic	Availability of a handwashing facility with soap and water at home
Limited	Availability of a handwashing facility lacking soap and water at home
No services	No handwashing facility at home

JMP, Joint Monitoring Program.

Diarrhoeal disease

Diarrhoeal disease is defined as having at least three loose or watery stools per day for at least 1 day within the 2 weeks before the survey’s launch, as reported by the child’s mother.⁴⁵

Wealth index

The household wealth index is a measure of a family’s overall standard of living. It was constructed using household asset data via a principal component analysis. All asset types and construction process were adopted from the EDHS.¹⁵

Patient and public involvement

None.

Results

Sociodemographic characteristics of the study participants

Of the 1807 mothers with at least one child under 5 years old who were invited to participate in the study, 99% agreed. Of the total, 16.5% of the participants had no formal education. The mean age of the respondents was 33 years, and 45.2% were housewives. The majority (68.2%) of the participants were residents of urban areas. In addition, 19% and 26% of participants were in the first or second lowest wealth quintile, respectively. Most (61.1%) participants had less than four family members. Of the total, 54.7% of the children were female. Additionally, most (75.7%) of the children were above 2 years (table 2).

Table 2.

Sociodemographic characteristics of participants (n=1791)

Study variables	Category	Frequency	Percentage
Mother’s education	University level	276	15.4
	Secondary/vocational	756	42.2
	Primary	464	25.9
	No formal education	295	16.5
Mother’s age	≤33 years	1126	62.9
	>33 years	665	37.1
Mother’s occupation	Farmer	112	6.3
	Merchant	125	7.0
	Government worker	281	15.7
	Private sector worker	357	19.9
	Housewife	809	45.2
	Retired	107	6.0
Residence	Urban	1222	68.2
	Rural	569	31.8
Households’ wealth index	Richest	245	13.7
	Richer	327	18.3
	Middle	417	23.3
	Poorer	461	25.7
	Poorest	341	19.0
Households’ family size	≤4	1095	61.1
	>4	696	38.9
Sex of the child	Female	979	54.7
	Male	812	45.3
Child’s age in years	>2	1356	75.7
	≤2	435	24.3

Prevalence of diarrhoea and child health

The 2-week prevalence of diarrhoeal disease among children under 5 years was 14.8%. In addition, 12.9% and 15.3% of the children had experienced fever and cough, respectively, during the 2 weeks before data collection, and 6.6% had experienced difficulty breathing, with faster breath than usual and short rapid breaths. Over 90% of the children had received the BCG vaccination against tuberculosis, and 83.37% had received a vaccination against diphtheria, pertussis and tetanus (DPT). A significant number (8.3%) of children had never been breastfed (table 3).

Table 3.

Child health and vaccination status (n=1791)

Study variables	Category	Frequency	Percentage
Had diarrhoeal disease in the past 2 weeks	Yes	265	14.8
	No	1526	85.2
Had fever in the last 2 weeks	Yes	231	12.9
	No	1560	87.1
Had a cough for the last 2 weeks	Yes	274	15.3
	No	1517	84.7
Had difficulty breathing or was faster than usual, with short rapid breaths	Yes	119	6.6
	No	1672	93.4
Took BCG vaccination against tuberculosis	Yes	1613	90.1
	No	178	9.9
Took DPT vaccination to prevent tetanus, whooping cough and diphtheria	Yes	1564	87.3
	No	227	12.7
Ever breast fed	Yes	1643	91.7
	No	148	8.3

DPT, diphtheria, pertussis and tetanus.

Households’ WASH services, food storage practice and exposure to health information

The vast majority of sample households (83.9%) had access to the JMP-defined ‘basic’ level of drinking water services, followed by 9.7% who had ‘limited’ water access. Only 5% of sample households used ‘safely managed’ drinking water, while 1% and 0.4% of the households used unimproved and surface water, respectively. Access to ‘safely managed’ sanitation services was higher than for water, at 25.4%, but the majority (54.6%) of households fell in the ‘limited’ sanitation services category, and 12.4% and 1.9% used unimproved sanitation services and open defecation practices, respectively. A large share (38%) of households lacked access to hygiene services, and a minority (20%) had access to basic services on this dimension. Further, 2.4% of the participants did not separate raw and cooked food. Few (38.6%) respondents had participated in any WASH education, awareness creation or discussions over the past year (table 4).

