Introduction
Childhood nutrition and food security play pivotal roles in child health as they directly impact the physical and cognitive development of children.(1,Reference Pasdar, Nachvak, Darbandi, Morvaridzadeh, Rezaeian and Daneshi Maskooni2) Malnutrition and food insecurity are persistent challenges, particularly in low- and middle-income countries, leading to detrimental effects on health, education, and productivity.(1) Despite efforts to address these issues, a comprehensive understanding of the determinants of food security and nutritional status among children in resource-limited settings remains elusive. Childhood undernutrition develops as a result of inadequate dietary intake, diseases, and inappropriate care practices, leading to serious negative effects on child growth and development.(Reference Ghattas3) The consequences of undernutrition in children are both short-term (mortality and morbidity) and long-term, including small adult size, low intellectual ability, reduced economic productivity and reproductive performance, and metabolic and cardiovascular diseases.(Reference Semazzi and Kakungulu4,Reference Fall and Kumaran5)
Household food security is one of the underlying determinants of the nutritional status of children under five.(Reference Semazzi and Kakungulu4) In Uganda, despite several interventions to improve household food insecurity, such as increased agricultural extension services to fight undernutrition, high levels of malnutrition still persist, as evidenced by statistics from the country’s most recent national demographic health survey where food insecurity affects 16.4 million people, with 24.4% of children under five stunted and 53% anaemic.(6,Reference Mpuuga7) Similar to the rest of the country, the Busoga sub-region still experiences high levels of household food insecurity with approximately 33.3% of the population reported to be food insecure.(8) Despite the link between food security and nutritional status, established factors associated with household food security in the subregion remain scarce.
Children below 23 months of age are most likely to be still breastfeeding and cared for by their mothers or caregivers and may not suffer from the negative impacts of food insecurity. However, as they age, they are given less care and may not be given sufficient food, yet family meals are their sole source of nutrients. Therefore, this study sought to assess the prevalence of food insecurity and malnutrition among households with children 24–59 months in Namutumba District in the Busoga subregion. The factors that influence the food security situation of households and the nutritional status of children were also examined to establish the link between household food security and the nutritional status of children aged 24–29 months in the district.
Methods and tools
This cross-sectional study was conducted in Namutumba, a rural district in the Busoga subregion of Uganda. The district is composed of nine sub-counties and one town council as administrative units, with a total population of 252,557.(9) The main economic activity of the populace is subsistence farming with crops such as rice, groundnuts, maize, millet, and cassava grown.
Sample size determination and selection
The population of interest included children aged 24–59 months together with their primary caretakers. A sample of 308 child-caregiver pairs was determined using the Leslie Kish formula (1964)(8,Reference Kish10) based on a 27.7% stunting rate among children aged under five years in Namutumba District.(Reference Kuziga, Adoke and Wanyenze11) The child-caregiver pairs were ascertained at the household level following multi-stage probability sampling. Namutumba District has nine sub-counties/town councils, 36 parishes, and 200 villages. The sub-counties were the strata from which two parishes per sub-county were selected by simple random sampling, thereby including a total of 18 parishes. At the parish level, two out of 4 villages per parish were also selected by simple random sampling, thereby including a total of 36 villages. An estimated 8.5 households per village to give 308 households was rounded off to nine households per village to give a final total of 324 households.
In each selected village, a list of households with children aged 24–59 months (sampling frame) was obtained from the respective Village Health Teams. Nine households were systematically selected using a sampling interval determined by dividing the total number of eligible households in each village by nine. One index child was selected from each household, along with their primary caretaker. In situations where a household had more than one eligible child, one index child was selected using simple random sampling (See Figure 1).
Figure 1. Summary of the multi-stage probability sampling framework followed to ascertain the study respondents.
Data collection methods
An administered structured questionnaire was used for face-to-face interviews with the children’s primary caretakers to collect data on household socio-demographic characteristics, including household size, age and gender of the household head (HHH), that of the child and their primary caregiver as well as household income and assets, source of food, access to water, land, and food stocks.
