The association of socio-demographic and environmental factors on childhood diarrhea in Cambodia

Background: Diarrhea diseases remain the leading cause of death among children under-five in lower and lower-middle-income countries. This study was conducted to investigate the factors related to diarrhea among children aged 12 to 35 months in Cambodia. Methods: We analyzed cross-sectional data from the Cambodia Demographic and Health Survey 2014 using a combination of household and children’s datasets. A generalized linear mixed model was used to analyze the determinant factors of diarrhea. Results: The survey included 2,828 children aged 12 to 35 months. The prevalence of diarrhea in the last 2 weeks was 16.44% (95% CI: 14.72%-18.31%). Factors significantly associated with childhood diarrhea were: maternal unemployment (AOR = 1.43; 95% CI: 1.14-1.78); the child being male (AOR = 1.25; 95%CI: 1.02-1.53); the presence of unimproved toilet facilities (AOR = 1.17; 95%CI: 1.05-1.31); and unhygienic disposal of children’s stools (AOR = 1.32; 95%CI: 1.06-1.64) when controlling for other covariates. Both maternal age (one year older; AOR = 0.85; 95%CI: 0.78– 0.93) and child age (one month older; AOR = 0.86; 95%CI: 0.78-0.94) had significant negative associations with the occurrence of childhood diarrhea. Conclusion: Childhood diarrhea remains a public health concern in Cambodia. Intervention programs should focus on reducing diarrheal diseases by constructing improved toilet facilities and promoting behavior to improve hygiene, specifically targeting younger mothers.


Introduction
Diarrhea is the second leading cause of death in children under the age of five years, with an estimated 1.7 billion cases of childhood diarrhea and 525,000 deaths caused by diarrhea each year 1,2 . Globally, 88% of diarrhea cases are attributable to poor water, poor sanitation or poor hygiene 3 . Childhood diarrhea is associated with multiple factors, including unimproved drinking water sources 4-7 , untreated water 8-10 , unimproved toilet facilities 6,8,9,11 , unhygienic disposal of children's stools 12-14 , lack of hand washing facilities 15,16 , type and location of residence 11,16 , the child's age 4,13,16 , the child's sex (male) 13 , maternal illiteracy 12,13,17 , the mother's occupation 9,12 , maternal age 14,18 , wealth index 4,19 , and whether or not the child is breastfed 10,15 .
In 2014, Cambodia still had one of the highest prevalence levels of diarrhea among children under the age of five amongst countries in South-East Asia, at 12.8% 20 . By comparison, Myanmar had a prevalence of 10.4% in 2015-16 21 , Malaysia 4.4% in 2016 7 , Laos 6.5% in 2017 22 , Philippines 6.1% in 2017 23 , and Indonesia 14.1% in 2017 24 . According to a 2014 report from UNICEF Cambodia, diarrhea alone accounted for one fifth of the deaths of children under the age of five in Cambodia, and an estimated 10,000 deaths overall each year 25 . However, according to a 2018 report from UNICEF, in 2016, Cambodia had 5,947 total neonatal deaths, of which 20 were due to diarrhea; 5,248 post-neonatal deaths, of which 672 were due to diarrhea (13%); and 692 deaths of children under five due to diarrhea (6%) 26 . This demonstrates that diarrhea is the most common cause of death in Cambodian children. According to the Cambodia Demographic and Health Survey (CDHS) 2014, the prevalence of diarrhea among children aged 12 to 35 months was high compared with other age groups and this age period is known to be crucial for child development and growth 20 .
It is of great importance to understand the factors related to the prevalence of diarrhea among children aged 12 to 35 months. There are no existing studies on the factors affecting the prevalence of diarrhea in this age group, and no national studies on the factors associated with childhood diarrhea in Cambodia have yet been published. This study was therefore conducted to investigate the factors associated with diarrhea among children aged 12 to 35 months in Cambodia.

Ethical statement
This research project received approval from the Khon Kean University Ethics Committee in Human Research (HE632097).
This study uses existing CDHS data and re-analysis was done under the original consent provided by the participants.

