Malnutrition among under-5 children and its determinants in the southwestern coastal region of Bangladesh: A community-based study

Abstract

Objective

The study examines the prevalence of stunting, the most common form of malnutrition and its determinants among children under-5 in the southwestern coastal region of Bangladesh.

Design

Cross-sectional survey.

Settings

Two coastal villages of the Bagerhat and Khulna districts of Bangladesh.

Participants

Participants were selected using the simple random sampling technique with the following criteria: mothers who had at least one under-5 child and resided in the study area for not less than three consecutive years.

Outcome measures

To assess the nutritional status particularly, stunting (height for age) in children under-5, a standardised anthropometric measurement tool, such as a height-measuring vertical scale, was used and categorised into severely stunted, moderately stunted and normal height.

Results

The study found that 57.5% of children were severely and 29% were moderately stunted. Findings revealed that mothers’ age, prior experience of under-5 mortality, mothers’ working status and age of the children were significant determinants of stunting prevalence among children under-5. Moreover, children of mothers aged 21–30 years (AOR = 2.190; 95% CI: 1.039 to 4.613; p = 0.039) and children of mothers without prior experience of under-5 mortality (AOR = 4.426; 95% CI: 1.461 to 13.405; p = 0.009) were more possibly to be severely and moderately stunted, respectively. Additionally, children of non-working mothers were more likely to be moderately (AOR = 4.037; 95% CI: 1.382 to 11.792; p = 0.011) and severely stunted (AOR = 2.538; 95% CI: 1.033 to 6.238; p = 0.042). Moreover, children aged ≤24 months (AOR = 0.151; 95% CI: 0.034 to 0.661; p = 0.012) and 25–36 months (AOR = 0.195; 95% CI: 0.046 to 0.829; p = 0.027) were less possibly to be moderately stunted.

Conclusion

The study recommends training young mothers on effective child-rearing practices, creating income opportunities for women in rural Bangladesh, implementing local awareness programmes about appropriate growth standards for children and expanding facility-based healthcare centres in rural areas for better access to quality healthcare.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• A community-based approach was employed to assess the prevalence of stunting and its determinants which are important for implementing policies at the local level.

• Different socioeconomic, demographic, household and health-related factors were considered to assess their influence on stunting among under-5 children which is crucial to decrease the prevalence of childhood stunting.

• The cross-sectional nature hindered the ability to establish causal relationships.

• Recall bias regarding the birth weight of the children and other health-related issues might have impacted the findings.

Introduction

Childhood malnutrition is a major public health concern worldwide¹ and stunting is one of the crucial indicators of malnutrition which is defined as a child’s height being <−2 SD below the median child growth standard for the same sex and age.² Malnutrition in earlier life affects children’s both physical and mental development as well as increases the risk of many diseases in later life.³

Global estimates of stunted children under-5 in 2022 were projected to reach 149 million.⁴ However, stunting is more common than wasting or underweight in developing nations and has the highest incidence (40%) in South Asia and sub-Saharan Africa.⁵,⁷ In 2019, more than half of all stunted children under-5 were from Asia, and two out of every five were from Africa as well in Southern Asia, 30.5% of children under-5 are stunted.⁸

In Bangladesh, stunting is also the most common form of childhood malnutrition which is a foremost public health concern and a prominent cause of under-5 child mortality.⁹ Bangladesh continued to be one of the developing nations in Asia with a greater stunting prevalence as 31% of under-5 children were stunted in 2017¹⁰ and 26.4% in 2023.⁸ Children who are stunted have worse health outcomes, such as a decreased ability to recover from illness, impaired cognitive function, including memory loss, reduced learning accomplishment and a higher attention deficit disorder.^{9 11} Additionally, stunting is a significant cause of illness burden and accounts for approximately 3.1 million or 45% of all child deaths globally each year.^{12 13}

