Effects of community-led total sanitation and hygiene implementation on diarrheal diseases prevention in children less than five years of age in South Western Ethiopia: A quasi- experimental study

Background Lack of improved sanitation is the most important contributing factor to diarrheal disease among under-five children in low and middle-income countries. There was no study to identify the effect of Community-Led Total Sanitation and Hygiene intervention on diarrheal diseases in the study area. Hence, this study was designed with the aim of finding the effects of Community-led Total Sanitation and Hygiene implementation for preventing diarrhea among under-five children. Methods A community-based Quasi-Experimental study was conducted among a sample of 846 households selected from intervention (kersa) and comparison (mana) districts using the four-stage random cluster-sampling method. A Semi-structured questionnaire was used to collect data. The collected data was cleaned, coded, and entered into EpiData version 3.1 and exported to SPSS version 20 for analysis. Difference-in Difference method with McNemar’s tests was used to compare the prevalence of diarrhea between the intervention and comparison districts, and the significance of change between the pre-test and post-test was declared at p-value less than 0.05 with 95% confidence interval. Results The intervention led to decreased diarrhea prevalence [pp = -8.2, 95% CI: -15.9, -0.5], increased latrine ownership [pp = 5.6, 95% CI: 0.5, 10.8], and increased latrine utilization [pp = 10.7, 95% CI: 4.7, 16.6] in intervention district at post-test compared to the baseline; while the presence of handwashing facility near the latrine, home-based water treatment, and proper water storage and handling practice were decreased at post-test compared to the baseline. Conclusion Implementation of Community-Led Total Sanitation improved sanitation and hygiene status of community that resulted in the reduction of diarrhea diseases in under-five children. Further implementation, evaluation, and scale-up of the interventions are needed to reduce diarrheal disease in under-five children.

of age. Further implementation, evaluation, and scale-up of the intervention is needed to reduce diarrheal disease in children less than five years of age. Background Diarrhea remains a major public health problem especially in developing countries where it is a leading cause of childhood morbidity and mortality (1). Although diarrhea has a low prevalence in high-income countries, it can occur in vulnerable populations anywhere (2). Provision of clean water and sanitation could substantially reduce the hundreds of thousands of deaths each year caused by diarrheal diseases (3).
The use of improved sanitation and excreta disposal were significantly associated with lower odds of diarrhea (4). An estimated 85% of diarrhea mortality is attributed to unsafe drinking water, inadequate sanitation, and substandard hygiene practices (5). The practice of open defecation is thought to be a major cause of the persistent worldwide burden of diarrhea and enteric parasite infection among children less than 5 years. Reducing open defecation requires access to and use of improved sanitation facilities, which are defined as facilities that prevent human feces from reentering the environment (6).
As a direct consequence of diarrheal pathogens excreted into the environment, in Ethiopia about 13% of children under age of 5 years had diarrhea, and 24-30% of all infant deaths and 25% of deaths among children aged between 1 and 4 years were due to diarrhea (7). Community Led Total Sanitation (CLTS) is an approach that focuses on sustained behavioral change through motivation and mobilization to facilitate and enhance community knowledge and understanding of the risks associated with open defecation. The approach is aimed at empowering the community to analyze the extent and risks of environmental pollution caused by open defecation and to construct and use toilets with their own resources (8).
Community-Led Total Sanitation is a sanitation promotion based on stimulating a collective sense of disgust and shame among community members as they confront the crude facts about mass open defecation and its negative impacts on the entire community. The basic assumption is that no human being can stay unmoved once they have learned that they are ingesting other people's feces.
Generally communities react strongly and immediately try to find ways to change this through their own effort based on different motivations (9).
Observational studies conducted to evaluate the effects of interventions that prevent human feces from entering the environment have shown that they reduce diarrheal diseases and enteric parasite infections (10)(11)(12)(13)(14)(15).Most of those researches, however, have focused on the construction and utilization of household level latrine. Few studies have been conducted in rural areas of low-income countries where CLTSH intervention is implemented (16,17).
Since all of kebeles (smallest administrative unit in Ethiopia) have declared open defecation free in Kersa district (intervention district), and there was no finding on the effect of CLTSH intervention in the district, a strong need exists for the evidence of the effects of implementation of community-led total sanitation and hygiene approach in the study area. This study was, therefore, designed with the aim of finding evidence for the effect of implementation of community-led total sanitation and hygiene approach on the prevalence of diarrheal diseases in children under five years of age.

