Social Determinants and Access to Water-Sanitation-Hygiene as Dominant Risk Factors of Stunting among Under-five Children in Rural Area of East Indonesia

(1) Background: Stunting remains a challenge in Indonesia, where 30.8% of under-five children are stunted and may never reach full potential height and cognitive function. We aimed to investigate the risk factors of stunting in Nangapanda subdistrict, East Nusa Tenggara; (2) Methods: The design was cross-sectional study located in rural area as part of PINTERMIDI UI project. We collected quantitative data on social determinants, weight, height, hemoglobin, ferritin, serum zinc, CRP, worm infection, history of deworming, co-morbidity, food security, and nutrient intake; (3) Results: A total of 196 under-five children included and 74% households were food insecure. The prevalence of stunting was 29.9% in this population (31.8% worm infection, 30.3% iron deficiency, and 28.1% zinc deficiency). Multivariate analysis showed household with 5-8 members (AOR 3.076; 95% CI 1.132 – 8.356) and unsafe drinking water (AOR 1.702; 95% CI 0.825 – 3.512) were significant independent risk factors of stunting after adjusted by child’s gender, father’s occupation, caregiver’s education, monthly expenses, sanitary facilities, and food security status; (4) Conclusions: The number of household members is the only independent risk factor of stunting among children age 24-59 months in rural area of Eastern Indonesia. Development of nutrition sensitive intervention and promotion of family planning are needed in order to increase adequate child’s care and feeding practices in rural area of Eastern Indonesia.


Study Design and Setting
The study design was cross-sectional study located in rural area, Nangapanda subdistrict, Ende, East Nusa Tenggara, Indonesia as a part of PINTERMIDI UI project. The data were obtained from 24 villages in selected subdistrict that covers coastal and upper land area.

Data Collection
Data collection was performed in October 2019. We trained local enumerators during two days before data collection. The standard questionnaires were pretested prior to implementation in the field and pilot survey was also conducted. We collected data on social determinants characteristics, anthropometric assessments, co-morbidity, history of deworming, WASH, food security status and nutrient intake. Biochemical measurements (serum ferritin, serum zinc, serum C-Reactive Protein) and parasitical examination were performed in sub-sample (± 100 subjects). Field coordinators were assigned to supervise enumerators collecting complete questionnaires in the field. Quality assurance steps were also considered during data entry and cleaning.

Interview Structured Questionnaires
The social determinants data included children's age and gender, parent's education and occupation, family household size, monthly expenses, and housing situation. The level of education was defined as years of education completed: 1) low (never went to school to junior high school graduates); 2) middle (graduated from senior high school); 3) high (graduated from diploma/bachelor degree and beyond). The type of occupation was grouped as 1) formal (public/private employee/Indonesian migrant worker); 2) semiformal (farmer/fisherman/seller/labor/self-employed, etc.); 3) not working (retired/not working).
Source of electricity was defined whether household used national electricity or not.
We recorded co-morbidity data by the occurrence of upper respiratory tract infection, fever, and diarrhea in last two weeks and measles in last six months as self-reported by caregivers. In addition, we identified the occurrence of worm infection (from mouth or feces) and history of received deworming in last one year. The indicators of WASH (water, sanitation, and hygiene) were source of drinking water, sanitary facility, and mother/caregiver's personal hygiene. We defined source of drinking water as 1) safe (tube well/ piped water); 2) unsafe (mineral water/ refill water/ well spring/ surface water/ Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 13 July 2021 doi:10.20944/preprints202107.0279.v1 rainwater/ water shelter) [13]. For sanitary facility, we observed whether household had: 1) private latrine; 2) public latrine; 3) no facility. Mother/caregiver's personal hygiene was self-reported whether they take handwash: 1) before food preparation; 2) before eat; 3) before feeding the children; 4) after from toilet; 5) after children's bathing. Food security was identified using the standard household food insecurity access scale in last month developed by Food and Nutrition Technical Assistance (FANTA) project [14]. It was then categorized as food secure and food insecure.