Household food security was assessed using the Household Food Insecurity Access Scale (HFIAS) questionnaire.(Reference Coates, Swindale and Bilinsky12) The HFIAS includes nine recall questions on the occurrence of a food insecurity condition experienced in the household and the frequency of occurrence of such a condition within the past 30 days.(Reference Coates, Swindale and Bilinsky12) The respondents were first asked if a given food insecurity condition occurred or was experienced, and if so, at what frequency. The occurrence questions were subdivided into three themes of food insecurity: experiencing anxiety and uncertainty about household food supply, insufficient quality of the diet in terms of variety and food preferences, and insufficient food intake together with its physical consequences. The questions on the frequency of occurrence assessed the severity of food insecurity based on the regularity of food insecurity conditions experienced.(Reference Coates, Swindale and Bilinsky12) The resulting responses were transformed into continuous (HFIAS) scores ranging from 0–27 to indicate the degree of food insecurity (access) within a household. Households were then categorized as food secure, mildly, moderately, or severely food insecure based on their experience of occurrence and frequency of food insecurity using the Household Food Insecurity Access Prevalence (HFIAP) indicator.
Anthropometric measurements (height and weight) were used to determine the children’s nutritional status. The measurements were expressed as Height for Age, Weight for Age, and Weight for Height Z-scores and compared to the WHO reference standards.(13,14)
Data analysis
To assess factors associated with household food insecurity and the nutritional status of children, crude and adjusted prevalence rate ratios (PRR) were computed in Stata 16.1 through multi-level mixed effects generalized linear models using family (Poisson) and link (log) at the 95% confidence interval. The multilevel model was chosen because of the presence of clustering at the sub-county level (intra-class correlation coefficient = 0.104), with likely significant variations in food security across different sub-counties in the district, but not within sub-counties. All covariates with p < 0.2 at the bivariate level were considered for the multivariable model.
Simple and multiple linear regressions (adjusted for sex and age) were used to evaluate the association between household food security and child nutritional status. Statistical significance was set at p < 0.05.
Ethical considerations
Ethical clearance was sought from the Mbale Regional Referral Hospital Research and Ethics Committee Ref. No.UG-REC-011 and registered with the Uganda National Council for Science and Technology on permit No. HS2088ES. Informed consent was obtained from the respondents prior to participation in the study.
Results and Discussion
Household and socio-demographic characteristics of respondents
A total of 299 out of the 324 child-caregiver pairs met the eligibility criteria and consented to participate in the study. As indicated in Table 1, the mean age of the household head was 40.5 years (SD: ±11.1), with the youngest being 18 years and the oldest 75 years. Most household heads (91.6%) were male married (89.3%) and had attained primary-level education (71.2%). Conversely, almost all child caregivers were female (99.0%), with a mean age of 32.4 years (SD± 9.6) and a range of 15–65 years. Similar to the HH heads, the majority (78.3%) of child caregivers attained primary-level education. The average household number was 6.07 members (SD ± 2.31) and ranged from 2-15 HH members. The majority (94.31%) of households reported having access to some form of agricultural land. In terms of food stocks, only 36.8% had sufficient food reserves to last them at least a month.
Table 1. Household and demographic characteristics
The average age of the children was 40.9 months (SD: 10.8; Range 24–59 months) with a slight female predominance (53.85%). The majority (84.23%) of the children were immunized; however, only 44.97% had their immunization status verified using immunization cards. Approximately three-quarters (75.59%) of the children had at least one illness two weeks before the study interview.
Household food security status
The mean HFIAS score for all households was 12.79 (SD: 5.47; Range 0–26). Categorized according to the frequency of occurrence of food insecurity conditions, results demonstrated that most households 60.2% were severely food insecure, with 1% of the households being food secure (see Table 2).
Table 2. Household Food Security Status (HFIAS)
Household food security according to the HFIAS 3 domains
The study findings indicated that approximately 90% of the households experienced anxiety and uncertainty concerning their household food supply 30 days before this interview. In addition, the highest percentage (96%) of HH members had insufficient food quality and consumed foods that lacked variety and were not of their desired choice. On the same note, 92% of the households faced a shortage of food, inadequate intake, and dealing with the associated physical consequences (see Figure 2).
Figure 2. Percentage of Households that reported experiencing either of the Household food insecurity access domains in the 30 days preceding the study including: feelings of anxiety and uncertainty concerning their household food supply, insufficient food quality and consumption of un-preferred foods which lacked variety and were not of their desired choice in addition to shortage of food hence inadequate intake.
Factors associated with severe food insecurity
Bivariable and multivariable models were used to identify the key factors associated with severe food insecurity.
