Determinants of acute diarrhoea among children under five years of age in Derashe District , Southern Ethiopia

Introduction: Diarrhoea kills more children than AIDS, malaria, and measles combined. Knowing the determinants of a disease enables us to design an effective intervention. The objective of this study was to identify the determinants of acute diarrhoea and associated factors among children under 5 years of age in Derashe district, south Ethiopia. Methods: A community based unmatched case-control study supplemented with Focus Group Discussions (FGDs) was employed in rural kebeles (neighborhoods) of the district. Collected data were entered in Epi Info v3.5.3 (http://wwwn.cdc.gov/epiinfo/) and descriptive data analyses were performed using SPSS v16.0 (www.spss.com). Binary logistic regression analysis was used to measure the association between dependent and independent variables, calculating odds ratios and 95% confidence intervals (CI). Statistical significance was set at α ≤0.05. Multivariable analyses were applied to identify the relative effect of explanatory variables on the dependent variable. Results: The study revealed that the occurrence of diarrhoea was significantly associated with lack of latrine ownership (adjusted [A] OR: 2.43, CI:1.19-4.87), lack of home-based water treatment (AOR: 2.25, CI:1.43-3.56), lack of improved water sources (AOR: 1.98, CI:1.162.23) and consumption of left-over food (AOR: 1.65, CI:1.01-2.71).


Introduction
Diarrhoea is a major health problem 1 .It is usually a symptom of an infection in the intestinal tract, which has a variety of causative agents including viruses, bacteria and parasites 2,3 .Diarrhoeal infection spreads through the ingestion of contaminated food or drinking-water, or person-to-person as a result of poor hygiene.There are three clinical types of diarrhoea: (i) acute watery diarrhoea which lasts several hours or days and includes cholera; (ii) acute bloody diarrhoea, also called dysentery; and (iii) persistent diarrhoea that lasts 14 days or longer 1,4 .
The vast majority of deaths from diarrhoea are among children under 5 years of age living in low-and middleincome countries 5 .Diarrhoeal disease due to unsafe water and lack of sanitation is the greatest cause of morbidity and mortality in under-five children in the world, especially in poor countries 5,6 .A child dies every 15 seconds from diarrhoea caused largely by poor sanitation and a contaminated water supply 7 .As a child's immune system is progressively compromised with each bout of diarrhoea, related illnesses kill millions more indirectly.In developing countries, approximately 2 million people, the vast majority of whom are under-five children, die from diarrhoea each year 6 .Nearly 90% of diarrhoea is attributed to unsafe drinking water, inadequate sanitation and poor hygiene 6 .
It is widely recognized that exposure to diarrhoea pathogens in developing countries is associated with the age of the child, quality and quantity of water, availability of toilet facilities, housing conditions, level of education, household economic status, place of residence, feeding practices, and the general sanitary conditions (personal or domestic hygiene) in the vicinity of the house [8][9][10][11][12] .Socioeconomic factors may directly and indirectly affect environmental, behavioral, nutritional, and demographic risk factors, with the exception of age and sex 9 .
Diarrhoea alone kills more children than AIDS, malaria, and measles combined 10 .Recent estimates indicate that the twoweek period prevalence of diarrhoea in under-five children in Ethiopia is approximately 24% 13 .Morbidity-Mortality-and Treatment (MMT) surveys conducted in Ethiopia at various times have revealed five diarrhoeal episodes per child per year, and the two-week incidence to be 16%.Studies conducted in central, rural Ethiopia have revealed diarrhoea to be one of the common causes of under-five mortality, accounting for approximately 8.4% to 27% of all deaths 4 .Diarrhoea is also responsible for 25% to 75% of all childhood diseases, and accounts for approximately 14% of outpatient visits and 16% hospital admissions in Ethiopia 13 .In addition to excess mortality and morbidity, diarrhoea predisposes children to malnutrition, which makes them highly susceptible to other infections, and this has been found to be a major contributor to illness and death, particularly among children in Sub-Saharan Africa 9,14 .Behavioral factors associated with acute childhood diarrhoea include lack of hand-washing, poor infant and young child feeding practices, and lack of child immunizations 8,[15][16][17][18][19] .The objective of the present study was to identify socioeconomic, environmental and behavioral determinants of acute childhood diarrhoea which can be modified to improve child health.