Table 4.

Households’ WASH services, food storage practice and exposure to health information (n=1791)

Study variables	Category	Frequency	Per cent
Households’ JMP ladders for drinking water service	Safely managed	89	5.0
	Basic	1502	83.9
	Limited	174	9.7
	Unimproved	18	1.0
	Surface water	8	0.4
Households’ JMP ladders for sanitation service	Safely managed	455	25.4
	Basic	103	5.8
	Limited	977	54.6
	Unimproved	222	12.4
	Open defecation	34	1.9
Households’ JMP ladders for hygiene service	Basic	375	20.9
	Limited	735	41.0
	No service	681	38.0
Raw and cooked foods storing	Separately	1748	97.6
	Together	43	2.4
Mother attended awareness creation sessions about hygiene and environmental health practices in the past year	Yes	158	8.8
	No	1633	91.2
Mother attended WASH education aided with poster presentations in the past year	Yes	238	13.3
	No	1553	86.7
Mother participated in community and school WASH discussions/meetings in the past year	Yes	165	9.2
	No	1626	90.8
Mother obtained WASH education at religious venues in the past year	Yes	130	7.3
	No	1661	92.7

JMP, Joint Monitoring Program; WASH, water, sanitation and hygiene.

Bacteriological and chemical drinking water quality of the households

As depicted in figure 1, 47.1% and 29.4% of the households’ drinking water samples were positive for total and faecal coliforms, respectively. In addition, 13.5% of the households’ drinking water samples were positive for E. coli. On the other hand, nearly all of the households’ drinking water samples had acceptable nitrate and fluoride concentration (figure 1).

Figure 1.

Household bacteriological and chemical drinking water quality.

Methods of disposing of children’s faeces

While the majority of households (82.4%) used safe waste disposal systems, the shared unsafe disposal practice (17.6%) was substantial (online supplemental file 3).

Multivariable logistic regression analysis

Nineteen predictor variables, including those shown in table 5, as well as the child’s BCG and DPT vaccination status, showed a significant association with diarrhoeal illness in binary logistic regression analysis and were therefore included in a multivariate logistic model. In the multivariable logistic regression model, vaccination status was not significantly associated with diarrhoeal disease. The following variables remained significant predictors of this outcome (p<0.05) in the multivariate model: mother’s level of education, age and occupation; household size, wealth index and type of residence; child’s sex, age and (ever) breastfeeding status; households’ raw and cooked food storage practice, JMP ladder indicators for access to drinking water, sanitation and hygiene services; and mother’s participation in awareness creation, training or discussions related to WASH and environmental hygiene (table 5).

Table 5.

Multivariable logistic regression analysis (n=1791)