Study setting
A Quasi-experimental study was conducted in Kersa and Manna districts, Jimma Zone, Oromia

Population and sampling
The source populations were all households residing in Kersa and Manna districts during data collection period and the study populations were selected households having children less than five years of age who were residents of the two districts during data collection period. Households with at least one child under 5 years of age were eligible for this study. The age of a child was verified with an immunization card, which shows the birth date of the child. Households with under five children, but who were unable to communicate (in the absence of appropriate respondents or when respondents were unable to hear or speak) during data collection were excluded from the study.
The required sample size was calculated using G*Power 3.1.9.2 software considering the following assumptions: 5% for type I error, 80% power, Odds ratio of 2, proportion of discordant pairs 0.255, prevalence of diarrhea from a previous study 22.22% (11), effect size of 10% , and design effect of 1.6. Finally, after adding 10% non-response rate, the total sample size was calculated to be 814 participants (407 for intervention district, and 407 for comparison district).
A four-stage random cluster-sampling technique was employed for selecting the study participants. In the first stage, Kersa and Manna districts were purposively selected. The implementation of CLTSH intervention was used as criteria for selecting and including districts in the study. Accordingly, Kersa district was selected as an intervention district. Simultaneously, non-intervention (Manna district) was selected as comparison group for comparison. In the second stage, 10 out of the 30 kebeles in the intervention and 10 out of 19 kebeles of the non-intervention district were selected randomly and included in the study. In the third stage, five "gares" (sub Kebele) were selected from each kebele by using probability proportional to population size (PPS) method. In the fourth stage, eight households were randomly selected from each selected "gare". The number of households in each Kebele was decided based on the district level household proportion. In the selected households, the youngest child in the households was selected to be included in the study.

Baseline data
The lack of appropriate baseline data was the main challenge we faced during conducting this study, as it is impossible to measure the impact of the intervention without reliable data on the situation before the intervention began. To overcome this problem, we reconstructed the baseline data using secondary data sources (i.e. survey and annual report).We used the report of survey conducted to evaluate the implementation of Community-led Total Sanitation and Hygiene Approach on the Prevention of Diarrheal Disease in Kersa District (11) as well as zonal and district administrative reports to reconstruct the baseline data.

Post-intervention data
Post-intervention data was collected from March 01-April 10, 2019. A semi-structured questionnaire was used to collect data. The data collection tool was adapted from various similar studies (19-21), and modified according to local context. Data collection tool was translated from English to Afan Oromo language, and then back to English to check for consistency by different professionals.
Data was collected using face-to-face interview of primary caretaker of the selected child and observation of sanitation facilities and household condition to assess indicators of behavioral, environmental, and socio-demographic factors at home level. Before declaring non-respondent household, up to two visits was made at different times by data collectors. Data collectors and supervisors were residents of kebeles other than the study area to minimize social desirability bias, and with educational level of secondary school. There were 8 data collectors and 4 supervisors, and the overall activities were coordinated by the principal investigator. To assure the quality of data, a one-day training was given for data collectors and supervisors. Training was given by the principal investigator on the objective of the study, the methods of data collection, how to recruit eligible households, with practical exercises. The principal investigator and the supervisors closely monitored the entire data collection processes. Completed questionnaires were collected and delivered to supervisors after checking for consistency and completeness on daily basis. Missing values and outlier were checked before analysis by running descriptive analysis.
The questionnaire was pre-tested for one day before the actual data collection takes place in a non-study area (Omo nada and Seka chekorsa districts) by taking 5% of the sample size and the necessary corrections were made before the actual data collection.

Data Analysis
Data were entered into Epi-data version 3.1 software and exported to SPSS Version 20 software for analysis. Descriptive statistics were calculated from household surveys and observations, and presented by frequency distribution, summary measures, tables and graphs. McNemar test (crosstabulations) and Generalized Linear Model (GLM) analysis were conducted to estimate the prevalence of primary and secondary outcome variables between intervention and comparison districts at baseline and posttest survey. The primary outcome was diarrheal disease prevalence on children less than five years of age, with the household as the unit of analysis. Secondary outcomes assessed were the availability of latrine, latrine utilization, availability of cover for latrine seat, availability of hand washing facility near latrine, availability of soap at hand washing, safe disposal of child feces, solid waste management, hand washing practice at critical times, presence of feces in the compound, and water storage and handling practice. Self-reported latrine use was validated by observation of latrines. Latrines that were full, collapsed structure, or had unstable flooring were categorized as open defecation. The prevalence of primary outcome between baseline and posttest surveys both in the intervention and comparison districts was compared using a Difference-in-Difference estimator. In addition, McNemar's test was used to compare the status of secondary outcomes between baseline and posttest surveys. Difference in baseline and posttest surveys was declared statistically significant with P-value less than 0.05.

