The study was conducted at Welkite town from 1 February to 30 March 2021. Welkite town is the administrative center of the Gurage zone of the Southern Nations, Nationalities, and Peoples’ Region (SNNPR). The town is located at a latitude and longitude of 8o17′N37o47′E and an elevation between 1910 and 1935 m above sea level (Figure 1). It is located 155 km west of capital city of the county, Addis Ababa. The town has three sub-cities and six kebeles (the smallest administrative unit) and it is one of the most densely populated towns in Ethiopia with an average population density of 283 people/km².

A community-based cross-sectional study design was employed. The source population was all the under-five year children living in Welkite town, and the study population was under-five children living in the selected kebeles of Welkite town.

Sample size was estimated for a 2 week recall period based on a previous report of 21.3% among under-five children in Halaba special district, in southern Ethiopia [22]. With a 95% confidence level, the margin of error of 5%, and design effect of 1.5 and 10% non-response. This resulted in the final sample size of 426.

Multi-stage sampling was used to obtain a representative sample of the study participants. First, three Kebeles (the smallest administrative unit) were selected from the total six kebeles in the town using a random sampling technique. Household lists having under-five children were taken from health extension workers’ registration books in each kebele. Codes were given to each household, and computer-generated random numbers were used to select the households. A lottery method was employed to select one child in case more than one child younger than 5 years old was found in the household (Figure 2).

Socio-demographic, behavioral, environmental, and other related data were collected from mothers/primary caretakers using a structured questionnaire and observation checklists. The questionnaire was adapted from similar studies and customized accordingly. One trained data collector was recruited in each selected kebele and the principal investigator monitored the overall data collection process. Data collectors have been trained about interview methods, consent, and ethical aspects required to be executed during data collection.

Data was collected by Open Data Kit (ODK) version 1.27.3 (https://apkpure.com/odk-collect/org.odk.collect.android/download/), and it was exported to SPSS version 25 (IBM Corp., Armonk, NY: IBM Corp.) for analysis. Depending on the distribution of the continuous variables, mean (standard deviation), or median (interquartile ranges) was used to describe the variables, and categorical variables were reported as a number and percentage. The Shapiro-Wilk test was used to evaluate the normality of the quantitative data. Logistic regression was used to identify factors associated with diarrhea status after checking all the preliminary assumptions of the model. Variables with a p-value < 0.20 in bivariate analysis were selected as candidate variables to be included in a multivariable analysis in order to control for confounders. Variables with a p-value ≤ 0.05 in multivariable analysis were reported as being statistically significant, and AOR with 95% CI was reported as measures of the strength of the associations. The explanatory variables were tested for multi-collinearity using the Variance Inflation Factor (VIF) and the tolerance tests.

Diarrhea is the passage of three or more abnormally loose, watery, or liquid stools over 24 h periods

Completely immunized children were those who received all childhood vaccines according to Ethiopia national immunization program for (polio, rotavirus, tuberculous meningitis, and military tuberculosis, pneumonia, diphtheria, pertussis, tetanus, Hemophilus influenza type b, Hepatitis B, and measles).

Partially immunized children were those who received one or more doses of the primary doses of the National Immunization Schedule but not completed all of them.

Safe child stool disposal practice: a child uses a toilet facility or the child’s feces are put into a toilet (regardless of the type of toilet) or buried.

Overcrowding is a condition in which one bedroom is shared by three or more people.

Disposal of household rubbish: Proper solid waste disposal entails burying or storing waste in a container and disposing of it at a designated location. Household trash is improperly disposed of by being placed in open spaces and in roadside ditches along with other garbage.

Environmental sanitation: Environmental sanitation includes human excreta control, managing solid waste and wastewater, and pest and vector control.

A total of 426 under-five-year-old children participated in the study with 100% response. Respondents were children’s mothers/primary caregivers and 363 (85.2%) of the respondents were biological parents. The majority 263 (61.8%) of the respondents had primary or secondary levels of education. Of the respondents, 227 (53.3%) were housewives and 101 (23.7%) were government employees. The majority 257 (60.3%) of the study respondents had an independent house with a roof constructed from a corrugated iron sheet. Of the total households included in the study, 352 (82.6%) and 143 (33.6%) had dwellings made of wood and mud wall and mud floor respectively. One hundred ninety-two (45.1%) respondents reported that >4 persons live per household. Households included in this study had 4.57 ± 1.50 SD average family size and each household on average had 2.59 ± 1.21 SD rooms.