CDHS 2014
The CDHS 2014 collected data nationally across the country, which is subdivided into 19 province domains. Its sampling frame consisted of 28,455 eligible enumeration areas (EAs), which comprised the 2008 Cambodian General Population Census (GPC). The sample was proportionately allocated to urban and rural in each domain with a power allocation preventing the oversampling of urban, areas, in order to represent the fact that Cambodia is mainly rural. The stratified sample was selected in two stages. In the first stage, a fixed number of EAs were chosen using probabilities weighted proportional to the size of the EA. In the second stage, 24 and 28 households were picked up from every urban cluster and rural cluster, respectively, through a systematic sampling process with equal probability weighting. 15,825 households, 17,578 women, and 5,190 men were interviewed between the 2 nd June and the 12 th December, 2014, across the country; further details can be found in the CDHS 2014 report 20 .
Population and Sample size Among 7,044 children aged under five years, in our analysis, we included only children aged 12 to 35 months (n=2,828) due to the high prevalence of diarrhea among this age group compared to other age groups. We analyzed the sample power and it was found to provide a suitable degree of power, and was sufficient for this study (0.9627, 0.9682).

Data use
Two raw CDHS 2014 datasets, comprising household data and children's data, were combined for use in this analytical cross-sectional study. All entries and variables in these datasets were included in the study.

Dependent variable
The operational definition of diarrhea used by the CDHS was the occurrence of three or more loose or liquid bowel movements over a 24 hour period, as reported by the mother/caregiver, in any given 24 hour period during the preceding 2 weeks, as described in a French article 27 cited by, and in agreement with, multiple other sources 1,9 . The prevalence of diarrhea was the dependent variable considered in this study. This is referred to the questionnaire thus: "Has (NAME) had diarrhea in the last 2 weeks?" The dichotomous variable childhood diarrhea can take values "1" representing a response of "yes" or "0" representing "no" and "don't know" responses.

Independent variables
Socio-demographic characteristics take the form of continuous variables such as maternal age, child's age, and number of household members and categorical variables such as maternal education (no education/primary/secondary/higher), maternal occupation (employed/unemployed), mother's knowledge of oral rehydration salts (ORS) (good/poor) 28 , exposure to media (yes/no) 29 , sex of the child, breastfeeding (ever/never), deworming (yes/no) 28 , vaccination (ever/never), residence (urban/rural) and wealth index (poorest/poorer/middle/richer/richest) 28 .
CDHS data were organized in 19 province domains, which we regrouped into four regions: Central Plain; Tonle Sap; Coastal and Sea; and Plateau and Mountains 30 (Figure 1). Environmental characteristics were also treated as categorical variables, including drinking water source (improved/unimproved) 31 , whether or not the same source of drinking water was used during wet and dry seasons (same/different), whether or not water was treated before drinking (always/no), type of toilet facility (improved/unimproved) 31 , hygiene (adequate/inadequate) 31 , and disposal of children's stools (sanitary/unsanitary) 32 . The World Health Organization (WHO) guidelines on water, sanitation and hygiene (WASH) were used to classify each WASH facility as either improved or unimproved, and either sanitary or unsanitary according to the WHO/UNICEF Joint Monitoring Programme (Table 1 and Table 2) 31,32 Statistical analysis Statistical data analyses were performed using STATA/SE 14.0 33 as follows.
Categorical variables were analyzed using frequency and percentage. Continuous variables were analyzed as means, standard deviations, and ranges. A weighting variable was used in the form of the woman's individual sample weighting. Cross-tabulations were run with the appropriate sample weights to provide nationally representative results 19 . The svyset command was used to test for complex survey sampling methods used in the original surveys, in order to adjust for differences in the probabilities of sample selection and to avoid using over-sampled strata within the survey data 28 .
The prevalence of diarrhea was estimated as a percentage. The numerator was the number of living children aged 12 to 35 months with an occurrence of diarrhea during the two weeks preceding the interview (i.e. an answer "yes" to, "Has (NAME) had diarrhea in the last 2 weeks?") and the denominator was the number of living children aged 12 to 35 months.
A bivariate analysis with simple logistic regression was performed using the svyset ( svy command). A linearity test was conducted between the continuous variable and dependent variable. Variables associated with diarrhea in the bivariate analyses at a level of p<0.25 were included in the multivariable model 34,35 . Multicollinearity assessment of the independent

