Prevalence and factors associated with childhood diarrheal disease and acute respiratory infection in Bangladesh: An analysis of a nationwide cross- sectional survey

Satyajit Kundu (  satyajitnfs@gmail.com ) Department of Biochemistry and Food Analysis, Patuakhali Science and Technology University, Patuakhali 8602, Bangladesh https://orcid.org/0000-0001-9610-1479 Subarna Kundu Statistics Discipline, Khulna University, Khulna, Bangladesh Md. Hasan Al Banna Department of Food Microbiology, Patuakhali Science and Technology University, Patuakhali8602, Bangladesh. Bright Opoku Ahinkorah School of Public Health, Faculty of Health, University of Technology Sydney, Sydney, Australia Abdul-Aziz Seidu College of Public Health, Medical and Veterinary Services, James Cook University, Australia Joshua Okyere Department of Population and Health, University of Cape Coast, Cape Coast, Ghana


Introduction
Protecting the health and wellbeing of children is a crucial component of public health and global health targets. This is exempli ed in the ended Millennium Development Goals (MDGs) and the fairly new Sustainable Development Goals (SDGs), especially SDG 3.2 which seeks to reduce under-ve mortality to as low as 25 per 1000 live births by 2030 [1]. Nonetheless, diarrhea and acute respiratory infection (ARI) remain a major cause of morbidity and mortality among children under-ve worldwide [2], with diarrheal disease constituting about 9% of under-ve mortality (UNICEF, 2016). Available evidence also indicates that ARI constitutes one-fth of all under-ve mortality [3].
The severity of diarrheal disease and ARI cannot be underrated. Beyond its association with childhood mortality, both diarrheal disease and ARI among children have been linked with many child health outcomes [4,5]. In the rst two years of a child where the incidence of ARI and diarrheal diseases is highest, it impedes the physical growth and development of the child, which may later translate into further adverse health events later in their adult life, that is, if the child survives [6].
Contextualizing the study, it is important to note that Bangladesh was successful in achieving the MDGs, speci cally target 4 by attaining a 74% decline in under-ve deaths from 1990-2015 [7]. However, the country remains among the top 15 countries with a high prevalence of childhood mortality attributable to ARI and diarrheal disease [7]. Furthermore, evidence from Bangladesh shows that about 39% of all pediatric hospital admissions and, between 40-60% of total pediatric outpatient department visits were as a result of ARI [8]. This situation calls the attention of researchers to investigate ARI and diarrheal disease among children from the Bangladesh context.
Existing body of literature from Ethiopia [9], Nepal [10], and Uganda [11] have found ARI and diarrheal disease among children to be associated with household socioeconomic status. Evidence from Vietnam [12] also shows that childhood ARI and diarrheal disease were associated with rural residency. Other studies conducted elsewhere have also posited that the sex of the child and access to safe drinking water [10], sanitation [13], level of maternal education and maternal age [11], complementary feeding practices [14], breastfeeding practices [15], waste disposal [9], and household cooking fuel[16] to be signi cantly associated with ARI and diarrheal disease among children.
Current evidence that has used nationally representative data to investigate ARI and diarrheal disease among children in Bangladesh is sparse. To the best of our knowledge, existing current evidence has not looked at ARI and diarrheal disease concurrently. For instance, the study by Sarker et al. [17] was limited to only childhood diarrheal disease (CDD) whereas study by Sultana et al. [7] was limited to ARI. Therefore, our study is the rst current evidence using nationally representative data that investigates both childhood diarrheal disease and ARI in Bangladesh. Hence, the aim of this study is to investigate the prevalence of ARI and CDD, and determine the factors associated with these two childhood morbidities in Bangladesh. Our ndings our timely and relevant in preparing Bangladesh to achieve SDG 3.2, and facilitate the country's exist from the top 15 countries with high prevalence of CDD. Knowing the prevalence of ARI and CDD will inform policy makers in their policy formulation and target setting.
Moreover, identifying the factors associated with ARI and CDD is critical to developing need-based strategies to combat ARI and CDD in Bangladesh.

