Predictors of Stunting and Underweight Among Children Aged 6 to 59 months in Bussi Islands, Wakiso District, Uganda: A Cross-Sectional Study

Background: Child undernutrition is a major public health concern in Uganda that can lead to increased risks of death with its prevalence higher in rural hard-to-reach areas than in urban areas. While it is assumed that the prevalence will be more concerning in islands with restricted accessibility to healthcare resources, limited research has been conducted on the prevalence and associated factors of stunting and underweight among children in Bussi Islands of Uganda. This study aimed to assess the prevalence and predictors of stunting and underweight among children aged 6 to 59 months in Bussi Islands of Wakiso District in Uganda. Methods: A cross-sectional study was conducted in Bussi Islands of Wakiso District. Sociodemographic and anthropometric measurements were obtained for randomly sampled 409 caretaker-child pairs from 409 households. Data was collected using pre-tested structured electronic questionnaires validated by the Uganda Ministry of Health. Anthropometric indices were calculated using ENA-SMART version 2011 and data analysis was conducted using STATA version 14. Modified Poisson regression was used to generate Unadjusted and Adjusted Prevalence Ratios (APRs) with 95% confidence intervals. Results: Prevalence of stunting and underweight among children in Bussi Islands were 29.8% and 16.1%, respectively. Independent predictors of stunting included: suffering from diarrhea (APR: 1.8; 95% CI: 1.3, 2.5); household food insecurity (APR: 1.7; 95% CI: 1.2, 2.4); and child age of 12 to 23 months and 24 to 35 months (APR: 2.3; 95% CI: 1.3, 4.0 and APR: 2.0; 95% CI: 1.1, 3.6 respectively). Protective factors against stunting were not suffering from measles (APR: 0.62; 95% CI: 0.42, 0.92); receiving deworming tablets every 6 months (APR: 0.58; 95% CI: 0.42, 0.81); and daily household utilization of more than 80 L of water (APR: 0.48; 95% CI: 0.24, 0.95). Predictors of underweight were suffering from diarrhea (APR: 2.2; 95% CI: 1.4, 3.4) and having more than 9 household members (APR: 2.8; 95% CI: 1.1, 7.5). Conclusions: Child stunting and underweight are prevalent public health problems in Bussi Islands of Wakiso District. Therefore, the study suggests that nutrition interventions in the Islands should focus on childhood vaccination, family planning, sufficient safe water coverage, household food security, and health education of child caretakers on optimal infant and young child feeding and development. A community-based cross-sectional study was conducted in April 2019 in Bussi Islands of Wakiso District in Uganda. Bussi Islands are a group of 5 Islands that make up Bussi Subcounty with an approximate population of 10 000 people. 22 They are surrounded by Lake Victoria and are located 41 km (24.6 mi) south of Kampala City.23,24 In this study, child caretakers were study respondents while children aged 6 to 59 months in Bussi Islands were the study population. Children who had physical body deformities that interfere with anthropometric assessment were excluded from the study. An adjusted sample size of 409 was estimated using the Kish Leslie formula 25 with a 95% confidence interval, expected stunting prevalence of 30% in rural Uganda, 6 5% maximum acceptable error, a design effect of 1.2, and a 5% potential non-response. A sampling frame was developed for 3 out of 5 islands, which were randomly selected. The number of households (HHs) selected per island was proportionate to the number of households with eligible children in each of the selected islands. Simple random sampling was used to select the households and a child in a household with more than one eligible child. Stunting was determined using height-for-age z-score (HAZ) and underweight was determined using weight-for-age z-score (WAZ). A child was classified as stunted when their HAZ was < −2 standard deviations (SD) and normal when their HAZ was ⩾ −2SD. A child was classified as underweight when their WAZ was < −2SD and normal when their WAZ was ⩾ −2SD.26-28 Socio-demographic and economic characteristics: Caretaker age, sex, type, marital status, education, maternal parity, pregnancy, and breastfeeding status (when the caretaker was the biological mother), income source of household head, number of household members, and children aged <60 months in households. Dietary intake and child characteristics: Introduction of solid, semi-solid, or soft foods, feeding method, breastfeeding status, meal frequency, age, sex, delivery place, type of delivery attendant at birth, number of children given birth to at the time of delivery, and birth order. Child health status characteristics: Vaccination status, sickness (diarrhea, malaria, and respiratory diseases within 2 weeks prior to the study, and measles within 6 months prior to the study), Vitamin A supplementation and deworming medications within 6 months prior to the study. A child was reported to have been vaccinated if their vaccines were up-to-date based on the UNEPI schedule 29 at the time of the study. Diarrhea was defined as more than 3 loose stools in 24 hours. Respiratory disease was defined as an episode of fever and cough with at least one of the following: running nose, wheezing, breathing difficulties, sputum, and chest pain. Measles was defined as having a positive Measles-specific IgM test or a generalized maculopapular skin rash with a fever and at least one of the following: Koplik’s spots, dry cough, runny nose, sore throat, and conjunctivitis. Malaria was defined as having fever with a confirmed positive test (rapid diagnostic or microscopy) or a history of fever plus at least one of the following: chills, headache, muscular aching, and vomiting. 