Factors associated with stunting among children according to the level of food insecurity in the household: A cross-sectional study in a rural community of Southeastern Kenya Global health

BMC Public Health, Volume 15, No. 1, Year 2015

Interpretation

Background: Chronic malnutrition or stunting among children under 5 years old is affected by several household environmental factors, such as food insecurity, disease burden, and poverty. However, not all children experience stunting even in food insecure conditions. To seek a solution at the local level for preventing stunting, a cross-sectional study was conducted in southeastern Kenya, an area with a high level of food insecurity. Methods: The study was based on a cohort organized to monitor the anthropometric status of children. A structured questionnaire collected information on the following: demographic characteristics, household food security based on the Household Food Insecurity Access Scale (HFIAS), household socioeconomic status (SES), and child health status. The associations between stunting and potential predictors were examined by bivariate and multivariate stepwise logistic regression analyses. Furthermore, analyses stratified by level of food security were conducted to specify factors associated with child stunting in different food insecure groups. Results: Among 404 children, the prevalence of stunting was 23.3%. The percentage of households with severe food insecurity was 62.5%. In multivariative analysis, there was no statistically significant association with child stunting. However, further analyses conducted separately according to level of food security showed the following significant associations: in the severely food insecure households, feeding tea/porridge with milk (adjusted Odds Ratio [aOR]: 3.22; 95% Confidence Interval [95% CI]: 1.43-7.25); age 2 to 3 years compared with 0 to 5 months old (aOR: 4.04; 95% CI: 1.01-16.14); in households without severe food insecurity, animal rearing (aOR: 3.24; 95% CI: 1.04-10.07); SES with lowest status as reference (aOR range: from 0.13 to 0.22). The number of siblings younger than school age was not significantly associated, but was marginally associated in the latter household group (aOR: 2.81; 95% CI: 0.92-8.58). Conclusions: Our results suggest that measures against childhood stunting should be optimized according to food security level observed in each community. A cross-sectional study was conducted in Kwale District in the Coast Province of Kenya in 2012, using a cohort nested to the Health and Demographic Surveillance System (HDSS) program, which follows about 50,000 residents periodically, in collaboration with Nagasaki University and the Kenya Medical Research Institute [20]. In this cohort, we recruited children under 5 years old and their caregivers, including non-biological mothers, from households located within a radius of 2.2 km from the Kizibe Health Center, one of the health centers in the HDSS program area. The radius was set in consideration of accessibility for children and their caregivers to the surveys in the nested-cohort study. We took into consideration the estimated sampling size (438 children) for a 2 sample comparison of proportions calculated in the study design stage assuming that 10% of children would become stunted during the observation period and there would be twice as many children with stunting in the comparison group, which has a factor (exposure) with a power of 80% and a significance level of 5% (2-tailed). This cohort program measured several indices, including anthropometric measurements such as height and weight, and asked questions of mothers related to health status and dietary intake. The measurements were to take place 3 times per year between 2011 and 2014. In this cross-sectional study, a structured questionnaire was additionally administered as part of the follow-up surveys of the cohort to investigate the relationship between intra-household environment and child nutritional status. During the survey period in 2012, 653 households were registered within a 2.2-km radius from the health center of the HDSS program; and among them, 516 children less than 5 years old were identified in 360 households. After carrying out a pre-test to revise the questionnaire for suitability, we conducted interviews of the caregivers by trained local investigators in the Kiswahili language at the health center. The interview required approximately 20 minutes to complete. The structured questionnaire consisted of the following variables: demographic characteristics; socioeconomic status; household food security; child health status, such as breastfeeding behavior and illness in the past 2 weeks including jigger flea (Tunga penetrans) infection; caregiver’s perception of child’s growth; and caregiver’s household chores as a proximal factor of availability for child rearing. The household food security level was measured using the Household Food Insecurity Access Scale (HFIAS) with scores ranging from 0 to 27 by household level [21]. The HFIAS scores obtained from households were categorized into 4 levels of food insecurity, namely, “food secure,” “mildly food insecure,” “moderately food insecure,” and “severely food insecure,” based on the HFIAS guideline [22]. The household socioeconomic status (SES) was parameterized by the principle component analysis (PCA) method using house properties confirmed by the questionnaire: property owned; source of drinking water; type of toilet facility; and type of flooring, wall material, and roof material. The items of household property were selected according to the Demographic Health Survey (DHS) [19]. The score in the first PCA component was used as an asset index of SES status for each household [23]. According to the PCA-based asset index, households were divided into 4 groups; the first quartile SES group was poorest and the fourth quartile SES group was richest in the study area. For data validation, Cronbach’s alpha coefficient, which is a measure of the internal consistency of a scale, was used to confirm the reliability of the HFIAS and household SES measure. An alpha value of more than 0.7 indicates that the measure is acceptable. Child age was confirmed using his/her maternal and child health (MCH) handbook or by the response from the caregiver if the MCH handbook was not available. Anthropometric measurement data were obtained from the child cohort dataset. In the child cohort study, height was measured by a length scale (Seca GmbH & Co.Kg, Germany). Weight was measured using trouser for baby weighing scale (G.S.T. Corporation, India) and portable electronic scale (Guangzhou Weiheng Electronics Co., Ltd, China) for babies; and KRUPS Baby Cum Child Weighing Scale (Doctor Beci Ram & Sons [MFG.], India) for children who could stand. For measuring the weight of caregivers a Tanita THD-650 scale (Tanita, Japan) was used. Chronic malnutrition (stunting) of children was defined as z-score below 2 standard deviations(SD) from the mean for length or height for age according to the Child Growth Standards published by the WHO in 2006 [6]. For this study, those who had a z-score above −2 SD were defined as children who did not have stunting. We excluded the following children from the analysis: those whose caregivers were unable to answer the questions due to hearing disability; those who were severely sick; and those whose birth date were not appropriate or unclear. Because 72.3% of children in this study were born at home according to our survey data, some birth dates were not clearly recorded. The association between potential predictors (child and caregiver characteristics, intra-household environment, food intake, and health history) and stunting status was determined by univariate logistic regression analyses. Because some children belong to the same household and may be correlated, cluster options by household were incorporated in the logistic regression. Multiple logistic regression analysis was also conducted to control confounding factors by backward stepwise selection with 0.2 of significant level of removal from the model as well as cluster option by household. Additionally, to identify associated factors of childhood stunting separately in severe and non-severe food insecurity groups, the analyses were independently conducted for the 2 groups in the same statistical manner. Stata statistical software (version 12.0: Stata Corporation, TX, USA) was used for data cleansing and data analyses. This study was approved by the Ethics Committee of Nagasaki University and authorized as a sub-study of the cohort study by the Ethical Review Committee of the Kenya Medical Research Institute (KEMRI SSC No.1964). Study permission was also obtained from the National Council for Science and Technology (NCST) in Kenya (Research Permit No. NCST/RCD/12A/012/59). We explained the study objectives and obtained written informed consent from all participants before collecting data. Participants were informed that participation in this study was voluntary and that they could stop participating at any time without experiencing negative consequences.