Study variables	Category	Diarrhoea		P value	AOR (95% CI)
		Yes	No
Mother’s education	University level	14	262		Reference
	Secondary/vocational	86	670	0.006	2.54 (1.30, 4.94)
	Primary school	85	379	0.003	2.90 (1.45, 5.78)
	No formal education	80	215	0.002	3.15 (1.54, 6.47)
Mother’s age	≤33 years	127	999		Reference
	>33 years	138	527	0.006	1.63 (1.15, 2.31)
Mother’s occupation	Farmer	47	65		Reference
	Merchant	19	106	0.418	1.38 (0.64, 2.97)
	Government worker	16	265	0.009	0.36 (0.17, 0.77)
	Private sector worker	76	281	0.749	1.10 (0.61, 2.01)
	Housewife	80	729	<0.001	0.35 (0.20, 0.62)
	Retired	27	80	0.153	1.73 (0.82, 3.68)
Households’ family size	≤4	118	977		Reference
	>4	147	549	<0.001	2.38 (1.68, 3.39)
Households’ wealth index	Richest	17	228		Reference
	Richer	40	287	0.054	2.01 (0.99, 4.08)
	Middle	59	358	0.002	2.90 (1.48, 5.69)
	Poorer	74	387	<0.001	3.14 (1.62, 6.10)
	Poorest	75	266	<0.001	5.91 (3.01, 11.59)
Residence	Urban	106	1116		Reference
	Rural	159	410	<0.001	1.98 (1.35, 2.90)
Sex of the child	Female	120	859		Reference
	Male	145	667	0.004	1.62 (1.17, 2.24)
Child’s age in years	>2	141	1215		Reference
	≤2	124	311	<0.001	3.52 (2.51, 4.93)
Ever breast fed	Yes	220	1423		Reference
	No	45	103	<0.001	2.83 (1.74, 4.59)
Households’ raw and cooked food storing practice	Separately	251	1497		Reference
	Together	14	29	0.004	3.49 (1.47, 8.26)
Households’ JMP ladders for drinking water service	Safely managed	4	85		Reference
	Basic	180	1322	0.184	2.09 (0.71, 6.17)
	Limited	67	107	0.009	4.68 (1.47, 14.95)
	Unimproved	9	9	0.009	8.16 (1.69, 39.46)
	Surface water	5	3	0.003	25.99 (3.05, 221.76)
Households’ JMP ladders for sanitation service	Safely managed	43	412		Reference
	Basic	8	95	0.791	1.13 (0.45, 2.84)
	Limited	134	843	0.016	1.71 (1.10, 2.65)
	Unimproved	61	161	<0.001	2.74 (1.60, 4.67)
	Open defecation	19	15	0.001	5.17 (1.95, 13.70)
Households’ JMP ladders for hygiene service	Basic	18	357		Reference
	Limited	87	648	0.013	2.13 (1.17, 3.86)
	No service	160	521	<0.001	3.43 (1.91, 6.16)
Mother attending awareness creation sessions about hygiene and environmental health practices in the past year	Yes	7	151		Reference
	No	258	1375	<0.001	2.93 (1.26, 6.77)
Mother attending WASH education aided with poster presentations in the past year	Yes	23	215		Reference
	No	242	1311	<0.001	2.06 (1.18, 3.59)
Mother participated in community and school WASH discussions in the past year	Yes	9	156		Reference
	No	256	1370	0.029	2.39 (1.10, 5.23)
Mother obtained WASH education at religious venues	Yes	6	124		Reference
	No	259	1402	0.002	4.52 (1.75, 11.68)

AOR, adjusted OR; JMP, Joint Monitoring Program; WASH, water, sanitation and hygiene.

Findings of the Shapley decomposition analysis

Online supplemental file 4 presents the findings of Shapley decomposition analysis to the determinants of child diarrhoeal disease. The models illustrate the extent to which variation in sociodemographic characteristics, access to WASH services and exposure to health information contribute to variation in the prevalence of diarrhoeal disease among children under 5 years. Child age, drinking water service, residence type and access to hygiene are the most important determinants of diarrhoeal disease, followed by mother’s education, access to sanitation, family size, household wealth index and mother’s age. Exposure to health information through hygiene and environmental health awareness creation sessions, WASH education at religious venues, community and school WASH discussions, and WASH education aided with posters were relatively less influential, and mother’s occupation, child sex and household food storage practices were the least important determinants of diarrhoeal disease in this study (online supplemental file 4).

Discussion

This investigation aimed to explore associations of JMP WASH ladder service levels and sociodemographic factors with diarrhoeal disease among children under 5 years of age in Bishoftu town. The 2-week prevalence of diarrhoeal disease among children under 5 years was 14.8%. This prevalence is consistent with the EDHS 2016 finding and cross-sectional studies conducted among Ethiopian children under 5 years in 2019 and 2021,^{15 25 46} but lower than a similar cross-sectional study performed among children under 5 years in Amhara regional state, north central Ethiopia,⁴⁷ and higher than that reported by a similar cross-sectional study performed among children under 5 years in the northeastern part of Ethiopia.⁴⁸

The prevalence of diarrhoea among children under 5 years of age in the study sample varied considerably by sociodemographic, WASH service and exposure to health information variables. We find that child’s age, drinking water service, residence type and hygiene service were the factors most strongly associated with the occurrence of diarrhoeal disease. More research is needed to determine why this occurs, as the cross-sectional, observational design of our study precludes causal analysis. However, it is worth nothing that variation in respondents’ exposure to community and school WASH discussions and WASH education, as well as their separation of raw and cooked foods, contributed a relatively small share of the variation in child diarrhoea.