30 Household food access: Household Food Insecurity Access Scale scores (HFIAS), Household Dietary Diversity Scores (HDDS), Food Consumption Scores (FCS), number of household members earning an income, and household food source. HFIAS was measured by asking 9 questions that assessed the risk of food insecurity in the past month. Upon indication of food insecurity in a respondent’s response, a score was given based on the frequency of the experience in the last month. An answer of “No” scored “0,” “yes” with 1 to 2 times scored “1,” “yes” with 3 to 10 times scored “2,” and “yes” with >10 times scored “3.” Then the total score was calculated by the sum of all scores of 9 questions. Total HFIAS scores of 9 and under indicated food security while scores between 10 and 16 indicated marginal food insecurity, and between 17 and 27 was categorized as food-insecure households. 31 HDDS was measured by following the Food and Agricultural Organization of the United Nations (FAO) recommendations. Out of the 12 food groups, a score of 1 point was given for each group consumed while 0 points were given for each group not consumed in the previous 24 hours, assuming usual consumption. If their consumption pattern in the previous 24 hours was unusual, a score for any other day was calculated. The total score was the sum of scores across 12 food groups. Households with HDDS greater than the mean HDDS of the respondents’ households were classified as having good HDDS whereas those with HDDS lower than the mean were classified as having poor HDDS. 32 FCS was used to assess the quality of food consumed by household members in the previous 7 days. FCS represents the number of days where the household diet included food from each of the 8 food groups multiplied by the assigned weight of each group. Total FCS was calculated by the sum of scores across all food groups, and it may range from 0 to 112. Assigned weights are as follows: milk weighting “4”; cereals and tubers weighting “2”; meat, fish, and poultry weighting of “4”; pulses weighting “3”; vegetables and fruits each weighting “1”; and sugar and oil each weighting “0.5.” Scores under 21.5 showed poor quality, between 21.5 and 35 showed borderline quality, and above showed acceptable quality. 33 Household food availability: Agricultural land access, current crop growth, purpose of crop production, fishing, food stocks, livestock ownership, and purpose of livestock. Health system characteristics: Transport mode to health facility (Bussi Health Centre III), ownership of transport means, and transport cost to the health facility. Water sanitation and hygiene (WASH) characteristics: Drinking water source, daily water usage of household, waste disposal method, child’s handwashing before meals, and drinking water treatment method. Research assistants interviewed caretakers face to face using a pre-tested structured electronic questionnaire. The questionnaire had inbuilt skips, validations, and mandatory fields that ensured high-quality data collection. The questionnaire was adopted from the Uganda Demographic Health Survey (UDHS), which was validated by the Uganda Ministry of Health (MoH). Research assistants were trained in the procedures of data collection and taking anthropometric measurements. Pre-tested and standardized digital SECA scales were used to measure the weight of the children to the nearest 0.1 kg. Stadiometers approved by the United Nations Children’s Fund (UNICEF) were used to measure the length of children 6 to 23 months horizontally and the height of children 24 to 59 months vertically. All anthropometric measurements were taken twice, and the mean was entered into the questionnaire. The data collection process was supervised by the principal investigator, and errors were rectified before the questionnaires were uploaded to the server. Five research assistants, each with a motorcycle provided by the study, were recruited to make house-to-house visits to collect data from each household. Anthropometric data were entered into ENA SMART version 2011 where stunting and underweight classification was based on the 2006 World Health Organization (WHO) Growth Standards. Combined with the rest of the data in Excel, it was exported to STATA version 14 for cleaning and subsequent univariate, bivariate and multivariate analysis. Prevalence of stunting and underweight were each obtained by dividing the number of children who were stunted or underweight by the total number of children in the study and expressed as a percentage. Associations between stunting and underweight with independent variables were calculated by Prevalence Ratios (PRs) at their 95% confidence intervals, and p-values of <0.05 showed statistically significant associations. PRs were used instead of Odds Ratios (ORs) since the prevalence rates of stunting and underweight were both greater than 10%, and the ORs tend to overestimate the strength of association in such scenarios.34,35 PRs at both the bivariate (Unadjusted PRs) and multivariate analysis level (Adjusted PRs) were estimated using the Modified Poisson regression analysis, with robust standard errors via generalized linear models with family (Poisson) and link (log). 36 Crude associations were determined at bivariate analysis, and a cut-off point of P ⩽ .2 was used to consider variables for multivariate analysis where confounders were controlled. If variables included in the multivariate model resulted in a loss of significance, they were removed. The final model selection was based on Akaike Information Criteria (AIC) with smaller AIC values suggesting a better model. Covariates with P-values < .05 were considered as predictors of stunting and underweight. This study was approved by Makerere University School of Public Health Higher Degrees Research and Ethics Committee (5th March 2019) and permission to carry out the study was then sought from the District Health Officer of Wakiso District (21st March 2019, Ref. No. 218/03/2019). Participation was voluntary and free from coercion. Written informed consent was obtained from all study respondents and all the information got was kept confidential. Parental consent was obtained before taking anthropometric measurements of all children. All children who were identified to have undernutrition were referred to different health facilities depending on the severity of their conditions.