AI Digest

Study Justification:

This study aimed to investigate the factors associated with stunting among children under 5 years old in a rural community in Southeastern Kenya. The study was conducted in an area with a high level of food insecurity, as chronic malnutrition or stunting is influenced by household environmental factors such as food insecurity, disease burden, and poverty. The study aimed to identify specific factors associated with stunting in different levels of food insecurity in order to develop targeted interventions to prevent stunting at the local level.

Highlights:

– The prevalence of stunting among the children in the study was 23.3%.
– The percentage of households with severe food insecurity was 62.5%.
– The study found significant associations between stunting and certain factors in different levels of food insecurity:
– In severely food insecure households, feeding tea/porridge with milk, age 2 to 3 years compared to 0 to 5 months old were associated with stunting.
– In households without severe food insecurity, animal rearing and lower socioeconomic status were associated with stunting.
– The number of siblings younger than school age was marginally associated with stunting in households without severe food insecurity.

Recommendations:

Based on the study findings, the following recommendations can be made:

– Optimize measures against childhood stunting according to the observed food security level in each community.
– Provide support and education to severely food insecure households on appropriate feeding practices for children, such as avoiding tea/porridge with milk and introducing nutritious foods.
– Promote animal rearing and improve socioeconomic status in households without severe food insecurity to reduce the risk of stunting.
– Consider the potential impact of the number of siblings younger than school age on stunting, especially in households without severe food insecurity.