Improved Unimproved
Drinking water Piped water, boreholes or tube wells, protected dug wells, protected springs, rainwater, and packaged or delivered water, and provided collection time is not more than 30 minutes for a round trip, including queuing Unprotected dug well, unprotected spring, surface water (river, reservoirs, lakes, ponds, streams, canals, and irrigation channels).
Sanitation Flush and pour flush connected to piped sewer, septic tanks or pit latrines; ventilated improved pit (VIP) latrine, composting toilets or pit latrines with slabs, and that are not shared with other households Flush and pour flush not to sewer/septic tank/pit latrine, pit latrine without slab/open pit, bucket, hanging toilet/hanging latrine, no facility/bush/field Hygiene Availability of a handwashing facility on premises with soap and water No handwashing facility on premises variables was performed by excluding those with a variance inflation factor (VIF) greater than four 36 . Finally, a multivariable analysis was performed using a generalized mixed linear model with four regions picked as 'random effects' corresponding to the various clusters in the sampling design 37 . The backward stepwise procedure was applied as the model fitting strategy. Statistical significance was considered at a threshold of p<0.05 and the adjusted odds ratio (AOR) with 95% confidence intervals (CI) was considered as the magnitude of the effect.
The result was used to map geographical regions in Cambodia, applied using the free and open source geographic information system, QGIS V 2.18.4.

Results
The majority of the children (84.12%) lived in rural areas. Nearly half (44.03%) lived in Central Plain and one third (33.32%) lived in Tonle Sap. The mean maternal age was 28.27 years (standard deviation, SD = 5.89). More than half the mothers (51.08%) attended primary school. Three quarters (75.10%) of the mothers were employed and the average number of household members was five. More than half (51.18%) of the children were male and the mean age was 23.33 months (SD = 6.79). Almost all (96.17%) children had been breastfed; 59.60% had received deworming treatment. Out of 2,828 households, more than half (54.07%) always had treated water to drink; 57.97% had an unimproved toilet facility; while 68.01% used adequate hygiene; and 70.25% used sanitary disposal of children's stool (Table 3).

Bivariate analysis of factors associated with childhood diarrhea in Cambodia
The result from the bivariable analyses revealed that as maternal age increased by a year, the odds of the child suffering from diarrhea decreased 18% (COR = 0.82; 95%CI: 0.73-0.92; p<0.001). The odds of suffering from diarrhea were 49% higher (COR = 1.49; 95% CI: 1.11-1.98; p=0.007) in children whose mother was unemployed compared to employed. As the child's age increased by a month, the odds of the child suffering from diarrhea decreased 17% (COR = 0.83; 95%CI: 0.75-0.92; p<0.001). The odds of suffering from diarrhea was 20% higher (COR = 1.20; 95%CI: 1.04-1.39; p=0.013) in children living in a household with unimproved toilet facilities compared with those with improved toilet facilities. The odds of suffering from diarrhea was 40% higher (COR = 1.40; 95%CI: 1.05-1.87; p=0.020) in children whose stools were disposed of unhygienically compared to children whose stools were disposed of hygienically (Table 4). Further, the child's sex, the number of household members, wealth index, source of drinking water during dry season, whether or not the same source of drinking water was used during wet and dry seasons, and the treatment/non-treatment of drinking water did not reach significance but did meet the pre-determined threshold of p<0.25 for inclusion in the multivariable model. Finally, region (p<0.25) also met the criteria for inclusion in the multivariable model and was used as a random effect. As such, the multivariable analysis was conducted using a generalized mixed linear model with each of the four regions of Cambodia treated as random effects.