Results of multivariate logistic analysis indicate that children whose mothers had no formal education had 3.15 times higher odds (adjusted OR (AOR)=3.15; 95% CI: 1.54 to 6.47) of experiencing diarrhoeal illness than children whose mothers had university education. This finding is consistent with previous research conducted in Ethiopia.¹⁷ Important caregiving practices such as timely weaning, good hygiene habits and appropriate child feeding may all benefit from higher levels of maternal education. In addition, children with mothers older than 33 years old had 1.63 times higher odds (AOR=1.63; 95% CI: 1.15 to 2.31) of developing diarrhoea than children with mothers younger than 33 years old. The reason behind the better outcomes for younger mothers could be that they have fewer other children to take care of, but we have no way of verifying that (and that should to some extent be taken care of by household size). In addition, this study indicated that children from households with more than four family members had 2.38 times (AOR=2.38; 95% CI: 1.68 to 3.39) greater odds of diarrhoeal disease than children from households with four or fewer family members. Economic differences may underlie this difference.

Indeed, we find that children living in the poorest families had 5.91 times (AOR=5.91; 95% CI: 3.01 to 11.59) higher odds of developing diarrhoeal disease than those living in the wealthiest homes. This result was consistent with cross-sectional research conducted among children under 5 years in northeast Ethiopia.⁴⁹ Further, diarrhoeal disease prevalence was higher (AOR=1.98; 95% CI: 1.35 to 2.90) among children living in rural areas compared with children living in urban areas. The cause could be due to disparities in infrastructure for WASH, as well as levels of knowledge about diarrhoeal disease.

We further find that diarrhoeal disease was more prevalent in male children (AOR=1.62; 95% CI: 1.17 to 2.24) than in female children, in line with the results of many cross-sectional studies conducted among Ethiopian and Bangladeshi children under 5 years.^50–53 The underlying reason behind the widely observed discrepancy in diarrhoeal disease between boys and girls is not well understood and requires further study. Age was strongly associated with diarrhoeal disease prevalence: children aged ≤2 years had 3.52 times higher odds (AOR = 3.52; 95% CI: 2.51 to 4.93) of being reported as experiencing diarrhoea than older children. This finding may be explained by the fact that consumption of complementary foods and mobility are new behaviours among the under-2 group, which could expose them to pathogens with which their immune systems have little prior experience, via contaminated foods and dirty surfaces. There is also a possibility of data quality differences between age groups, as mothers may be less likely to observe the stool of older children, and so may under-report diarrhoeal episodes for this group.

Additionally, we find that the risk of diarrhoeal disease was greater (AOR=2.83; 95% CI: 1.74 to 4.59) in children who had never been breastfed compared with those who had been breast fed at some point. This was supported by a cross-sectional study conducted in 2019 among children under 5 years in Bahir Dar city, northwest Ethiopia.²⁵ Breastmilk contains several disease-preventive nutrients and may help strengthen the immune system. Furthermore, children from families who practised inappropriate food storage had 3.49 higher odds of diarrhoeal disease (AOR=3.49; 95% CI: 1.74 to 8.26) than children from households who practised proper food storage. This demonstrated that mixing cooked and raw foods can lead to cross-contamination and serious public health issues. Furthermore, children with mothers who described their occupation as housewife had over one-third the odds of developing diarrhoeal disease (AOR=0.35; 95% CI: 0.20 to 0.62) compared with children of farmer mothers. This finding echoes that of a cross-sectional study conducted among Ethiopian children under 5 years in 2020.⁵⁴ This could be due to housewives spending more time and giving attention to their children, potentially reducing children’s exposure to the faecal–oral transmission route.