Key Role Players:

To address the recommendations, the following key role players are needed:

– Local community leaders and organizations: They can help implement and promote interventions to address childhood stunting, especially in severely food insecure households.
– Health professionals and nutritionists: They can provide guidance and support on appropriate feeding practices and nutrition education.
– Government agencies: They can allocate resources and develop policies to improve socioeconomic conditions and food security in the community.

Cost Items:

While the actual cost of implementing the recommendations will vary depending on the specific interventions and context, the following cost items should be considered in planning:

– Training and capacity building for local community leaders, health professionals, and nutritionists.
– Development and dissemination of educational materials on appropriate feeding practices and nutrition.
– Support for animal rearing programs and initiatives to improve socioeconomic conditions.
– Monitoring and evaluation of the interventions to assess their effectiveness and make necessary adjustments.

Strength of Evidence

The strength of evidence for this abstract is 7 out of 10.
The evidence in the abstract is based on a cross-sectional study conducted in a specific rural community in Southeastern Kenya. The study collected information on demographic characteristics, household food security, socioeconomic status, and child health status. The associations between stunting and potential predictors were examined using bivariate and multivariate logistic regression analyses. The study found that there was no statistically significant association between child stunting and the predictors in the multivariate analysis. However, further analyses conducted separately according to the level of food security showed significant associations in certain groups. The evidence is based on a specific population and may not be generalizable to other settings. To improve the evidence, future studies could consider conducting a longitudinal study to establish causal relationships and include a larger and more diverse sample to enhance generalizability.

Abstract

A cross-sectional study was conducted in Kwale District in the Coast Province of Kenya in 2012, using a cohort nested to the Health and Demographic Surveillance System (HDSS) program, which follows about 50,000 residents periodically, in collaboration with Nagasaki University and the Kenya Medical Research Institute [20]. In this cohort, we recruited children under 5 years old and their caregivers, including non-biological mothers, from households located within a radius of 2.2 km from the Kizibe Health Center, one of the health centers in the HDSS program area. The radius was set in consideration of accessibility for children and their caregivers to the surveys in the nested-cohort study. We took into consideration the estimated sampling size (438 children) for a 2 sample comparison of proportions calculated in the study design stage assuming that 10% of children would become stunted during the observation period and there would be twice as many children with stunting in the comparison group, which has a factor (exposure) with a power of 80% and a significance level of 5% (2-tailed). This cohort program measured several indices, including anthropometric measurements such as height and weight, and asked questions of mothers related to health status and dietary intake. The measurements were to take place 3 times per year between 2011 and 2014. In this cross-sectional study, a structured questionnaire was additionally administered as part of the follow-up surveys of the cohort to investigate the relationship between intra-household environment and child nutritional status. During the survey period in 2012, 653 households were registered within a 2.2-km radius from the health center of the HDSS program; and among them, 516 children less than 5 years old were identified in 360 households. After carrying out a pre-test to revise the questionnaire for suitability, we conducted interviews of the caregivers by trained local investigators in the Kiswahili language at the health center. The interview required approximately 20 minutes to complete. The structured questionnaire consisted of the following variables: demographic characteristics; socioeconomic status; household food security; child health status, such as breastfeeding behavior and illness in the past 2 weeks including jigger flea (Tunga penetrans) infection; caregiver’s perception of child’s growth; and caregiver’s household chores as a proximal factor of availability for child rearing. The household food security level was measured using the Household Food Insecurity Access Scale (HFIAS) with scores ranging from 0 to 27 by household level [21]. The HFIAS scores obtained from households were categorized into 4 levels of food insecurity, namely, “food secure,” “mildly food insecure,” “moderately food insecure,” and “severely food insecure,” based on the HFIAS guideline [22]. The household socioeconomic status (SES) was parameterized by the principle component analysis (PCA) method using house properties confirmed by the questionnaire: property owned; source of drinking water; type of toilet facility; and type of flooring, wall material, and roof material. The items of household property were selected according to the Demographic Health Survey (DHS) [19]. The score in the first PCA component was used as an asset index of SES status for each household [23]. According to the PCA-based asset index, households were divided into 4 groups; the first quartile SES group was poorest and the fourth quartile SES group was richest in the study area. For data validation, Cronbach’s alpha coefficient, which is a measure of the internal consistency of a scale, was used to confirm the reliability of the HFIAS and household SES measure. An alpha value of more than 0.7 indicates that the measure is acceptable. Child age was confirmed using his/her maternal and child health (MCH) handbook or by the response from the caregiver if the MCH handbook was not available. Anthropometric measurement data were obtained from the child cohort dataset. In the child cohort study, height was measured by a length scale (Seca GmbH & Co.Kg, Germany). Weight was measured using trouser for baby weighing scale (G.S.T. Corporation, India) and portable electronic scale (Guangzhou Weiheng Electronics Co., Ltd, China) for babies; and KRUPS Baby Cum Child Weighing Scale (Doctor Beci Ram & Sons [MFG.], India) for children who could stand. For measuring the weight of caregivers a Tanita THD-650 scale (Tanita, Japan) was used. Chronic malnutrition (stunting) of children was defined as z-score below 2 standard deviations(SD) from the mean for length or height for age according to the Child Growth Standards published by the WHO in 2006 [6]. For this study, those who had a z-score above −2 SD were defined as children who did not have stunting. We excluded the following children from the analysis: those whose caregivers were unable to answer the questions due to hearing disability; those who were severely sick; and those whose birth date were not appropriate or unclear. Because 72.3% of children in this study were born at home according to our survey data, some birth dates were not clearly recorded. The association between potential predictors (child and caregiver characteristics, intra-household environment, food intake, and health history) and stunting status was determined by univariate logistic regression analyses. Because some children belong to the same household and may be correlated, cluster options by household were incorporated in the logistic regression. Multiple logistic regression analysis was also conducted to control confounding factors by backward stepwise selection with 0.2 of significant level of removal from the model as well as cluster option by household. Additionally, to identify associated factors of childhood stunting separately in severe and non-severe food insecurity groups, the analyses were independently conducted for the 2 groups in the same statistical manner. Stata statistical software (version 12.0: Stata Corporation, TX, USA) was used for data cleansing and data analyses. This study was approved by the Ethics Committee of Nagasaki University and authorized as a sub-study of the cohort study by the Ethical Review Committee of the Kenya Medical Research Institute (KEMRI SSC No.1964). Study permission was also obtained from the National Council for Science and Technology (NCST) in Kenya (Research Permit No. NCST/RCD/12A/012/59). We explained the study objectives and obtained written informed consent from all participants before collecting data. Participants were informed that participation in this study was voluntary and that they could stop participating at any time without experiencing negative consequences.