In developing nations, stunting prevalence starts to increase at around 3 months of age and then starts to decline at around 2 years of age.¹⁴ Existing literature revealed several factors related to the household, parents and children are linked to stunting prevalence. With reference to Bangladesh, previous research suggests that specific factors increased the stunting prevalence among children like children’s age, sex, birth order, and size at birth, mother’s height, body mass index (BMI), age at first birth, occupation, and exposure to mass media, parental education level, household wealth, place of residence, household food insecurity, receiving prenatal care, place of delivery, and toilet facility.⁹¹⁵,²¹

Studies in other Asian and African countries including Nepal, Indonesia, Ethiopia and Pakistan stated that household-related factors including household wealth, socioeconomic status, household size, joint family system, the number of children, access to improved sanitation facilities, household food insecurity are related to stunting and stunting is less common in children from higher socioeconomic status than those from lower socioeconomic status.²²,²⁷ Besides, parental factors, such as parental education level, mothers’ age at the child’s birth, height, BMI and birth interval, are associated with stunting as narrated in Pakistan, Indonesia, Tanzania and Ethiopia.^{2324 28},³⁰ Furthermore, biological factors include infections (such as diarrhoea), low birth weight, birth order, breastfeeding, child’s age and sex.^{24 29 31 32} Other factors such as ethnicity, geographical region, unprotected drinking water, place of delivery, type of delivery assistance, inadequate dietary intake, poor dietary diversity, history of bottle feeding, unvaccinated status, poor health status and environmental factors (such as climate change) are the significant predictors of stunting incidence among children under-5.²⁵,^{2729 33 34}

Bangladesh is one of the South Asian countries with a higher prevalence of stunting. Identifying the possible determinants of stunting is an essential first step in accelerating the stunting reduction rate. Despite several studies conducted in Bangladesh at the national level¹⁵,^{1935 36} which are mostly secondary data based, the prevalence and determinants at the community level have not been sufficiently emphasised,³⁷ which generates interventions challenging in such situations. Therefore, it is imperative to have comprehensive and tangible data that can fill these gaps and provide useful information that helps policy-makers develop effective community-based intervention strategies to reduce stunting prevalence among children under-5 and enhance the health of the coming generation. Thus, the study intends to investigate the stunting prevalence, a form of malnutrition in children under-5, and its determinants in the southwestern coastal region of Bangladesh.

Methods

Study area

The study intends to investigate the prevalence of stunting and its determinants in children under-5 in the southwestern coastal region of Bangladesh. Therefore, the study was carried out following the cross-sectional survey design. Two coastal areas namely Joymoni village under Mongla Upazila in Bagerhat District and Sutarkhali village under Dacope Upazila in Khulna District of Bangladesh were selected to carry out the study. These areas are selected because the socioeconomic, cultural and environmental aspects are different compared with other areas of Bangladesh. In addition, these are the remote areas where people are struggling to fulfil their basic needs such as food, water, sanitation and access to health facilities. Besides, the scarcity of safe drinking water is a major issue of concern for the local population in coastal regions³⁸ along with sanitation facilities which might influence children’s health and nutritional status. Furthermore, no study has been conducted on this issue among the younger age group in this region.

Participants, sampling and data collection

The following criteria were considered while selecting participants in this study: (1) mothers who had at least one under-5 child and (2) mothers who had been residing in the study area for not less than three consecutive years. A census with two personnel was conducted in May 2023 to get information about the population considering the criteria of the participants. During the census, children’s names and ages as well as their parents’ names were gathered. Following the census, the aforementioned details were included in a population list that was assigned a serial number. The study involved a total population of 1198 mothers based on the conducted census in the study area. Using Cochran’s formula for simple random sampling, a sample of 400 mothers was determined with a 95% confidence level and a margin of error of 5%.³⁹ The selection process employed a lottery method to choose participants from the population list, promoting a fair and random selection. Data collection occurred at the household level, using a replacement method to enhance validity. For mothers with multiple under-5 children, the study focused on collecting information about the youngest child, allowing for more precise insights into malnutrition determinants.