Multivariable analysis of factors associated with childhood diarrhea in Cambodia
The multivariable analysis (Table 5) showed that as maternal age increased by a year, the odds of the child suffering from diarrhea decreased 15% (AOR = 0.85; 95%CI: 0.78-0.93; p=0.001). The odds of suffering from diarrhea was 43% higher (AOR = 1.43; 95% CI: 1.14-1.78; p=0.002) in children whose mother was unemployed compared to employed. As the child's age increased by a month, the odds of the child suffering from diarrhea decreased 14% (AOR = 0.86; 95%CI: 0.78-0.94; p=0.001). The odds of suffering from diarrhea was 25% higher (AOR = 1.25; 95%CI: 1.02-1.53; p=0.031) in males compared to females. The odds of suffering from diarrhea was 17% higher (AOR = 1.17; 95%CI: 1.05-1.31; p=0.004) in children living in a household with unimproved toilet facilities compared with those with improved toilet facilities. The odds of suffering from diarrhea was 32% higher (AOR = 1.32; 95%CI: 1.06-1.64; p=0.011) in children whose stools were disposed of unhygienically compared to children whose stools were disposed of hygienically.

Discussion
This is the first study to report factors associated with diarrhea in children aged 12 to 35 months at the national level in Cambodia. Younger maternal age, maternal unemployment, younger child age, being male, lack of unimprovement to toilet facilities, and unhygienic disposal of children's stools were found to be associated with childhood diarrhea.
Socio-demographic characteristics such as maternal age were significantly associated with reduced incidence of diarrhea, in line with studies conducted in Brazil that found younger mothers to be associated with a higher prevalence of diarrhea among their children 18 . It is likely that older mothers have more experience in childcare and feeding. The association of maternal unemployment with the incidence of diarrhea is consistent with a study in Senegal that found children of housewives to have a higher risk of diarrhea compared to children of women who worked in the public or private sector 9 . It is likely that the employment status of the mother will improve a child's quality of living standards and as well as improving hygienic practice and sanitation in the home during feeding and childcare. The child's age had a significant, negative association with the incidence of diarrhea, in line with many studies in Ethiopia and Tanzania 4,14,16 . This might be due to the development of the immune system throughout childhood. Males were more likely to suffer from diarrhea than females, which may simply reflect a natural predisposition of males to develop diarrhea more frequently than females 38 , and is also supported by a previous study conducted in India 13 .
Environmental characteristics such as the lack of improvements to toilet facilities were significantly associated with the incidence of diarrhea, consistent with many studies including a systematic review 4,6,8,11 . Finally, disposal of children's stools was significantly associated with the incidence of diarrhea, consistent with previous studies in Ethiopia, India, and Tanzania 12-14 .
The present findings demonstrate that the quality of sanitation facilities strongly influences the prevalence of childhood diarrhea. Increasing the number of toilet facilities that receive improvements is likely to reduce direct contact with children's stools, and consequently reduce the occurrence of childhood diarrhea in Cambodia.
A limitation of this research study was that it used a crosssectional design with just one outcome measure (diarrhea prevalence) taken as a snapshot at a given point in time and cannot be used to infer a causal relationship. Future longitudinal studies may improve on this. The CDHS 2014 was not fully comprehensive in that it did not cover the WASH factors of hand washing before preparing meals and after defecating. The inclusion of these questions in the survey would give a more comprehensive analysis of hygiene practices in the population. Despite all efforts to prevent bias in the data collection process, the use of self-reporting measures and recall bias may have had an effect on the study findings. Further, the CDHS 2014 captured data by household, rather than by individual person, which may introduce a confound in that it has a tendency to under-estimate the quality of both drinking water source and sanitation facility available.

Conclusion and recommendations
Diarrhea still remains a public health concern among children in Cambodia. The probability of developing diarrhea is strongly associated with maternal unemployment, being male, not having access to improved toilet facilities, or practicing hygienic disposal of children's stools. Conversely, increasing maternal and child age is associated with a reduction in the probability of developing diarrhea.
"Based on these findings, the authors provide the following recommendations.
National: The WASH program should prioritize their efforts in reaching out to younger mothers, mothers of younger children, boys, and unemployed mothers. Guidance should include the use of sanitary methods for disposing of children's stool, as well as water treatment methods, the importance of practicing good sanitation, and maintaining one's health.
Intervention programs should focus on the construction of new sanitary toilet facilities, making improvements to existing toilet facilities, and promoting hygienic behaviors.
Local: Younger mothers should be encouraged to enroll in health education programs. Additional community sanitation facilities should be constructed, and existing facilities should be improved and properly maintained to ensure continued access to sanitation.
Future study: Longitudinal studies are needed to measure the impact of these interventions on multiple aspects of public health, not necessarily limited to the incidence of diarrhea in children.