Materials

N/A

Innovations

Based on the provided information, here are some potential innovations that could be used to improve access to maternal health:

1. Mobile Health (mHealth) Applications: Develop and implement mobile applications that provide pregnant women and new mothers with access to important health information, reminders for prenatal and postnatal care appointments, and educational resources.

2. Telemedicine Services: Establish telemedicine services that allow pregnant women in remote or underserved areas to consult with healthcare professionals via video calls, reducing the need for travel and improving access to prenatal care.

3. Community Health Workers: Train and deploy community health workers who can provide basic prenatal and postnatal care, conduct health education sessions, and refer women to healthcare facilities when necessary.

4. Maternal Health Vouchers: Introduce voucher programs that provide pregnant women with financial assistance to cover the costs of prenatal care, delivery, and postnatal care, ensuring that cost is not a barrier to accessing essential maternal health services.

5. Transportation Support: Develop transportation initiatives that provide pregnant women with affordable and reliable transportation to healthcare facilities, particularly in rural areas where access to transportation is limited.

6. Maternal Health Clinics: Establish dedicated maternal health clinics that offer comprehensive prenatal and postnatal care services, including regular check-ups, screenings, and counseling, to ensure that women receive specialized care throughout their pregnancy and after childbirth.

7. Health Education Programs: Implement community-based health education programs that focus on maternal health, covering topics such as nutrition, breastfeeding, hygiene practices, and the importance of prenatal and postnatal care.

8. Maternity Waiting Homes: Set up maternity waiting homes near healthcare facilities, providing a safe and comfortable place for pregnant women to stay as they approach their due date, ensuring they are close to medical care when labor begins.

9. Public-Private Partnerships: Foster collaborations between public and private sectors to improve access to maternal health services, leveraging the resources and expertise of both sectors to address gaps in healthcare delivery.