In the bivariate analysis, results showed that households where the head possessed a high school qualification (Crude PRR: 0.73; 95% CI: 0.57–0.93; p = 0.011) and those within the medium (Crude PRR: 0.78; 95% CI: 0.65–0.95; p = 0.014) and richest (Crude PRR: 0.63; 95% CI: 0.47–0.83; p = 0.001) wealth groups experienced less severe food insecurity compared to those whose heads possessed less than a high school qualification or were within the poorest wealth group. Similarly, households with at least one member earning an income had reduced severe food insecurity (Crude PRR: 0.82; 95% CI: 0.68–0.99; p = 0.045) compared to those with no income earners.
On the contrary, households without access to agricultural land (Crude PRR: 1.50; 95% C.I: 1.23, 1.84; p < 0.001), those without access to food stocks (Crude PRR: 1.69; 95% C.I: 1.33, 2.15; p < 0.001), and those with treated drinking water (Crude PRR: 1.38; 95% C.I: 1.16, 1.65; p < 0.001) were more severely food insecure, as indicated in Table 3.
Table 3. Factors associated with severe food insecurity among households of children 24–59 months
HHH = household head; HH = household; * Significant association; #One household had missing data (n = 298).
All covariates with a p < 0.2 at the bivariate level were considered for multivariable analysis. The results demonstrated a positive association between household size and food insecurity, that is, for every additional household member, the prevalence of severe food insecurity increased by 4% (Adjusted PRR: 1.04; 95% C.I: 1.01, 1.08; p = 0.005). There was also a significant positive association between severe food insecurity and lack of access to agricultural land (Adjusted PRR: 1.26; 95% C.I: 1.07, 1.49; p = 0.005). Similarly, 52% of households that lacked food stocks were also food insecure compared to those that had stocks (Adjusted PRR: 1.52; 95% C.I: 1.20, 1.92; <0.001). Conversely, households within the highest wealth category had a 29% lower prevalence of severe food insecurity than those in the lowest wealth category (Adjusted PRR: 0.71; 95% C.I: 0.52, 0.96, p = 0.027). No caregiver factor was associated with household food insecurity.
Child nutritional status
Overall, the prevalence of wasting, underweight, and stunting was 3.1%, 9.7%, and 28.0%, respectively, while 2.8% of the children were overweight/obese (Table 4).
Table 4. Nutritional status of children 24–59 months in Namutumba District
Z score cut offs: Severe < −3; Moderate < −2 and ≥ −3; Moderate & Severe <-2; Mild < −1 and ≥ −2; Normal < 2 and ≥ −1; Overweight < 3 and ≥ 2; Obese: ≥ 3; †Some children (n = 10) excluded due to flagging; #One child had missing data.
Stratified by sex, boys showed a higher prevalence for all forms of malnutrition (wasting: 3.76% vs. 2.56%, underweight: 10.53% vs. 8.97%, and stunting: 31.58% vs. 25.00%) than girls (Table 5).
Table 5. Nutritional status of children stratified by sex
Z score cut-offs: Severe < −3; Moderate < −2 and ≥ −3; Moderate & Severe <-2; Mild < −1 and ≥ −2; Normal < 2 and ≥ −1; Overweight < 3 and ≥ 2; Obese: ≥ 3; †Some children (n = 10) excluded due to flagging; #One child had missing data.
By age group, wasting was found to be most prevalent among the youngest (24–35 months) and oldest children (48–59 months) children at 3.2% for both age groups. The oldest children (48–59 months) also had the highest prevalence of underweight (14.5%), while stunting was the highest among the youngest children (24–35 months) at 34.4% (see Figure 3).
Figure 3. Prevalence of undernutrition among children aged 24–59 months in Namutumba District disaggregated by age group.
Factors influencing nutritional status
Bivariate analysis of factors associated with stunting did not reveal any household, caregiver, or child factors to be associated with stunting. The two variables (i.e. ‘Number of household members who earn income’ and ‘child morbidity’) that met the p < 0.2 cut-offs for multivariable analysis were modelled but were also not found to be significantly associated with stunting. Similarly, no child factors, including sex, age, immunization status, and child morbidity in the previous two weeks, were associated with child malnutrition at both the bivariate and multivariate levels (p > 0.05).