A semi-structured interview schedule was used for data collection from July to August 2023. The interview schedule contained eight different parts covering a variety of topics including socioeconomic, demographic, health-related and household information, the stunting status of under-5 children, mothers’ knowledge about child health, mothers’ decision-making power, household wealth and household food insecurity. To effectively assess children’s nutritional status, particularly in relation to stunting (height for age), we implemented a standardised anthropometric measurement tool, such as a height-measuring vertical scale. Data collectors carefully measured each child’s height while ensuring that the child’s head, shoulders, buttocks and heels were aligned with a flat surface. Children’s heights were recorded in centimetres.

Measures

Socioeconomic information

Socioeconomic information includes religion (Muslim and non-Muslim), place of residence (Joymuni and Sutharkhali village), parental education (non-literate, primary, secondary and tertiary), occupation of mother (not working and working mother), father’s occupation (business, self-employed, day labour, employee and others), father’s monthly income (Bangladeshi Taka [BDT] ≤10 000, 10 001–20000 and ≥20 001) and monthly family income (BDT ≤10 000, 10 001–20000 and ≥20 001). Mother’s mass media exposure was categorised into yes = 1 (if the mother was exposed to any of watching television, listening to radio, and reading newspapers and magazines) and no = 0 (if she was not exposed to anyone).

Demographic and health-related information

Demographic factors such as mother’s age at first birth (<18 years, 18–22 years and ≥23 years), the number of living children (1, 2 and ≥3), the number of under-5 children (1 and 2), history of under-5 child mortality (no and yes), age of the children (≤ 24 months, 25–36 months, 37–48 months and 49–59 months), sex of the children (girl and boy) and birth order of the children (1, 2 and ≥3). Besides, health-related information including mothers’ BMI was measured by their height and weight and grouped into underweight (≥18.5), normal weight (18.5–25), overweight and obesity (≤25).⁴⁰ Besides, children’s birth weight was measured in kilograms (kg) and classified according to the WHO⁴¹ recommended categories such as underweight (< 2.5 kg), normal weight (2.5–3.9 kg) and overweight (4 kg or above). Additionally, the child’s primary immunisation status (completed, ongoing and no), exclusive breastfeeding (no and yes) and children’s health problems were assessed whether they had been suffering from any health problems during the last month and categorised into (no and yes).

Household information

Household information such as the source of drinking water was categorised into improved sources (such as deep tubewell/rainwater/bottled water) and non-improved sources (such as unprotected well/surface water, e.g.pond, canals and rivers). Similarly, sanitation facilities were classified as either improved toilet facility (such as flushing to pit latrine/pit latrine with slab) and ‘non-improved toilet facility’ (pit latrine without slab/open pit, hanging toilet/open space/no facility).⁴²

Stunting prevalence among under-5 children

The stunting status of children under-5 is the outcome variable in this study. This was determined by measuring the children’s height for age and classified into three categories: severely stunted (Z-score is below −3.0), moderately stunted (Z-score is below −2.0) and normal height (Z-score is ‘0’ to less than +2).⁴³

Mothers’ knowledge about child health

An index with 10 statements related to child health issues (with dichotomised responses: yes = 1 and no = 0 was developed) to assess mothers’ knowledge about child health, and categorised into low (1–6), moderate (7–8) and high (9–10).

Mothers’ decision-making power

Adapted from Mahmud et al,⁴⁴ a 10-item dichotomised scale was used to assess mothers’ decision-making power with yes = 1 and no/do not know = 0. The total score was classified into low (0–5), moderate (6–7) and high (8–10).

Household wealth

The household wealth was assessed following the wealth index of 27 items used in the Bangladesh Demographic and Health Survey¹⁰ measured on a dichotomised scale of yes = 1 and no = 0. The highest score is 16 and the lowest score is 1 with a median score of 7. Therefore, the household wealth index was categorised into three groups such as low (below the median score of 1–6), moderate (7–11) and high (12–16).

Household food insecurity

Household food insecurity access scale (index) was used which was developed by Coates et al⁴⁵ to measure household food insecurity. Then it was classified into food secure, mild food insecure, moderately food insecure and severely food insecure households in accordance with the scale.