Data availability
Our study used raw children's and household data from the DHS, Cambodia 2014. Data are free to access for research purposes and can be obtained through the DHS Program after registering and obtaining an approval letter from the Inner City Fund (ICF) (https://dhsprogram.com/data/Access-Instructions. cfm).

Open Peer Review
Bias: Kindly provide efforts to address potential bias 5.
Study size: kindly explain how the sample size was arrived at. How did you arrive at the sample of 2,828? You selected 2,828 sample which represent 565 households based on five persons per household. Recall that you stated that the total number of households was 15,825. This means you selected 4% of the entire household in the study location. This is small and this affects the generalization of the study. Kindly review this along side the eligibility criteria. Also, how were 15,825 households, 17,578 women, and 5,190 men interviewed? What is the significant of interviewing men? The study sample size is clumsy. Kindly review.

6.
Statistical method: How were variables handled in the analyses. The statistical method was not well explained. What is the meaning of improve and unimproved sanitation? Describe the statistical method and sampling strategy used for the study. How were data handled? What kind of technique was used to calculate the prevalence of diarrhea (see abstract)?

Conclusion:
The method for the study is not sufficiently detailed to understand the approach. The author is advised to provide detailed approach and statistical techniques applied.

Results
The demographics of the respondents was not sufficiently reported. 1.
Descriptive statistics was used for the study. Recall in the methodology, under the 2 nd stage of the sampling technique, 24 and 28 households were picked from urban and rural clusters, respectively using equal probability weighting. However, result shows that 84% of the children live in rural areas. Kindly align this narrative.

2.
Kindly explain the result succinctly.

Interpretations/Discussion
Contextually, why are male more susceptible to diarrhea when other studies such as Nwokocha, Chukwudeh, and Damian (2020) have found otherwise. Kindly justify this contextually, this makes the work unique.

1.
If male gender is uniquely susceptible to diarrhea as stated in this study, why was it not included in the conclusion and its implications reflected in the recommendation. However, gender status and disease prevalence are contextual and highly contested in literature.

2.
Validity of the results was not included.

3.
The results were not engaged with other studies to know the gap that was filled in this study.

4.
The contribution to knowledge of the study is not clearly outline in the body of the work. 5.
The implication and contribution of findings to the body of knowledge is missing. 6.

Study design and setting:
The design, settings, location and relevant dates was not clearly stated. This should be stated in the methodology section. Where is Central Plain, and Tonle Sap? These were not described in the methodology. Are they cities or communities in rural or urban clusters? Reply: Taken care. Please, see Figure 1 as well as please, kindly see the CDHS 2014 and Data use.

Participants:
Eligibility criteria was totally missing in this study. Kindly fulfil this requirement. The Author should provide the population size. What is the total number of under-five children (0-5 years) in the country? What is the total number of children 12 to 35 months in the country? What is significant about 12 to 35 months in Cambodia? Reply: Taken care.

Variables:
The variables, indicators, predictors and confounders for the study are not clearly outline. Kindly outline this in alignment with recent WASH analyses pattern e.g Water ladder scale, and Sanitation ladder services (this should be added to what author have on the work). Reply: The authors have retained our originals. Please feel free to check it in limitation.

Data:
The data for each variable of measurement is entirely missing. Kindly provide sources of data for each variables and indicators. Reply: Some of the variable is not provided the sources because it is the secondary data which has been grouped or categorized already. i.e. source of drinking water during wet and dry seasons (same/different), residence (urban/rural), etc.

Bias:
Kindly provide efforts to address potential bias Reply: The authors have retained our originals. Please feel free to check it in limitation.
6. Study size: kindly explain how the sample size was arrived at. How did you arrive at the sample of 2,828? You selected 2,828 sample which represent 565 households based on five persons per household. Recall that you stated that the total number of households was 15,825. This means you selected 4% of the entire household in the study location. This is small and this affects the generalization of the study. Kindly review this along side the eligibility criteria. Also, how were 15,825 households, 17,578 women, and 5,190 men interviewed? What is the significant of interviewing men? The study sample size is clumsy. Kindly review. Reply: DHS used large survey which is not only interviewing for only one outcome. It provides information related to fertility, family planning, maternal and child health, gender, HIV/AIDS, malaria, and nutrition. Therefore, you will get the answer. Please, kindly understand the DHS/CDHS report.