Household food security and nutritional status
A simple linear regression of the association between household food security (HFIAS scores) and the three nutritional status indicators showed a significant positive association between HFIAS scores with weight for height (WHZ) scores (β = 0.75, p = 0.011), while height for age (HAZ) scores were found to be negatively associated with HFIAS scores (β = -0.38, p = 0.035) indicating that an increase in food insecurity (high HFIAS scores) resulted in high levels of stunting (HAZ scores). In the multiple linear regression models after adjusting for child sex and age, the results still demonstrated a positive association between WHZ scores (β = 0.81, p = 0.007) and HFIAS scores, whereas HAZ scores were negatively associated with HFIAS scores (β = −0.37, p = 0.039). The WAZ scores were not significantly associated with the HFIAS scores (Table 6).
Table 6. Association between household food security and nutritional status
Discussion
Household food security and associated factors
The findings of this study indicate that almost all households (99%) faced some form of food insecurity comparable to 75.8% status found by a study conducted in Ethiopia,(Reference Betebo, Ejajo, Alemseged and Massa15) with a high prevalence of severe food insecurity (60.2%) in the district compared to the Busoga regional prevalence of 56% and the national average of 11%.(16) Most households (>90%) experienced the three domains of food insecurity: anxiety and uncertainty about the next food source, poor quality and variety of food, and insufficient intake of food. This predisposes HH members, especially young children, to the risk of malnutrition and irreversible poor nutritional outcomes limiting them from reaching their full potential in adulthood.(1,Reference Ghattas3)
The findings at the bivariate level i.e. secondary level education (p = 0.011), medium level (p = 0.014), and highest-level wealth percentile (p < 0.001), and at least one household member earning an income (p = 0.045) were positively associated with improved household food security, while lack of access to agricultural land, lack of food stocks, and households with treated drinking water were positively associated with severe food insecurity (p < 0.001), are similar to those of previous comparable studies including a population-based cross-sectional study of food insecurity and the influential factors in households in Kermanshah, Iran that found 69.5% of the households had food insecurity with significant correlations observed between food insecurity and family size, occupation status of the household head, number of rooms, monthly income, and education level (P = 0.001).(Reference Pasdar, Nachvak, Darbandi, Morvaridzadeh, Rezaeian and Daneshi Maskooni2) Another study conducted in Laos on food security revealed that household size, food price, drought, shock, household income per month, number of labourers, gender of the household head, and farmland areas are important factors for household food insecurity.(Reference Phami, He, Liu, Ding, Silva and Li17) Similarly, a study in the eastern region of Nepal on the determinants of food security emphasized access to land as one of the key necessities to improving food security status.(Reference Joshi and Joshi18) However, the association between treated water and household food insecurity was surprising. We assumed that water treatment could be a time-consuming and expensive activity that leaves less time for household members to grow or look for food and less disposable income to buy food.
At the multivariate level, all factors linked to severe food insecurity at the bivariate level held true, except for treated drinking water (Adjusted PRR: 1.29; 95% C.I: 0.98, 1.71; p = 0.068). Additionally, household size was positively associated with severe food insecurity. For every additional household member, the prevalence of severe food insecurity increased by 4% (Adjusted PRR: 1.04; 95% C.I: 1.01, 1.08; p = 0.005). Considering that the average age of the household head was 40.5 years implies that these were middle-aged adults(Reference Lachman, Smelser and Baltes19,Reference Klimczuk20) who may still have school-going children who would not contribute to family labour to grow food. Hence, the larger the household, the more mouths to feed yet with less labour for production.
Namutumba is a rural district, where land is a major resource for household food production; therefore, access to an adequate size of land increases the likelihood of improved food stocks; hence, households become more food secure(Reference Joshi and Joshi18) while the reverse increases the likelihood of food insecurity. However, in our study, although more than 90% of the households had access to agricultural land, this did not translate into ample food stock. Studies have shown that a larger portion of agricultural land in Busoga sub-region is usually allocated to sugarcane growing for economic purposes hence affecting food production and household food security.(Reference Kasango21,Reference Mwavu, Kalema, Bateganya, Byakagaba, Waiswa and Enuru22) This could have been the case in our study, therefore the agricultural outputs were not sufficient for households to have a surplus for stocks. Additionally, the surplus food stocks could have been sold for economic gain. In either case, the nutritional status of the household members was greatly compromised.