Data analysis

To analyse the data, Statistical Packages for the Social Sciences V.21. was used. Percentage analysis was performed to determine the stunting prevalence in children under-5. Besides, Pearson’s χ² test of independence and Fisher’s exact test (F-test) (when the cell count is less than 5) were done to test the significant relationship between the outcome variable and predictors using p < 0.10 significance level. Subsequently, significant variables in bivariate analyses were considered in conducting the multinominal logistic regression analysis. Multinominal logistic regression results were presented by adjusted odds ratio (AOR) with 95% CIs and a significance level of p < 0.50.

Patient and public involvement

None.

Results

Prevalence of stunting among under-5 children

Table 1 depicts information about stunting prevalence among children under-5. Findings reveal that 57.5% of the children are severely stunted, and 29%t are moderately stunted whereas only 13.5% of the children have normal height in relation to their age.

Table 1. Prevalence of stunting among under-5 children.

Stunting status (height for age)	Frequency	Percentage
Normal height	54	13.5
Moderately stunted	116	29.0
Severely stunted	230	57.5

Bivariate analysis of the association between stunting among under-5 children and its determinants

Findings from bivariate analysis (see online supplemental table 1) showed that place of residence (χ² = 5.443, p = 0.066), mothers’ age (χ² = 9.037, p = 0.060), mother’s occupation (χ² = 6.196, p = 0.045), mothers’ decision-making power (χ² = 10.604, p = 0.031), the number of the under-5 children (F-test = 5.211, p = 0.071), history of under-5 child mortality (χ² = 7.039, p = 0.030), age of the children (χ² = 39.075, p < 0.001), sex of the children (χ² = 4.968, p = 0.083) and children’s primary immunisation status (F-test = 12.820, p = 0.009) were statistically significantly related to stunting prevalence among children under-5.

Multinominal logistic regression analysis of the predictors of stunting among under-5 children

Out of 30 variables, 9 were found statistically significant in bivariate analysis in relation to stunting prevalence among children under-5. These variables were considered to conduct the multinominal logistic regression analysis (see table 2). Here, the stunting status of the children (normal height, severely stunted and moderately stunted) was the outcome variable and normal height was the reference category. On the other hand, the predictors were place of residence, mother’s age, occupation, decision-making power, history of under-5 child mortality, the number of under-5 children, age and sex of the children, and children’s primary immunisation status. Results revealed that the mother’s age and occupation, history of under-5 child mortality and age of the children were significantly related to stunting prevalence among children under-5.

Table 2. Multinominal logistic regression analysis of the predictors of stunting among under-5 children.

Predictors	Moderately stunted					Severely stunted
	B (SE)	p value	AOR	95% CI (lower-upper)	B (SE)	p value	AOR	95% CI (lower-upper)
Place of residence
Joymoni village	−0.074 (0.352)	0.834	0.929	0.466 to 1.850	0.312 (0.320)	0.329	1.367	0.730 to 2.559
Sutarkhali village (R)
Mothers’ age
≤20	0.702 (0.704)	0.318	2.019	0.508 to 8.021	1.055 (0.612)	0.085	2.872	0.866 to 9.526
21–30	0.803 (0.421)	0.057	2.232	0.978 to 5.095	0.784 (0.380)	0.039*	2.190	1.039 to 4.613
≥31 (R)
Mothers’ occupation
Non-working mother	1.395 (0.547)	0.011*	4.037	1.382 to 11.792	0.931 (0.459)	0.042*	2.538	1.033 to 6.238
Working mother (R)
Mothers’ decision-making power
Low (0–5)	0.140 (0.436)	0.749	1.150	0.489 to 2.701	0.134 (0.387)	0.729	1.143	0.535 to 2.443
Moderate (6–7)	0.212 (0.419)	0.613	1.236	0.544 to 2.810	−0.525 (0.397)	0.185	0.591	0.272 to 1.287
High (8–10) (R)
Number of under-5 children
1	−0.328 (0.750)	0.663	0.721	0.166 to 3.137	−0.728 (0.655)	0.266	0.483	0.134 to 1.743
2 (R)
History of under-5 child mortality
No	1.488 (0.565)	0.009†	4.426	1.461 to 13.405	0.751 (0.452)	0.096	2.120	0.875 to 5.137
Yes (R)
Age of the children
≤24 months	−1.891 (0.753)	0.012*	0.151	0.034 to 0.661	−0.572 (0.720)	0.427	0.565	0.138 to 2.315
25–36 months	−1.637 (0.739)	0.027*	0.195	0.046 to 0.829	−0.862 (0.720)	0.231	0.422	0.103 to 1.730
37–48 months	−0.513 (0.758)	0.499	0.599	0.136 to 2.646	−0.626 (0.755)	0.408	0.535	0.122 to 2.351
49–59 months (R)
Sex of the children
Girl	0.170 (0.348)	0.626	1.185	0.599 to 2.344	−0.333 (0.317)	0.293	0.717	0.386 to 1.333
Boy (R)
Children’s primary immunisation status
No	0.453 (1.197)	0.705	1.573	0.151 to 16.425	0.441 (1.122)	0.694	1.554	0.172 to 14.003
Ongoing	−0.025 (0.562)	0.965	0.975	0.324 to 2.936	0.270 (0.474)	0.569	1.310	0.517 to 3.318
Completed (R)