Statistical method:
How were variables handled in the analyses. The statistical method was not well explained. What is the meaning of improve and unimproved sanitation? Describe the statistical method and sampling strategy used for the study. How were data handled? What kind of technique was used to calculate the prevalence of diarrhea (see abstract)? Reply: Our statistical method is already explained. The meaning of improved and unimproved sanitation was outlined in Table 1 and Table 2. Please, kindly see. Our outcome calculation was also outline there. Please, see it carefully.

Conclusion:
The method for the study is not sufficiently detailed to understand the approach. The author is advised to provide detailed approach and statistical techniques applied. Reply: As we explain above in statistical method, the authors have retained our originals in this point.

Introduction
The writing quality of the introduction needs substantial improvement. I am trying to help, but it is difficult and time-consuming. Here are some examples: The 1 st sentence: the definition of diarrhea is not very important in the background as everyone is familiar with the basic definitions, and it takes unnecessary space. However, the authors should provide its operational definition specific to this study in methods. This would ensure the reader that the measurement of the outcome variable was valid. 1.
The 2 nd and 3 rd sentences, paragraph 1, tell almost the same thing and should be combined.
The writing quality makes the background very confusing. For example, the authors stated that, in 2016, 20/5,947 (0.34%) total neonatal deaths were due to diarrhea; 672/5,248 postneonatal deaths were due to diarrhea (13%); and 692 deaths of children under five due to diarrhea (6%). These are hard to understand -how were 'neonatal deaths, post-neonatal deaths, and under-five child deaths' defined? Was it possible that the age groups can be overlapping? The authors continued by stating that "High rates of diarrhea alone account for one fifth of the deaths of children under the age of five in Cambodia, and an estimated 10,000 deaths overall each year." One-fifth means 20%, which is much higher than any percentages mentioned above.

4.
The rationale of study is also relatively weak. For example, the authors stated that, "It is of great importance to understand the factors related to the prevalence of diarrhea among children aged 12 to 35 months." Please elaborate why it is important to study diarrhea among children in this age group. This should also be explained in the methods (study population).

5.
The study's objective is missing. Please provide a clear objective at the end of the introduction. 6.

Methods
Sampling procedures: it is great that the authors described the CDHS-2014's sampling methods. However, the reader needs the information on how the research team reached children aged 12 to 35 months, this study's population.

1.
Please provide operational definition of diarrhea used in the CDHS 2014. This would ensure the reader that the measurement of the outcome variable was valid.

2.
I understand that the sample size was sufficiently large to address the research question. However, it would make more sense to mention minimum required sample size calculation, rather than saying that "The final sample size comprised 2,828 children aged 12 to 35 months, providing a suitable degree of power (0.9627, 0.9682)," which was not understandable.

3.
Statistical analyses: The sentence, "Any independent variables significant at p<0.25 were entered into the initial model" is inaccurate and incomplete. Consider: "Variables associated with diarrhea in bivariate analyses at a level of p<0.25 were included in the initial model." 4.