Study area and period
The study was conducted in rural kebeles (or neighborhoods, the smallest administrative unit of Ethiopia) of Derashe The sample size calculation was based on the following assumptions: P1=proportion of diseased with disposal of refuse in pit; P2= proportion of non-diseased with disposal of refuse in pit.From a similar study conducted in Nekemte town with refuse disposal method as the main predictor of outcome (diarrhoea), the sample was 42.86% for cases and 61.47% for control 13 .Therefore, P 1 =0.4286, z 1 =1.96 (95% CI) and P 2 =0.6147, z 2 =0.84 (power of 80%) the proportion of case and control was assumed to be 1:2 With a design effect of 2 and 10% non-response rate, the total sample was estimated using Epi Info v3.5.3 software (http://wwwn.cdc.gov/epiinfo/) as 612 (cases=204, controls=408).
Multi-stage sampling procedure was employed, first by selecting five kebeles from the fifteen using a lottery method.Then all households with under-five children in the selected kebeles were registered through a house-to-house survey by 10 trained enumerators, in order to register all under-five children with diarrhoea or without diarrhoea.After identifying cases and controls by the census, the sample was selected by using a simple random sampling technique.Purposive sampling was used to select participants for qualitative study/FGDs in order to have participants who could provide adequate information.

Study variables
The dependant variable of the study was the occurrence of acute diarrhoea.Independent variables included socioeconomic and demographic characteristics such as income of a family, maternal age, age of child, housing condition, family size, and maternal education.Environmental factors were the availability and utilization of a latrine, waste disposal, and types of water source.Behavioral factors included patterns of hand-washing at critical times, water storage, water treatment, breast feeding status, duration of breastfeeding, and time of introducing supplementary feeding.The definition of some variables follows: • Improved water source: protected spring, protected dug well and piped water • Unimproved water source: pond, river, unprotected dug well and unprotected spring • Hand-washing at critical times: hand-washing after visiting toilet, before eating food and after feeding children.
• Adult member defecation: latrine utilization by adult members of the household, both at home and when outside the home.

Data methods and analysis
Data collection: Quantitative data were collected from the study population using a pre-tested, structured questionnaire via face-to-face interviews in the local language.The qualitative data were collected by using an interview guide.
Data quality assurance: A properly designed and pretested questionnaire was used.Interviewers and data clerks were trained and closely supervised during data collection and entry.Double data entry was used to ensure data quality.
Data analysis: Quantitative data were entered into a predrafted coding sheet in Epi Info by two data clerks.All data analysis was performed using SPSS v16.0 (www.spss.com).Binary logistic analysis with enter method calculating odds ratios (OR) and 95% confidence intervals (CI) was used to estimate the association between the dependent and independent variables.Statistical significance was set at α ≤0.05.In an attempt to identify the relative effects of explanatory variables on the outcome variable, hierarchical multivariable analysis was applied.Explanatory variables with p-value <0.2 were entered into the final regression model based on the likelihood ratio for further analyses in the three different models.
The FGD were tape recorded.Qualitative data were collected using an interview guide that asked participants' perceptions of different determinants of diarrhoeal diseases.
The interview guide was used to guide the flow of discussion and collect data.Careful attention was given to establish the frequency of the occurrence of themes, phrases and expressions as participants described their opinions relative to the specific research questions.The tape recorded data were analyzed according to selected themes based on the guide, and summarized manually according to thematic analysis.
Open code software was used to facilitate the analysis.The written-notes and tape recorded data were transcribed line by line and translated from the local language (Dirayta) into Amharic and then to English.