^*Significant at 5%

^†Significant at 1%

AORadjusted odds ratioBUnstandardized coefficientCIConfidence intervalRReference categorySEStandard error

Furthermore, children of mothers aged 21–30 years were 2.190 times more possibly to be severely stunted compared with mothers aged ≥31 years (AOR = 2.190; 95% CI: 1.039 to 4.613; p = 0.039). In addition, children of non-working mothers were 4.037 times and 2.538 times more likely to be moderately (AOR = 4.037; 95% CI: 1.382 to 11.792; p = 0.011) and severely stunted (AOR = 2.538; 95% CI: 1.033 to 6.238; p = 0.042) compared with children of working mothers, respectively. Besides, children of mothers who did not have a history of under-5 mortality were 4.426 times more likely to be moderately stunted (AOR = 4.426; 95% CI: 1.461 to 13.405; p = 0.009) than the mothers who experienced under-5 child mortality. Moreover, children aged ≤24 months were 0.151 times more prone to be moderately stunted (AOR = 0.151; 95% CI: 0.034 to 0.661; p = 0.012) and children aged 25–36 months were 0.222 times less likely to be moderately stunted (AOR = 0.195; 95% CI: 0.046 to 0.829; p = 0.027) compared with those who were within the age group of 49–59 months.

Discussion

Stunting is still a serious public health issue in Bangladesh like in other developing nations.¹⁶ The present study reported that 57.5% of the under-5 children are severely stunted and 29% are moderately stunted which is higher compared with the national stunting prevalence (30.7%) in Bangladesh.¹⁰ The stunting prevalence is also higher compared with other South-Asian and African countries like 40% in Pakistan,²⁵ 47% in Nepal⁴⁶ and 38% in Ethiopia.³⁰ However, consistent with the current study, another community-based study conducted in Nepal found a higher stunting prevalence among children at 56.7%.³³ Nevertheless, the discrepancies in the findings may be due to the differences in the nature of the study area, livelihood strategies of the coastal people, socioeconomic factors, healthcare accessibility and differences in education and cultural practices.

Moreover, the findings of the current study uncovered that mothers’ age and occupation, history of under-5 child mortality, and children’s age were the significant determinants of stunting prevalence among under-5 children. It has been found that children of mothers aged 21–30 years were more likely to be severely stunted compared with mothers aged ≥31 years. This finding is consistent with previous studies conducted in Ghana,⁴⁷ Tanzania⁴⁸ and Pakistan²⁸ that found adolescent mothers had a higher risk of having stunted children than adult mothers. This could be interpreted by the fact that adolescent mothers might lack nutrition knowledge, the time of complementary feeding initiation and dietary diversity which increases the risk of stunting prevalence among children under-5.^{49 50} However, another study carried out in South Africa noted that maternal age was not a significant predictor of malnutrition, suggesting that other factors, like poverty and lack of access to healthcare and education, maybe more important determinants of child malnutrition.⁵¹