10. Data-driven Approaches: Utilize data analytics and digital health technologies to identify areas with high maternal health needs, monitor health outcomes, and target interventions to areas with the greatest need, ensuring resources are allocated effectively.

These innovations can help address barriers to accessing maternal health services, improve health outcomes for pregnant women, and reduce maternal and infant mortality rates.

AI Innovations Description

Based on the information provided, the recommendation to improve access to maternal health and prevent childhood stunting in food insecure communities is to optimize measures according to the level of food security observed in each community. This means tailoring interventions and strategies to address the specific needs and challenges faced by households with severe food insecurity and those without severe food insecurity.

For households with severe food insecurity, the following measures can be considered:

1. Promote the inclusion of nutrient-rich foods: Encourage the feeding of tea/porridge with milk, which was found to be associated with a lower risk of stunting. This can help provide essential nutrients to children in food insecure households.

2. Focus on age-specific interventions: Target children aged 2 to 3 years, as they were found to be at a higher risk of stunting compared to younger children. Implement age-appropriate interventions to address their nutritional needs and promote healthy growth.

For households without severe food insecurity, the following measures can be considered:

1. Promote animal rearing: Animal rearing was found to be associated with a lower risk of stunting in these households. Encourage households to engage in small-scale animal husbandry, such as raising chickens or goats, to improve access to animal-source foods and enhance dietary diversity.

2. Improve socioeconomic status: Socioeconomic status (SES) was found to be a significant factor in households without severe food insecurity. Implement interventions that aim to improve household income, education, and access to resources to uplift the socioeconomic status of these households.

3. Address the impact of household size: Although not statistically significant, the number of siblings younger than school age was marginally associated with stunting in households without severe food insecurity. Consider interventions that address the challenges faced by larger households, such as providing support for family planning and reproductive health services.

Overall, it is important to take into account the specific context and level of food security in each community when developing and implementing interventions to improve access to maternal health and prevent childhood stunting.

AI Innovations Methodology

The methodology used in the study described is a cross-sectional study design. Here is a brief summary of the methodology:

1. Study Location: The study was conducted in Kwale District, Coast Province, Kenya.

2. Study Population: The study included children under 5 years old and their caregivers from households located within a radius of 2.2 km from the Kizibe Health Center.

3. Sample Size: The estimated sample size was 438 children, based on a 2-sample comparison of proportions. A total of 516 children were identified in 360 households within the study area.

4. Data Collection: A structured questionnaire was administered to the caregivers of the children. The questionnaire collected information on demographic characteristics, socioeconomic status, household food security, child health status, caregiver’s perception of child’s growth, and caregiver’s household chores.

5. Measurements: Anthropometric measurements such as height and weight were taken from the children. Height was measured using a length scale, and weight was measured using different scales depending on the age of the child. Chronic malnutrition (stunting) was defined as a z-score below 2 standard deviations from the mean for length or height for age.

6. Data Analysis: Univariate logistic regression analyses were conducted to determine the association between potential predictors and stunting status. Multiple logistic regression analysis was also conducted to control for confounding factors. Cluster options by household were incorporated in the logistic regression to account for correlation between children within the same household.

7. Stratified Analysis: The analysis was stratified by the level of food security to identify associated factors of childhood stunting separately in severe and non-severe food insecurity groups.

8. Statistical Software: Stata statistical software (version 12.0) was used for data cleansing and analysis.

9. Ethical Considerations: The study was approved by the Ethics Committee of Nagasaki University and the Ethical Review Committee of the Kenya Medical Research Institute. Informed consent was obtained from all participants.

The study aimed to investigate the relationship between intra-household environment and child nutritional status, specifically focusing on factors associated with stunting among children in a rural community in Southeastern Kenya. The findings suggested that measures against childhood stunting should be optimized according to the level of food security observed in each community.

Author

Dima Health

Statistics:

Citations: 38

Identifiers:

DOI: 10.1186/s12889-015-1802-6

Research Areas:

Community Interventions, Disability, Disparities, Environmental, Food Security, Health System and Policy, Maternal Access, Maternal and Child Health, Quality of Care, Social Determinants, Technology and Innovations

Study Countries:

Kenya

Study Design:

Case-Control Study, Cohort Study, Cross Sectional Study, Quasi Experimental Study

Study Approach:

Quantitative