The median age of children was 30.5 months with an Interquartile range (IQR) of (19.75–42.25). Most 313 (73.5%) of participants were in the age group of 24–59 months. There were slightly more 217 (50.9%) males than females.

About two-thirds, 266 (62.4%) of the study participants were not breastfeeding at the time of the study but on average they have fed for 21.38± (6.73 SD) months. Most, 331 (77.7%) of participants were fully vaccinated (Table 1).

Two hundred ninety-two (68.5%) of the respondents had frequent handwashing habits. Of the total respondents, 288 (67.6%), 185 (43.4%), and 255 (59.9%) wash their hands after the toilet, after cleaning, and before feeding the child respectively. Household members of 33 (7.7%), 139 (32.6%), and 190 (44.6%) had the habit of consuming raw milk, raw vegetable, and overnight stored foods after its preparation, respectively. From the observation of household environment, it was found that 206 (48.4%) of water storage area and 88 (20.7%) of water-storing materials were not clean and not covered, respectively. With regards to the food preparation area, 251 (58.9%) households prepare food in the unclean area and there were flies in 89 (20.9%) of households’ food preparation area. The majority 288 (67.6%) of households had improved latrine and almost half 217 (50.9%) of them had private latrine used only by a given household. About three-fourth, 318 (74.6%) of households dispose of their child’s stool in the toilet. Faeces were observed outside the toilet in 71 (16.7%) of the households and there were flies in 145 (34.1%) households. Almost all households 385 (90.8%) had no handwashing facility. There were human and animal faces lying around the house in 10 (2.3%) and 24 (5.6%) of households respectively. Children playing with animal feces lying around the house were observed in 23 (5.3%) of the households (Table 2).

Respondents from 109 (25.6%) households reported the presence of domestic animals in their house and most 82 (75.2%) reported that animals stay outside the dwelling at night. Of the reported animal type present in the household, 42 (38.5%) and 34 (31.2%) were cats and chickens, respectively. Children had close contact with animals in about one-third 37 (33.9%) of households where the presence of animals was reported.

The 2-week prevalence of diarrhea among under-five children was 20.7% (88/426) (95% CI: 17.1%, 24.6%). In bivariate analysis, child's immunization status, mother's/caregiver's educational status, mother's/caregiver's occupation, number of people sharing a sleeping room, breastfeeding, child's stool disposing practice, eating prepared food after overnight storage, cleanness of water storage area, water storage material being covered, cleanness of food preparation area, presence of flies in food preparation area and detection of feces outside a toilet were identified as candidate variables for the multivariable analysis. In the multivariable analysis the odds of having diarrhea among under-five years age children during the last 2 weeks was about five times higher among children whose mothers/caregivers are merchants as compared to children whose mothers/caregivers are housewives [AOR = 5.34; 95% CI: 2.06, 13.82]. Likewise, the odds of a child having diarrhea is 2.67 times higher among partially vaccinated children compared to fully vaccinated counterparts [AOR = 2.67; 95% CI: 1.23, 5.83]. Furthermore, the odds of a child having diarrhea is about two times higher in children living in a household where flies are found in a food preparation area than their counterpart [AOR: 2.24; 95% CI: 1.05, 4.78]. Additionally, the odds of diarrhea occurrence was about five times higher among children whose mothers/caregivers dispose of their stool into garbage than those who put child stools in the toilet [AOR = 5.05; 95%CI: 1.09–23.32]. Children living in households where water was stored without cover were 2.35 times more likely to develop diarrhea than their counterparts ([AOR = 2.35 (1.18, 4.66)]. The difference in the prevalence of diarrhea across categories of potential risk factors as compared to respective reference categories and findings from multi-variable logistic regression analysis are summarized in Table 3.