Introduction
The writing quality of the introduction needs substantial improvement. I am trying to help, but it is difficult and time-consuming. Here are some examples: 1. The 1 st sentence: the definition of diarrhea is not very important in the background as everyone is familiar with the basic definitions, and it takes unnecessary space. However, the authors should provide its operational definition specific to this study in methods. This would ensure the reader that the measurement of the outcome variable was valid. Reply: Taken care.
2. The 2 nd and 3 rd sentences, paragraph 1, tell almost the same thing and should be combined. A: Taken care 3. The 5 th sentence, paragraph 1: consider, "Childhood diarrhea is associated with multiple factors, including…" Reply: Taken care 4. The writing quality makes the background very confusing. For example, the authors stated that, in 2016, 20/5,947 (0.34%) total neonatal deaths were due to diarrhea; 672/5,248 postneonatal deaths were due to diarrhea (13%); and 692 deaths of children under five due to diarrhea (6%). These are hard to understand -how were 'neonatal deaths, post-neonatal deaths, and under-five child deaths' defined? Was it possible that the age groups can be overlapping? The authors continued by stating that "High rates of diarrhea alone account for one fifth of the deaths of children under the age of five in Cambodia, and an estimated 10,000 deaths overall each year." One-fifth means 20%, which is much higher than any percentages mentioned above.
Reply: Taken  -"...and 692 deaths of children under five due to diarrhea (6%)." What is the denominator of the 6%? In the following sentence, the authors stated that 'diarrhea alone account for one fifth of the deaths of children under the age of five in Cambodia." Please clarify these.
-"High rates of diarrhea alone account for...'Diarrhea alone? -It is confusing that this study used data from CDHS 2014, but also cited the prevalence of diarrhea in the same population and from the same data, while claiming that no national studies on childhood diarrhea in Cambodia have yet been published. -The rationale of the study needs improvement.
A: The denominator of the 6% is the "number of under five children", however in the data from UNICEF, they do not put it. They just put only "under five deaths due to diarrhea: 692"; and "% underfive deaths due to diarrhoea: (6%)". According to calculation by the authors, denominator of (6%) is about 11,533. A: Please, see the reference on number 26 which mentioned as "High incidences of diarrhoeal diseases alone account for one fifth of the deaths of children age five and under in Cambodia, and an estimated 10,000 overall deaths annually, largely owing to lack of sanitation and poor hygiene practices". A: Agree with the reviewer. As suggested, we have revised the text to "There are no existing studies on the association in this age group, and no national studies on the associated factors with childhood diarrhea in Cambodia have yet been published". See 2 nd sentence, para 3 on page 3. A: Agree with the reviewer. As suggested, we have added the text "World Health Organization (WHO) guidelines on water, sanitation and hygiene (WASH) were used to classify WASH as either improved or unimproved according to the WHO/UNICEF Joint Monitoring Programme (Table1 and Table 2). See 3 rd sentence, para 2 on page 4; and we also added the Table 1 and Table 2. See on page 11 and 12. 5. Data analyses: -It is not accurate to state this "Continuous data were treated as means, standard deviations, and ranges for analysis." Perhaps something like 'For continuous variables, mean and standard deviations were calculated..." -I am not sure what authors wanted to tell by this "A weighting variable was used in the form of the woman's individual sample weighting." -Any independent variables significant at p<0.25 in bivariate analyses were entered into the initial model. -Multicolinearity assessment was performed...

Methods
A: Agree with the reviewer. As suggested, we have revised the text "Categorical variable were analyzed to provide frequency and percentage. Continuous variable were calculated as means, standard deviations, and ranges". See 1 st sentence, para 3 on page 4. A: Because it is survey data, by providing national representative, we used Woman's individual weighting because child data was accessed by asking for their mother. A: Yes, please see in the "Result" on 2 nd sentence, para 2 on page 6. A: Taken care "Multicolinearity assessment was performed..." Results: ...and one third (33.32%) in Tonle Sap region? A. Taken care. We have revised by adding "lived". See on 2nd sentence, para 1 page 5 2. It should be mean (SD xx). A. Taken care. 3."More than half the mothers (51.08%) attended primary school." Did this include mothers who had no education? A: No, It did not include mothers who had no education 4. Any details to define the breastfeeding -duration, exclusivity...? A: No, in the data showing only "ever breastfed (not currently breastfeed, never breastfed, still breastfeeding"; so we group as "ever and never". 5. Please check this data: "...and 77.97% of them had never been vaccinated." This could be very wrong as the immunization coverage in Cambodia has been globally recognized as very high. A: Agree with the reviewer. As in the data of 2828 showed that 2.15% was no, 22.03% was yes, 0.27% was don't know, and 75.55 was missing. The authors request to delete this variable since it is too much missing, moreover according to literature review this variable is not related with childhood diarrhea, only rotavirus vaccination is associated. Discussion: 1. This section can be improved by extending more in-depth literature in this area and link to the policy implication of the findings. A: Agree with the reviewer. However, we do not extending more in-depth literature. We have retained our original.