Dietary Assessments
Dietary intake was estimated by using the first 24-h food recall for all samples and the second 24-h food recall for sub samples in order to obtain the percentage of nutrient adequacy in this population. We provided Indonesian food picture book to estimate the children's consumption. The dietary data was then inputted to Nutrisurvey software and was analyzed using MSM (Multiple Source Method) program to estimate usual intake of nutrients [15]. The Estimated Average Requirement (EAR) was used to categorize as [16]: 1) adequate (nutrient intake ≥ EAR); 2) inadequate (nutrient intake ˂ EAR) for mineral (calcium, iron, zinc) and vitamin (A, B1, B2, B3, folate, B12, C).

Anthropometric Measurements
The survey employed trained anthropometric staff to ensure data quality with respect to the all the procedures and survey activities. Weight was measured using SECA weight scale and height was measured by Shorr-board. The weight scale was calibrated daily for any possible errors. The data was then calculated height for age z-score (HAZ-score) and weight for age z-score (WAZ) by using WHO Anthro software [17]. Stunting defined as HAZ-score of ˂ -2 and underweight defined as WAZ-score of ˂ -2 based on the WHO Child Growth Standards [18].

Biochemical Measurements
Trained laboratory analyst obtained venous blood from the children using single-use sterile syringe and placed the blood in a red top microtainer. We used Hemocue Hb 201+ that needs disposable microcuvatte to measure hemoglobin levels in a drop of blood samples after 5 hours of blood collection. Hemocue was calibrated daily for any possible errors. Anemia defined as hemoglobin ˂ 11 mg/dL [19]. The blood was centrifuged approximately 5 hours after blood collection. Serum was then transported to Jakarta and was stored at -20 0 C for analysis of zinc, ferritin, and C-Reactive Protein (CRP) in Laboratory of SEAMEO RECFON. To evaluate zinc deficiency, we used serum zinc analysis using Atomic Absorption Spectrophotometry (AAS). Serum zinc concentrations was then adjusted for CRP to consider inflammation response [20]. Zinc deficiency defined as serum zinc < 9.9 µmol/L [21]. Serum ferritin were measured using Fluorescence Enzyme Immunoassay (FEIA) for iron deficiency identification. The ferritin concentrations was adjusted for CRP to consider that ferritin is an acute phase protein and is raised in inflammatory conditions [22]. Iron deficiency defined as serum ferritin < 12 mg/dL [23].

Parasitical Examinations
A single stool sample was collected from each child using a stool container. Kato-Katz is the most widely used microscopy technique to detect a soil-transmitted helminth (STH) species and to quantify the infection [24]. A single Kato-Katz smear was used to identify eggs and larvae of helminths. A 41.7 mg template was placed on the microscopy slide and filled with sieved stool material. The sample was covered then with a cellophane slip soaked with glycerol and water. After that, the sample was flattened by pressing it. The slide has to be examined under the microscope within 30-60 minutes after preparation. In addition, the microscopic examination of protozoan cysts was measured using direct smear technique with lugol staining. These two examinations were performed on the same day as stool sample collected. The remaining sample was aliquoted into several tubes and was transported to Laboratory of Parasitology Universitas Indonesia in Jakarta and was stored at -20 0 C for others analysis. Positive worm infection was identified if we found one of the three STH species that are Ascaris Lumbricoides, hookworm, and whipworm.

Statistical Analysis
Data analysis was conducted using SPPS version 20. First, univariate analysis was used to summarize descriptive data. Second, chi-square and Fisher test was performed to analyze the association between stunting and each of risk factors. Pearson or Spearman correlation was used to analyzed the continues data (HAZ-score, serum zinc, serum ferritin, hemoglobin, zinc intake). Third, variables at OR > 1.5 in all samples (n=196) were selected to be included in the multivariate model. Logistic regression was then used to assess determinants of stunting with adjusted odd ratios (AOR) and 95% confident intervals (CI) for variables.

Characteristics of Respondents
A total of 196 children between the ages of 24-59 months were identified that boys was more than girls (  Table).