A plethora of studies,(Reference Botreau and Cohen23–Reference Botreau and Cohen28) have shown that factors such as age, gender, and marital status of household heads are known to influence household food security; however, the findings of our study revealed that these factors did not significantly impact household food security in Namutumba District.
Child nutritional status
The average age of the children (40.9 months) means that most of them were older and most likely to experience reduced care from their mothers/caretakers, especially concerning feeding recommendations such as being fed less nutritious food, lack of variety, not being given special diets, and eating together with other family members.(Reference Huong, Xuan, Phuong, Huyen and Rocklöv29) This could most likely explain the high prevalence of wasting (14.5%) among children 48–59 months. On the other hand, the youngest children 24–35 months experienced the highest prevalence of stunting. At this age, they are heavily reliant on their caregivers for feeding and other forms of care, which may not be provided adequately. As they grow older, they are more likely to feed themselves and care for themselves, without much reliance on their caregivers.
It is important to note that most children (87.3%) reported at least one illness in the two weeks before the study. Coupled with limited evidence of immunization (45%), such children are highly susceptible to infections(Reference Farhadi and Ovchinnikov30) which could result into inadequate intake and utilization of energy and nutrients(Reference Humphries, Scott and Vermund31) thereby contributing to poor nutritional status.
Although poor feeding and other child care practices are risk factors for malnutrition for both boys and girls, studies(Reference Thurstans, Opondo, Seal, Wells, Khara and Dolan32,Reference Samuel, Osendarp, Feskens, Lelisa, Adish and Kebede33) have shown that boys are more likely to be malnourished compared to girls. The underlying biological mechanisms for this difference remain poorly understood.(Reference Wamani, Åstrøm, Peterson, Tumwine and Tylleskär34) Many have hypothesized that boys are more susceptible to infections and are more fragile in the first year of life.(Reference Kraemer35) The findings of this study were not any different and showed that boys were at a greater risk than girls for all forms of malnutrition.
The lack of significant association between household, caregiver, and child characteristics and stunting indicates that child malnutrition may be attributed to other factors such as poor dietary practices and lack of access to quality health services. Other factors include long-term shocks and humanitarian crises.(36)
However, the study revealed that HFIAS scores were positively associated with WHZ scores, implying that the more food insecure a household was, the lower the prevalence of wasting. This is consistent with a report by(13) which noted that food-insecure households mostly have scarce resources for food, often resorting to cheap, less nutritious, and high-calorie foods, which increases the risk of becoming overweight and obese. In addition, according to the Uganda IMAM guideline (2020),(37) children wasting from food insecure households when identified and enrolled into therapeutic feeding programmes tend to cure and their normal WHZ score is restored, yet their households may still remain food insecure.(37) Conversely, a negative association was established between HFIAS scores and HAZ scores, implying that prolonged household food insecurity resulted into a higher prevalence of stunting among children, and the reverse was true. This also resonates with the World Health Organization’s state of food security and nutrition report(1) which indicated that household food security influences the dietary intake of children in terms of quality and quantity of diets thus known to have a strong influence on the nutritional status of children.
Conclusion
Food insecurity in Namutumba District was highly influenced by several factors such as household size, lack of access to agricultural land, lack of household food stocks, and low wealth index, leading to all forms of undernutrition (stunting, wasting, and underweight), with stunting dominating at 28%. The positive relationship between household food insecurity and wasting among children and the negative association between household food insecurity and stunting underscores the importance of household food security to the nutritional well-being of children.
However, the lack of association between household food insecurity and being underweight implies that food security is not the only predictor of a child’s nutritional status. Other factors that may influence the nutritional status of children should be studied.
Finally, the high food insecurity levels in Namutumba District not only predispose children but also the entire population to malnutrition and associated social and economic outcomes, such as poor cognitive development and poverty, leading to a vicious cycle. Therefore, a holistic food system approach to improving household food security should be used to address population malnutrition in the district.
Acknowledgements
The authors thank all caregivers and children who participated in this study. Additionally, they are grateful to the district local government officials for guiding them in the study area during the data-collection process.
Authorship Contributions
JKA, KN, and JN conceptualized the study design. KN collected and analyzed the data, JKA wrote the first draft of the manuscript; JKA, JN, and VN revised and improved the final manuscript.
Financial support
This study did not receive any specific grant from funding agencies in the public, commercial, or non-profit sectors.
Competing interests
No Conflict of interest to declare.
Open access