Data, sampling design, and study population
In this study, the latest Bangladesh demographic and health survey (BDHS) data 2017-18 was used which is the eighth national survey conducted by the National Institute of Population Research and Training (NIPORT) of Health Education and Family Welfare Division of the Ministry of Health and Family Welfare under Training, Research and Development operational plan of 4th HPNSP (Health Population and Nutrition Sector Program) [18]. The BDHS 2017-2018 is a nationally representative cross-sectional household survey data, covering all the 7 administrative divisions of Bangladesh. Two-stage strati ed sampling was used where 675 (227 in urban areas and 448 in rural areas) enumeration areas (EAs) were selected with probability proportional to size at the rst stage and then a systematic sample of 30 households was selected from each EAs which constitute a sample of approximately 20,250 households. Detailed sampling and data collection procedures were given in the nal BDHS report 2017-2018 [18]. In this survey, ever-married women aged 15 to 49 years were approached for an interview in order to collect information on reproductive health, child health, and nutritional status. This leads to a total sample of 8,402 living children aged under ve years born to women living in these households. From these, missing cases were removed and replaced due to missing information leaving a sample size of 7222 for diarrheal disease and 7215 for acute respiratory infection (ARI) of children < 5 years old.

Variable speci cation Outcome variable
The current study focuses on two binary outcome variables: childhood diarrheal disease ("1" indicated the occurrence of diarrhea for the indicated period and "0" indicated no occurrence) and acute respiratory infection of children < 5 years old ("1" indicated the experience of ARI for the indicated period and "0" indicated no experience). A child was considered to suffer from diarrhea if the mother or primary caretaker reported that the child had diarrhea either in the last 24 hours or within the last 2 weeks. In the survey, childhood diarrheal disease was determined if the children had three or more loose or watery stools per day, in the 2 weeks preceding the survey. Similarly, symptoms of ARI of children were identi ed by asking their mothers if their children were ill with cough, and/or short rapid breathing, and/or di cult breathing two weeks prior to the survey [18][19][20]. For analysis, we combined "Yes, last two weeks" and "Yes, last 24 hours" into "Yes" for both ARI and Diarrhea.

Independent variables
The exposure (explanatory variables) of the current study consisted of administrative division (Barisal, Chittagong, Dhaka, Khulna Mymensingh, Rajshahi, Rangpur, and Sylhet), Sex of child (male, and female), current age of child (in months), mothers' age (in years), educational quali cation of the parent, occupation of parent, type of place of residence, number of household members, household wealth index, household access to television and refrigerator, household oor materials, type of cooking fuel, source of drinking water, type of toilet facilities, drugs for intestinal parasites in last 6 months, birth order and nutritional status of the children (wasting, stunting, and weight for age). Our variable selection was based on the previous studies [17,21,22] and available information in BDHS data 2017-18. Nutritional status was measured by three child growth standards including stunting, wasting, and weight for age proposed by the World Health Organization (WHO). A child was said to be stunted whose height-for-age Z-score is < -2 standard deviation (-2SD) from the median. Similarly, A child was said to be wasted and underweighted whose weight for height Z-score and weight for age Z-score is < -2 standard deviation (-2SD) from the median, respectively [23]. Both mother's occupation and father's occupation was categorized as "Home maker/ No formal occupation (Not working, unemployed, student, retired)", "Poultry/Farming/Cultivator (land owner, farmer, agricultural worker, sherman, poultry raising, cattle raising, home-based handicraft)", and "Professional" (Professional/Big business/Technical, Small business/semi-skilled & unskilled) [17].
The source of drinking water was categorized as "Improved (piped into dwelling, piped to yard/plot, public tap/standpipe, piped to neighbor, tube well or borehole, protected well, protected spring, rainwater, tanker truck, cart with small tank, bottled water)" and "Unimproved (unprotected well, unprotected spring, surface water (river/dam/lake/pond/stream/canal/irrigation channel, and other)" for the current study [24,25]. Type of toilet facilities was recategorized into "Improved ( ush -to piped sewer system, ush -to septic tank, ush -to pit latrine, ush -don't know where, pit latrine -ventilated improved pit (VIP), pit latrinewith slab, composting toilet)" and "Unimproved ( ush -to somewhere else, pit latrine -without slab / open pit, bucket toilet, hanging toilet/latrine, others)" [19]. Children under age of ve years are the respondents of the current study whose ages were categorized into 5 categories (< 12 months, 12-23 months, 24-35 months, 36-47 months, 48-59 months). Mother's age was coded as below 20 years, 20 to 34 years, and above 34 years [17]. Father's and mother's education had four categories no education, primary, secondary and higher education. Type of cooking fuel used was recategorized into "Clean fuel (electricity, lique ed petroleum gas (LPG), natural gas, and biogas)" and "Polluted fuel (coal/lignite, charcoal, wood, straw/shrub/grass, agricultural crops, and animal dung)" [26]. Birth order of the respondent was categorized as rst child, second child and third and above. The household wealth index is a measure of living standard. DHS calculated household wealth index using Principal component analysis (PCA) based on household's ownership of selected assets, such as televisions and bicycles; materials used for housing construction; and types of water access and sanitation facilities which had ve wealth quintiles (poorest, poorer, middle, richer, richest) [27]. Family size or number of household family members were divided into two categories (≤ ve members and > ve members). Floor materials were categorized into "Improved (cement, ceramic tiles, vinyl asphalt strips, parquet, polished wood)" and "Unimproved (earth, sand, dung, wood planks, palm, bamboo)"[28].