Consistent with an earlier study, the current study showed that mothers’ employment status was significantly connected to stunting prevalence and children of non-working mothers have a higher likelihood of being moderately and severely stunted than children of working mothers.⁵² One plausible justification could be that working women possess better education and decision-making power.⁵³ Additionally, they tend to be more aware of their children’s health and nutrition, promote improved hygiene practices, are better equipped to give care and use health services more effectively.¹⁶ Besides, parental socioeconomic status influences their children’s right to better healthcare and welfare⁵⁴ particularly, educated mothers may earn more money, have better access to nutritious food, and have an improved standard of living, all of which would improve their children’s nutritional status. Furthermore, the majority of the rural women in Bangladesh are unemployed which is attributed to their lower career aspirations⁵⁵ as a result they remain unemployed and often face financial difficulties and struggle to afford proper healthcare for their children’s illnesses, leading to frequent illness and inadequate healthcare, resulting in child malnutrition.⁵⁶ On the contrary, an Indian study found that children of working mothers particularly those who were cultivators or labourers had a greater risk of stunting⁵⁷ as women who work in agriculture in India typically have lower educational backgrounds and come from poor families. However, some previous studies found no evidence of a substantial relationship between parental work status and children’s nutritional status.^{15 58}

Moreover, the current study depicted that children of mothers who did not experience under-5 child mortality have higher chances of being moderately stunted than the mothers who experienced under-5 child mortality. This could be illustrated by the fact that mothers’ prior experience of under-5 mortality may foster their awareness about child health, healing and nutritional requirements. As a result, they are more concerned about their children’s overall growth and development, a supplement of nutritious foods and appropriate utilisation of healthcare services during a child’s illness. The current study explores this issue as an influencing factor of childhood stunting, a crucial aspect that has been overlooked in existing literature, necessitating further investigation.

The age of the children is a critical predictor of stunting prevalence, and the current study observed that children aged ≤24 months and 25–36 months are less prone to be moderately stunted than their counterparts. The lower prevalence of stunting among younger age groups might be due to the protective effects of breastfeeding since all children in Bangladesh receive breast milk during the first 2 years of their lives.¹⁹ Besides, the higher incidence of stunting among older children (aged 49–59 months) may be caused by an improper food supply that provides insufficient quantities and quality of nutrients along with the increasing pattern of infectious diseases among children with age. However, this finding is inconsistent with existing literature that found children aged more than 6 months had a higher likelihood of being stunted compared with children aged ≤6 months.^{9 19 59}

Limitations

Despite some notable findings, the study has the following imitations. The cross-sectional nature hindered the ability to establish causal relationships, which is one of its limitations. There might be a recall bias regarding the birth weight of the children and other health-related issues. Notwithstanding these drawbacks, the study’s strength lies in the fact that our findings have pinpointed the most critical contributing factors for stunting, which will contribute to the existing literature regarding the incidence of stunting and its determinants in children under-5.

Conclusion

The study intended to investigate the prevalence of stunting and its determinants in children under-5 in the southwestern coastal region of Bangladesh. Findings depicted that the mother’s age and occupation, history of under-5 child mortality and age of the children were the significant determinants of stunting among children under 5. Moreover, the study recommends that government and non-government organisations should proactively foster programmes aimed at training young mothers on effective child-rearing practices, which can help to reduce stunting among under-5 children. Furthermore, creating more income-generating opportunities for women in rural Bangladesh which can empower their families and contribute to the overall well-being of their children. Implementing local awareness programmes that inform parents about appropriate growth standards for their children’s age and sex can be instrumental in promoting healthy development. Additionally, expanding facility-based healthcare centres in rural areas will ensure better access to quality healthcare for both mothers and children, paving the way for healthier communities.

Data availability statement

Data are available in a public, open access repository.