Socio-demographic characteristics
Samples of 423 households were included in this study each from intervention and comparison districts. Of all households interviewed, the response rate was 99.5% and 98.1% in intervention and comparison districts, respectively. Majority of respondents in intervention and comparison districts (63.9% and 74.7% respectively), were married. More than 75% of respondents in both intervention and comparison districts were in the age range of 26-40 years. In both districts, the family size of more than half of households was in the range of 3-5. More than 60% of households in both districts had one child under five years of age. On average, almost half of heads of the households in both intervention and comparison districts were illiterate. Majority of households in both intervention and comparison districts (96.4% and 89.9% respectively), were Oromo by ethnicity. About 70% of heads of households in both intervention and comparison districts were farmers. Regarding their religion, more than 90% of them were Muslims in both intervention and comparison districts. The average monthly income of about 40% of households in both intervention and comparison districts were 350-550 Ethiopian Birr (Table 1). Effects of the intervention on primary outcome district were more likely to report a significant decline in prevalence of diarrhea during the pre-test and post-test periods than those in comparison district [pp = -8.2, 95% CI: -15.9, -0.5] ( Table 2 and Fig. 1).

Effects of the intervention on secondary outcomes
This study revealed that the intervention had significant effects on several secondary outcome measures. Latrine coverage, availability of cover for latrine seat hole, latrine utilization, availability of hand washing facility near latrine, availability of soap at hand washing facility, safe management of solid waste, safe disposal of child feces, hand washing practice at critical times, presence of cover for water container, safe drawing of water from the storage, and water treatment at home are expected to increase in intervention district as compared to comparison district, while presence of feces around home is expected to decrease in intervention district. Accordingly, the prevalence of secondary outcomes at baseline and post-test for the intervention and comparison districts showed that eight out of twelve indicators suggested differences between intervention and comparison districts in the intended directions.  (Table 3 and Table 4).

Discussion
This study was conducted to evaluate the effectiveness of the implementation of Community-Led Total Sanitation and Hygiene (CLTSH) approach on the prevention of diarrheal disease in children less than five years of age. Statistical analyses done on the differences of diarrheal disease prevalence and factors associated with diarrhea between intervention and comparison districts were found to be significant.
This study showed that there was 8.2 percentage point greater reduction in the prevalence of diarrhea after the implementation of the intervention in intervention district. This is supported by a study conducted in Hadaleala district, Ethiopia, which showed that human excreta management was associated with childhood diarrheal disease (7).
In our study, in the intervention district, implementation of CLTSH resulted in a 5.6 percentage point increase in household private latrine ownership than in comparison district. This is supported by a cluster-randomized controlled trial conducted on the effect of a community-led sanitation intervention on child diarrhea and child growth in rural Mali, which showed latrine ownership rose more steeply as a result of CLTS; latrine ownership increased by 39 percentage points (22). Another study conducted on the effect of CLTS on latrine ownership in Mozambique showed that the proportion of people owning latrine is increasing with increasing extend of CLTS-related information and highest in the group of CLTS participation (79%) (23).
The coverage of private latrine ownership is also greater than the finding of a study conducted rural settings of Dangla District, Northwest Ethiopia, which showed that majority of households in both ODF (89.7%) and OD (92.8%) kebeles had a private latrine (24). The possible explanation for this variation might be due to the difference in the extent of the implementation of WASH intervention or due to differences in the study design employed.
In addition to increased possession of latrine, participants in the intervention district were more likely to report latrine utilization at posttest assessment compared to baseline. This is supported by a study conducted on the effect of community led total sanitation and hygiene approach on improvement of latrine utilization in Laelay Maichew District, North Ethiopia, which showed that implementation of CLTS improves the utilization of latrine by 16.2 percentage points (16).
This study revealed that the availability of hand washing facility near latrine was reduced in both intervention and comparison districts from baseline to posttest survey. It was reduced from 70.3-62.7% (-7.6 pp) and from 74-54.8% (-19.2 pp) in intervention and comparison districts respectively.
The finding is higher than the study from Yaya Gulele district, which showed that more than half of the participants, (54%) in CLTSH implemented and (63%) unimplemented kebeles had no handwashing facility in or close to the latrine (25). The possible explanation for this variation might be due to the difference in the extent of the implementation of the intervention or due to differences in the study period.
Our study showed that, from baseline to posttest, the availability of soap at hand washing facility was more increased in intervention district than in comparison district as compared to the baseline, which is greater than the study conducted on Community-Level Sanitation Coverage in rural Mali (22). This variation might be due to the difference in the level of the sanitation coverage and geographical variation between the study populations.
The practice of open defecation was more decreased by 3.5 percentage points in intervention district, at posttest assessment compared to baseline. This is almost similar with a study conducted to evaluate the sustainability of community-led total sanitation outcomes in Ethiopia and Ghana, which shows that open defecation practice was decreased by 12 and 17 percentage points respectively, in villages receiving CLTS interventions (26). (1.11%) of children's stools was buried (32). It is also greater than the finding from Benishangul Gumuz Regional State, North West Ethiopia, which showed 55% of the households disposed children's' stool in an improper manner (33). This variation might be due to the difference in the study period and level of implementation of WASH intervention.