Data processing and analysis
Data management and analyses were done using SPSS version 25.0, and R version 4.0.1 for the children's data set (KR le). Descriptive weighted prevalence was computed to show the prevalence of diarrhea and ARI among children under 5 years of age accounting the strati cation and sampling weights. The weights were obtained from the women's individual sample weight dividing by 1000000. Frequencies and category-based percentages were showed to present the descriptive characteristics of study participants. Chi-square test was performed to identify the association between considered risk factors and Diarrhea as well as ARI. Binary logistic regression was carried out to assess the adjusted and crude effect of risk factors on diarrhea and ARI among children of age under ve years. Adjusted odds ratio (AOR) and crude odds ratio (OR) with 95% CI were performed in the analysis of the current study. A p value of less than 0.05 was considered to be statistically signi cant.

Patient and public involvement
No patient involved

Background characteristics
After data cleaning, a total of 7222 mothers having children < 5 years old were included in case of diarrheal disease, and 7215 mothers who had children < 5 years were included in case of ARI in the present study. The age of the children was categorized with an 11 months interval and was almost equally distributed for the age category. More than half of the mothers were home maker who had no formal occupation. Most of the children (64.8%) in the study were from the rural area. Considering the measurement of nutritional statuses, 30.2%, 22.4%, and 8.2% of children were identi ed to be stunted, underweight, and wasted, respectively. Most of the households had an improved toilet facility (68.2%), and an improved source of drinking water (97.7%) ( Tables 1 & 2). The results and the associated χ 2 tests shown in Table 1 indicate that the incidence of childhood diarrheal disease in Bangladesh is signi cantly associated with the age of children, mothers' age, household wealth index, and drug intake for intestinal parasites. The associated χ 2 tests regarding ARI of children in Bangladesh shown in Table 2 reveal that region, age and sex of children, mothers' age, and household having television and refrigerator were signi cantly associated with ARI.

Prevalence of diarrheal disease and ARI
The overall prevalence of diarrheal disease among children < 5 years old was 4.91%. The highest diarrheal prevalence was found among children from Barisal region (6.78%), followed by Rajshahi region (5.93%) (Fig. 1). Among the age groups, children aged between 12 to 23 months (9.36%) were most vulnerable to diarrhea, followed by < 12 months old children (5.92%). Children of young mothers aged between 20 to 34 years old suffered from diarrhea more (6.50%) than those of older mothers aged above 34 years old (2.91%). Children of mothers with no formal education (6.28%) were found to be more vulnerable to diarrheal disease. Based on the ve quintiles of the household wealth index, the diarrheal prevalence was higher among children from the poorest families (6.52%). A high prevalence was observed in children (5.37% vs 4.27%) who did not intake drugs for intestinal parasites in the last 6 months prior to data collection, and who were stunted (5.01% vs 4.66%). A high prevalence was observed in households that had unimproved oor materials (5.40% vs 4.04%) ( Table 1).
The overall prevalence of ARI among children < 5 years old was 3.03%. The highest prevalence of ARI observed in Rangpur region (5.47%), followed by Barishal (4.11%) region of Bangladesh (Fig. 2). Children aged between 12 to 23 months (4.10%) were found to be more vulnerable to ARI, followed by < 12 months old children (4.07%). A higher prevalence of ARI was found among children of mothers aged 20 to 34 years (5.28%). ARI prevalence was higher among male (3.63%) than female children (2.36%). The prevalence of ARI is highest (3.56%) among the children whose mothers had no formal education, and a similar pattern was also observed with the educational status of fathers. Based on the socioeconomic status of the households, ARI prevalence was higher (3.98%) in the households with lower socioeconomic status ( Table 2).   (Table 3).  The bolded values (ORs) indicate the statistical signi cance.