Bivariate and Multivariate Analysis between several factors and stunting
From bivariate analysis (

Discussion
This study intended to investigate the important interaction between risk factors and stunting among children age 24 to 59 months in Ende, East Nusa Tenggara. In general, the prevalence of stunting, zinc deficiency, anemia, nutrient inadequacy, and food insecurity were very high in children. We identified that serum ferritin have significance association with HAZ-score. We also found that household with more than four members and unsafe drinking water after controlling for confounders are the main independent factors that related to stunted children in rural area of Indonesia.
We observed the greater prevalence of anemia and iron deficiency among all child's participants. Further analysis showed 34% and 30.3% were anemic and iron deficient among stunted children. Interestingly, we found association between serum ferritin and HAZ-score ( Table 2) and it was supported by our results that showed iron deficient children had two times risks to be stunted ( Table 3). This finding was consistent with a systematic review in Ethiopia that pointed out poor dietary diversity, stunting, food insecurity, and not dewormed were predictors of anemic under-five children [25]. Food insecurity has consequences in both macro and micronutrient deficiencies that increased the risk of child anemia [26]. In fact, we identified high prevalence of household food insecurity ( Table 1) and inadequacy of iron intake in our study (Supplementary table).
Moreover, we found 19% children infected with worm (Table 1), and one study demonstrated that helminths destroy red blood cells, reach hemoglobin and result in anemia in Ethiopian under-five children [25].
In this present analysis, we revealed weak association between serum zinc concentrations and HAZ-score but we found strong association between serum ferritin and HAZ-score ( Table 2), suggesting that other factors might be playing more important role in the etiology of stunting than previously thought. Study in rural area of Guatemala explained the odds of being anemic were 3 times higher among zinc deficient children than those who were not and it was indicated the existence of multiple micronutrient deficiencies [27]. In our case, those given similar conditions were occurred in children age 24-59 months. The mechanism potentially linking zinc concentrations and HAZ-score is the presence of zinc deficiency reduced the lifespan of red blood cells due to the compromised function of zinc enzymes [28]. Then, it leads to iron deficiency, identifying have 2 times risks of being stunted children in our study ( Table 3). Given these explanations, the interaction between parasitic infection, anemia, iron deficiency, and zinc deficiency and stunting is clearly determined.
According to WHO Conceptual Framework on Childhood Stunting [5], basic and underlying causes are the most affected factors of stunting in this study. Children living in rural communities had risk to suffer malnutrition more than in urban settings [7]. In our study, the selected setting was rural area with monthly expenses approximately IDR 600,000,--IDR 1,400,000,-. Moreover, household with ≥ 5 members had 3 times risks to have stunted children after controlled by several factors ( Table 3). It was consistent with previous studies in South Africa and Brazil that showed children living in big family (≥ 5 members) had 2-3 times more likely to develop stunting [29,30]. The poor condition led to lack of purchasing good of food quality, improper food allocation, limit access to adequate Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 13 July 2021 doi:10.20944/preprints202107.0279.v1 health services and WASH and it manifested to poor health condition [9,11,31]. In our current study, the percentage of household with food insecure was high ( Table 1) and most of vitamin and minerals was inadequate (Supplementary table). In addition, children living with household food insecure had 1.3 times risks of being stunted after controlling for confounders ( Table 3). This condition was related to poor household that contributed to high prevalence of stunting in Indonesia [6][7][8][9].
Our study highlights the emerging issue of child stunting in Eastern Indonesia. The data presented provides valuable insight into the prevalence of stunting in children which has remained fairly consistent in Indonesia over the last decade. Many specific and sensitive nutrition interventions implemented by Ministry of Indonesia should be improved in respond to specific problem in each condition. However, the observations of the study should be interpreted in the context of certain limitations. First, the crosssectional nature of this study means temporal relationship between stunting and its associated factors could not be established. Second, biochemical measurements and parasitical examination is only 50% of total sample so that it is not sufficient to include in logistic regression. Third, our present analysis did not use own's asset to define wealth index.

Conclusions
In summary, our findings indicate that the number of household members is the only independent risk factor of stunting among children age 24-59 months in rural area of Eastern Indonesia. Development of nutrition sensitive intervention on agriculture to improve food security especially for large-sized households might be needed in this area.
Furthermore, promotion of family planning need to be emphasized in order to increase adequate child's care and feeding practices in rural area of Eastern Indonesia.
Supplementary Materials: Table 1. Percentage of nutrient inadequacy in underfive children.
Author Contributions: GW, UF, and TS conceived the study. DS, GW, EI, and TS trained and supervised the enumerators in the field. DS, GW, EI, UF, and TS implemented and interpreted the study.
DS and EI conducted data cleaning. DS and GW provided data analysis and drafted the manuscript.
UF and TS provided advice and revision of the final manuscript. All authors critically reviewed and approved the final manuscript.