Ethics approval
The ethical approval and clearance for this research study was obtained from Haramaya University College of Health Sciences Institutional Research Ethics Review Committee.At all levels, officials were contacted and permission from administrators was secured.Participants were provided with an explanation of the purpose of the study and its procedures, and confidentiality was assured.Both written and verbal consent was obtained from the study participants.

Results
Socio-economic and demographic characteristics  1).

Environmental factors associated with diarrhoea
Statistically significant associations were found between acute childhood diarrhoea and sources of water (p<0.001),availability of home-based water treatment (p=0.003),latrine ownership (p<0.001),patterns of adult member defecation (p<0.001) and disposal of infant feces (p<0.001)(Table 2).

Behavioral factors associated with diarrhoea
The acute childhood diarrhoea had significant associations with duration of breast feeding of children (p=0.008),handwashing (p=0.005),covering of food after preparation (p<0.001),consumptions of left-over food stored at room temperature (p=0.009) and measles vaccination (p=0.003)(Table 3).

Qualitative results
Three FGDs of 12 participants in each group were conducted with parents who could give adequate information.Qualitative data analysis revealed that perceived causes of diarrhoea were: child eating dirt, tooth eruption, 'falling of the pancreas' and intestinal parasites.Eating dirt was among the most frequently mentioned causes of diarrhoea, explained by participants as occurring when a child of 5-6 months starts to play on the ground and is likely to eat dirt frequently.Participants also said that 'a child develops diarrhoea after five or six month of age when the child is teething.If the tooth is extracted then the diarrhoea will stop'.'Fallen pancreas' was described as a shift in the position of pancreas that occurs during a fall.When this occurs a child will cry and experiences pain when the legs and abdomen are touched.If the pancreas is returned to its position by traditional healers, both the diarrhoea and the pain cease.
Concerning care of a child with diarrhoea, a number of issues were raised.A female participant said that 'since diarrhoea is caused by worms that cause liquid stool, restricting fluids can prevent it'.Participants also recommended taking the child to a traditional healer to discover the main cause of diarrhoea and to obtain the required care.Taking a child with diarrhoea to a health facility in the first 6 months of life is generally not accepted because the cause of diarrhoea is linked to teething.However, after 6 month of age the child can be taken to a health facility.
Regarding prevention of diarrhoeal diseases and the role of water treatment, one participant said: It is not easy to control diarrhoea using treating water because some of the germs may not be killed with boiling; use of chemical alters the taste of water and it could not be used in this culture.

Multivariable analysis
Hierarchical logistic regression technique was used to assess the relative effect of the explanatory variable on the outcome variable.To avoid an excessive number of variables and unstable estimates in the subsequent model, only variables with a p-value <0.20 were kept in the analyses.
Three nested logistic models were used to estimate the effects of explanatory variable on the outcome variable: • Model 1: examined the joint effects of socioeconomic and demographic factors From the total variables entered in the multivariable regression, four were found to have significant independent association with the diarrhoea (Table 4): 1.The odds of developing diarrhoea was