We used an internationally validated study tool (JMP core questions) to enable for international comparisons of the role of access to WASH services.⁵⁵ The link of diarrhoeal disease with the new JMP ladders for WASH services is strongly reflected in our results. The influence of the level of water services used by households is particularly striking. Children living in households using surface water had 25.99 times higher odds (AOR=25.99; 95% CI: 3.05 to 221.71) of diarrhoeal disease than children living in households using safely managed drinking water, those living in households using unimproved drinking water had 8.16 higher odds (AOR = 8.16; 95% CI: 1.69 to 39.46) and children living in households using limited drinking water services had 4.68 higher odds (AOR = 4.68; 95% CI: 1.47 to 14.95) of diarrhoeal disease than the reference group. These findings revealed that if more than 99% of households were to have access to safely managed drinking water service, according to the 2030 SDG agenda, the prevalence of diarrhoeal disease among children under 5 years would be significantly reduced.

Access to sanitation services was similarly influential. Children living in households using open defecation practices had 5.17 times higher odds (AOR = 5.17; 95% CI: 1.95 to 13.70) of diarrhoeal disease than those living in households using safely managed sanitation services, while those in households using unimproved sanitation services had 2.74 times higher odds (AOR = 2.74; 95% CI: 1.60 to 4.67) and those using limited sanitation services had 1.71 times higher odds (AOR = 1.71; 95% CI: 1.10 to 2.65) of diarrhoeal disease.

Additionally, we find that children who lived in households that lacked hygiene services had 3.43 times higher odds (AOR=3.43; 95% CI: 1.91 to 6.16) of diarrhoeal disease than those who resided in households with basic hygiene services, and children who live in households with limited hygiene services had 2.13 times higher odds (AOR=2.13; 95% CI 1.17 to 3.86). This indicates that basic hygiene services substantially reduced the risk of diarrhoeal disease.

Finally, we find that mothers’ participation in hygiene education programmes is associated with lower diarrhoeal disease. Children whose mothers had not attended environmental health awareness creation sessions had 2.93 higher odds of experiencing diarrhoea than those whose mothers had attended such sessions (AOR=2.93; 95% CI 1.26 to 6.77). Furthermore, children whose mothers did not participate in WASH education aided by poster presentations were 2.06 times more likely to develop diarrhoeal disease (AOR=2.06; 95% CI 1.18 to 3.59) than children whose mothers did. In addition, children whose mothers did not participate in community and school WASH discussions were 2.39 times (AOR=2.39; 95% CI 1.10 to 5.23) more likely to develop diarrhoeal disease than children whose mothers did participate. Likewise, the odds of developing diarrhoeal disease were 4.52 times (AOR=4.52; 95% CI 1.75 to 11.68) higher in children whose mothers did not get WASH education at religious settings than in children whose mothers did. This suggests that various approaches to awareness creation significantly reduce diarrhoeal illness among children in this setting by increasing understanding and improving hygienic behaviour among mothers. This is supported by a quasi-experimental study conducted in Central Sudan, which found that health education activities such as lectures and focus group discussions resulted in a significant improvement in mothers’ knowledge, attitude and practice regarding the prevention of diarrhoea in children under 5 years.⁵⁶

Conclusion

Two-week prevalence of diarrhoeal illness among children under 5 years is a serious health concern in the study population and in Ethiopia more generally. In this study, we have shown that child’s age, drinking water access, residence type and access to hygiene services were the largest contributors with respect to explaining variation in the prevalence of diarrhoeal disease. Many additional associated factors were also identified. This study provides information that public health programmers and decision-makers could use to inform innovative interventions to reduce childhood diarrhoeal disease. The findings of this study suggest that state authorities and partner organisations should initiate strong WASH policies and strategies to achieve the SDG 3 agenda.

Data availability statement

Data are available upon reasonable request.

Ethics statements

Patient consent for publication

Not required.

Ethics approval

The Ethiopian Public Health Institute scientific and ethical review board approved the study proposal with reference number EPHI-IRB-358–2021. Written consent was acquired from the participants. Those who did not desire to participate in the study were not forced to participate. The confidentiality and privacy of participants were ensured throughout the investigation.