Discussion
Although Bangladesh met the MDG targets, it still remains among the top 15 countries with high cases of CDD and ARI [7]. Therefore, to ensure that there is a greater understanding of the situation of ARI and CDD in Bangladesh, as well as facilitate its potential to achieve SDG 3.2, we investigated the prevalence of ARI and CDD, and determined the factors that are associated with these two childhood health events. Our study indicates that the prevalence of CDD and ARI was 4.91% and 3.03% respectively, with the prevalence for both outcomes being highest for children born to younger mothers (20-34 years), mothers with no formal education, those in lower socioeconomic status. The ARI prevalence observed re ects a trend of decline in the prevalence of ARI from previous rounds of the BDHS survey reports [7,29,30].
Concerning the factors associated with ARI and CDD, the results of our study show that it was signi cantly associated with the sex of the child, with male children being at higher risk of ARI or CDD.
This nding is in line with earlier studies from Bangladesh [7], Ethiopia [14], Nepal [10], Sudan [31], and Thailand [32] that have reported higher risk of ARI and CDD among male children. This could probably be due to differences in genetics that places males at higher risk of diseases and other health events compared to women [7]. Another plausible explanation could be due to higher reporting for male children, which is reinforced by mothers' preference for the male child [33]. As such, they are able to notice changes in the health status of the male child early and report to the hospital accordingly.
There is a myriad of evidence suggesting that ARI and CDD are most prevalent in the rst two years of a child's life, thus, making children < 12 months and those between 12-23 months being at higher risk of ARI or CDD [13,17,34]. Our nding provides con rmation of this association. Moreover, the nding from this study indicates that although the prevalence of ARI and CDD is higher within the rst two years of a child's life, the risk of developing ARI or CDD is highest in children between 12-23 months, which supports Sarker et al. [17] ndings that the prevalence of CDD is highest for children aged 1 to 2 years compared to those less than a year old. However, our ndings that younger child age is associated with higher prevalence and risk of ARI and CDD could be explained from the point that, the immune system of the child is delicate at that early age, thereby putting them at increased risk of infections [35]. Furthermore, children which such early years tend to be heavily dependent on their mothers and therefore, require appropriate feeding that is proportional to their age [17]. Hence, when mothers slack in their responsibilities to provide safe and appropriate feeding to the children at that age, then their risk of ARI and CDD increases.
We found a signi cant association between household wealth status and risk of CDD, with children belonging to the poorest household having greater likelihood of developing the diarrheal disease. This corroborates previous related studies from Bangladesh[16,17] and Nepal [10] that also reported higher risk of CDD among children belonging to poor households. This may possibly be justi ed from the perspective that; poorer households have di culty in meeting their nutritional needs and adopting appropriate feeding practices which may exacerbate their risk of diarrheal infection [10]. This is further iterated in our nding that stunted children had a higher risk of CDD.
Congruent to existing literature [3,11], our study indicates that there is signi cant association between formal education and ARI, with lower odds of ARI being reported among children whose mothers had formal education compared to those whose mothers had no formal education. A plausible justi cation for this nding could be that children spend more time with their mothers; therefore, the mother's educational attainment will re ect in the quality of care that they will provide to their child, which may either increase the risk or protection against ARI [36]. Hence, emphasizing the need to promote formal education among women.
Beyond these individual and household factors, we found statistically signi cant association between geographical region and the risks of ARI and CDD. It was found that children who lived in Rangpur region and Barisal region were at higher risk of developing ARI or CDD. This is consistent with previous studies from Bangladesh [17] that also found similar ndings in relation to the regional differences in the prevalence of ARI and CDD. Begum and her colleagues also reported a higher diarrheal prevalence among children < 5 years old in the similar setting and found that water, sanitation and hygiene (WASH) education to the mothers was effective to reduce the burden of diarrhea [37]. According to Sarker et al [17], regions like Barisal are densely populated and is also characterized by the existence of more rivers and water reservoirs that create an enabling environment for diarrheal disease to spread among the population. Perhaps, this could be the reason for the high prevalence of ARI and CDD within the Barisal region.

Conclusion
Bangladesh met the MDG targets but still remains among the top 15 countries with high cases of CDD and ARI. This study sought to investigate the prevalence of ARI and CDD, and determine the associated factors. Based on the ndings from the study, we conclude that the prevalence of ARI and CDD in Bangladesh has reduced when compared with previous studies and previous rounds of the BDHS. We also conclude that there are individual, household and geographic factors that exacerbate the risk of ARI and CDD (children born to mothers of younger age, mothers with no formal education, belonging to lower socioeconomic households, being a male child, being stunted, and residing in Barisal and Rangpur regions). Therefore, we recommend that the government of Bangladesh commit resources, policies and interventions geared towards ARI and CDD reduction to the identi ed at-risk groups. Also, there is the need to augment formal education for women in Bangladesh to accelerate the realization of SDG 3.2, and complete eradication of ARI and CDD in the country. Further studies can be conducted to explore how culture also permeates the dynamics of ARI and CDD in Bangladesh, in order to ensure that interventions and policies developed are culturally sensitive to facilitate acceptance and adherence.

Declarations
Author contributions