Participants in both intervention and comparison districts reported a significant increase on hand
washing practice at critical times. But there is a 0.5 percentage points greater increase in intervention district from baseline to posttest survey, as compared to comparison district. This is greater than the finding of a study in Sheko district, Southwest Ethiopia and Jabithennan district, Northwest Ethiopia, which shows that 61.5% and 72.6% of households practiced hand washing at critical times respectively (34,35). The possible explanations for these variations might be due to the difference in the extent of the implementation of the intervention on the study populations and the study period in which the data collection period for our study is recent than that of these studies.
In our study, the availability of cover for water container during storage was 97.9% and 98.6% in intervention and comparison districts, respectively, but declined as compared to the base line. In intervention district, it was decreased by 1.1 percentage points, which is the reverse of the intended effect of the intervention. This is higher than the finding of a study from Mali, which shows that 96% of households had stored water covered at the time of sample collection (12). Another similar study with our finding from rural Ethiopia, shows that drinking water was stored in the household primarily in a container with a lid (98%) (36).
In our study, the practice of safe drawing of water from the storage was 60.1% and 72% in intervention and comparison districts, respectively, and declined in both intervention and comparison districts as compared to the base line. In intervention district, it was decreased by 16.4 percentage points, which is the reverse of the intended effect of the intervention. This is lower than the finding of a study conducted on Diarrheal status and associated factors in under five years old children in relation to implemented and unimplemented community-led total sanitation and hygiene in Yaya Gulele district, which shows that the practice of safe drawing of water from the storage was 99% and 84% for CLTS implemented and non-implemented districts respectively (25). This might be happened due to the difference on the awareness level of the study populations.
This study also revealed that, the practice of water treatment at home level was 15.9% and 20.7% in intervention and comparison districts, respectively, and declined as compared to the base line. In intervention district, it was more decreased by 4.6 percentage points than in comparison district, which is the reverse of the intended effect of the intervention. This is higher than the finding of a study conducted on appropriate household water treatment methods in Ethiopia, which shows that the number of households treating their water prior to drinking with any treatment options was 8.0% in 2005, 10.2% in 2011, and 9.4% in 2016 (37). The difference might be happened due to the difference on the water sources used by the study populations, and difference on the time of data collection. The possible explanation of the reduction of this practice between the baseline and posttest survey in our study might be due to the expansion of safe water supply between the baseline and posttest survey.
There were some limitations such as lack of random assignment which is the major limitation of quasiexperimental study. Besides, since the baseline data was not collected prior to the implementation of the intervention for comparison district, there are threats to the internal validity.

Conclusion
In this study, we presented the effects of implementation of community-led total sanitation and hygiene approach on the prevalence of diarrheal disease in children less than five years of age.
Despite the limitations of the present study, including the nature of the study design and reconstructing of baseline data for comparison district, the results identified levels of improvements in sanitation to deliver the expected health benefits within the intervention district. The findings also suggested that the implementation of community-led total sanitation and hygiene approach can be an effective tool to reduce diarrheal disease prevalence in children under the age of 5 years. Hence, the local health authorities need to work on improvement of some indicators as per CLTS guideline.
Moreover, provision of refreshment activities to sustain these observed changes as well as to improve the above mentioned indicators is recommended. More investigation is needed to evaluate this intervention in different settings and Experimental study would also help to provide a stronger evidential base for the sustainability of the intervention's effect. University. Permission letter was also obtained from Jimma zonal health department and respective District Health Offices to conduct the study. In addition, Kebele Administrators were informed about the purpose of the study. Before the commencement of data collection, caregivers of a child (parents or guardians 18 years and above) were well informed by data collectors about the objectives and importance of their involvement in the study, and the confidentiality of the information they provide.
Finally, after taking informed consent, caregivers (parents or guardians) of a child who were willing to take part in the study were interviewed.

Consent for publication
Not applicable as it does not contain an individual person's data

Availability of data and materials
All data generated or analyzed during this study are included in this article