Discussion
This study found the determinants of acute diarrhoea to be the source of household water, availability of home-based water treatment, latrine ownership and the consumption of left-over food stored at room temperature.
An unimproved water source is among the potential sources of transmission of diarrhoeal diseases.The probability water source contamination depends on whether or not the source is protected.Households using unprotected water sources were three times more likely to have a child with diarrhoea.This finding is in agreement with a study conducted in Gonder 19 but it contradicts a study conducted in the Gojam zone 12 (both locations are in northern Ethiopia).This may be due to the use of water from both protected and unprotected sources, contamination during transport and storage or a lack of homogeneity in water sources in the Gojam study.
Availability of home-based drinking water treatment was an independent predictor of diarrhoeal morbidity in this study.
Children whose families used home-based drinking water treatment such as boiling, the use of chemicals (eg aqua tabs and wuha-agar) and/or filtering were found to have lower odds of getting diarrhoea compared with the children of families not treating water.Because collected water may be contaminated during collection, transportation and storage, this may increase the risk of diarrhoeal diseases.This finding is in agreement with study from Kenya 20 but contradicts previous studies in Ethiopia 12 .This may be due to varying methods of home-based water treatment, and overall differencse in the sanitation of the environment.
In the present study children from households without toilet facilities were more likely to develop diarrhoea compared with children from households with latrine facilities.The presence of toilets increases the chance of safe disposal of feces, one way to decrease contact between the causative organisms of diarrhoea and the host.This finding was similar to other studies 8,14,[20][21][22][23] .The consumption of left-over food stored at room temperature is also an independent predictor of acute childhood diarrhoea due to contamination with and the proliferation of diarrhoea-causing bacteria.The present study cases were almost twice as likely to consume left-over food stored at room temperature than controls.
It has long been established that bottle-fed children are at greater risk of diarrhoea than those who are breast fed, due to milk contamination.Even though it was not statistically significant in the multivariable analysis, the odds of bottle-fed children having diarrhoea was approximately one and a half times greater than for children who were not bottle fed.
Bottle-feeding was significantly associated with diarrhoeal morbidity in the bivariate analysis.Similar findings have been observed in India and Nekemte, western Ethiopia 8 .
The number of under-five children and family size were not independently associated with diarrhoeal morbidity.This is in agreement with a study from Jimma in west Ethiopia 21 but contradicts the finding from the locations Meskanena Mareko Woreda (central Ethiopia) and Gojam Hullet Ejju Ense Woreda (north Ethiopia) where greater family size was associated with diarrhoeal morbidity (OR=2.07& OR=2.3, respectively) 12,23 .This might be due to differences in maternal attention, general living conditions and differences in study design.It is known that as the number of children in a family becomes greater, overcrowding adversely affects hygiene conditions, increasing the chance of contact with pathogens.There is also likely to be competition for the mother's attention and other resources.
The results from FGDs revealed that, due to participants' outdoors work, open defecation was common.This affects individuals' sanitation and is a cause of water source contamination.Where children play in the dirt, it may be contaminated by feces and act as a reservoir of pathogens.A further issue is that using treated water is not a common cultural practice in the society studied, increasing the likelihood of unsafe drinking water.Comprehensive diarrhoeal disease prevention is not easy in a poor community.Changing beliefs and habits, such as avoiding open defecation is difficult when there are no latrines close to markets, farms and other areas of work.

Strength and limitation
The strength of this study is that it was community based and specifically addressed acute childhood diarrhoeal morbidity in a rural community.Additionally, the study employed qualitative and quantitative methods, and had a high response rate (96.7%).However, the retrospective study design was a limitation.As this was a case-control study, there was potential for recall bias; however, this was minimized by using reported-incident cases within a 2 week period.

Conclusion
The determinants of acute childhood diarrhoea in this study were latrine availability, home-based water treatment, source of water and consumption of left-over food stored at room temperature.Also, a very high proportion of the adults interviewed defecated in open land, which can also increase the risk of diarrhoeal disease.The authors recommend the protection of water sources for the wellbeing of the community, and that the community is advised to treat drinking water to reduce the risk of diarrhoea among children.Food hygiene and general sanitation improvement is also highly recommended, as is further research into this vital area of child health.

•
Model 2: used sociodemographic factors with pvalue <0.20 in Model 1, and environmental factors • Model 3: built on variables with p-value <0.20 in Model 2 and included behavioral factors.

Table 2 :
Environmental factors related to acute diarrhoea among study participants in Derashe District, South Ethiopia, 2012 COR, Crude Odds Ratio

Table 3 :
Behavioral factors associated with acute diarrhoea in bivariate analysis, Derashe District, south Ethiopia, 2012 BF, Breast feeding; COR, Crude Odds Ratio.

Table 4 :
Independent predictors of acute diarrhoea in Derashe District, South Ethiopia, 2012