Introduction

Stunting in infancy is the best indicator for measuring children’s welfare and reflects the existing social inequalities accurately. Stunting in toddler is a chronic nutritional problem, exacerbated with morbidity, infectious diseases, and environmental concern^1,2. It is associated as weel with delays motor development and lower intelligence levels³. Furthermore, it leads to the depression of the immune function, metabolic changes, decreased motor development, low cognitive scores, and low academic scores. As an adult, the stunted toddlers will be at risk of degenerative diseases, decreased intellectual capabilities and productivity, low birth weight (LBW) in delivery, and preterm ^4–6. The factors related with the problem of stunting in toddlers are poverty, including nutrition, health, sanitation, social and cultural, increased exposure to infectious diseases, food insecurity, and access to health services ^7,8.

Stunting is a global incident. The Report Global Nutrition Targets 2025 estimates that around 171 million to 314 million children under five year are stunted and 90% of them are in Asian countries. Indonesia is included in 17 countries among 117 countries, which having three nutritional problems, stunting, wasting, and overweight in toddler ⁹. While the result of Riskesdas showed that the proportion of stunting in Indonesia in 2007 was 37.6% declined in 2010 to 35.8%, which declined by 1.2%, but in 2013 increased again to 37.2% and declined to 30.8% in 2018 ¹⁰. However, this proportion is still above the threshold (cut-off) non-public health problem agreed by WHO that stunting above 20% is still a public health problem ^11,12.  The prevalence of stunting in children aged 0-59 months is higher than school-aged children. This period is a crucial period that would affect intelligence and work productivity in the future. If the quality of life of the children is not cared properly in this period, it can lead to various disorders at a later age, one of which is a disruption in the physical growth of children ¹³.

The existence of various problems of stunting will be a negative impact on the next generation proved by several studies which revealing the future danger of stunting. So far, the government has attempted to reduce stunting rates through two approaches, a specific nutritional approach to address the direct causes of stunting and a sensitive nutrition approach to address the indirect causes. Both approaches are integrated efforts to reduce stunting if carried out simultaneously. Nonetheless, there are various obstacles in its implementation. The existence of these conditions encourages the government to set a breakthrough in addressing and reducing stunting. One of the main pillars of the intervention of stunting is recommendations in monitoring the control action plan and evaluation regularly to ensure the delivery and quality of stunting intervention programs as well as result-based planning and budgeting (results-based budgeting and planning) through national and regional programs ^12,14.

One of the purposes of this research is to make easier for policy makers in running the programs to accelerate the handling of stunting in Indonesia especially in developing a responsive predictive index model for stunting in toddler as part of one of the pillars of the stunting intervention action plan mentioned above. This responsive prediction model of stunting is methods that can be used to help set priorities and strategies for public services especially for decision makers in formulating, implementing and evaluating stunting prevention policies. Wardani et al. (2020) have made a simple and responsive model that can predict the incidence of stunting in toddler in IndonesiaHowever, there has been no research on toddler in Indonesia ¹². Therefore, this research is to develop a responsive prediction index model of stunting in toddler in Indonesia.

Method

This research was conducted from June to August 2019 at the Department of Public Nutrition, FEMA – IPB University, Bogor. This study used a model development design through the mathematical formulation of aggregate data from TPB / SDGs metadata indicators through Bapennas and Riskesdas data with a span of 4 years (2015–2018). This indicator is a time-series data cited from the research of Wardani et al, (2020)¹², the results of the literature review related to stunting in Indonesia since the last 17 years. Time series-aggregate data were selected 14 variables based on data availability from 34 provinces in Indonesia. In this study, subjects were children aged 0-59 months. This indicator metadata is one of the ways to measure the achievement of sustainable development in Indonesia by evaluating the progress of development programs that have been implemented¹². Analysis of data used the backward regression method with IBM SPSS Statistics software version 26. Backward regression was used to determine which factors are most influential (independent variable) on the dependent variable with the final form of a multiple linear regression equation. The equation was selected as the best model to predict stunting on toddlers. The general regression mathematical formulation used the equation below:

Y_i = β₀ + β₁ X_1i + β₂ X_2i + ⋯ + β_p-1 X_p-1,1 + ℇ_i

In this case :

Yi is the dependent variable for the observation of i, for i = 1, 2,…, n.

β0, β1, β2,…βp – 1 are parameters

X1i, X2i,…, − 1, i are independent variables, for i = 1, 2,…, n.

ℇi is an error, which is independent and distributes N(0, σ2).

Backward regression was used to determine which factors are most influential (independent variable) on the dependent variable with the final form of a multiple linear regression equation. The selected equation is the best model to predict stunting in toddlers. The presentation of the research results was carried out in tabulated and textual forms. The selected variables are presented in Table 1.

Table 1: Research Variable.

Source: Bappenas RI and Riskesdas (2015-2018)

Result

This study used multiple linear regression models (backward) to produce a selected equation model by eliminating one by one of the independent variables from the regression model formed. The first stage was carried out a regression analysis between the dependent variable (Y), the stunted toddler prevalence in 2018 with some independent variables (Xi), or the value of k following the rules of k + 1 <n. In this study, n = 34 provinces, then the value of k <33. It means that the first stage regression analysis only includes the independent variable (Xi) fewer than 33 variables. The rule is a further consideration to only 14 independent variables include that exist in the closest the year 2017 to the first stage regression analysis. The significance value in the first regression analysis with the selected significance level is 95% (α = 0.05).

Table 2: The First Stage of Stunting Prediction Model in Toddler.

Source: Bappenas RI and Riskesdas (2015-2018)

Table 2 shows that from 14 variables in the first regression results, there are four variables that are eligible to participate in the second stage regression models, such as the incidence of malaria per 1000 people (X3), with p <α (0.000), the literacy rate of population aged ≥15 years (X10), with p <α (0.001), Households had access to proper sanitation services (X13) with p <α (0.001), all methods of CPR to married couple aged 15-49 (X14) with p <α (0.006).

Table 3: Selected Stunting Prediction Model in Toddler.

Source: Processed Secondary Data

After the first stage, the selected variables are obtained, then performed the second stage of advanced test between the dependent variable (Y) which consist of the prevalence of stunted toddler in 2018 with four independent variables from the results of the first stage regression models (X3, X10, X13, and X14) for the data in 2015, 2016, and 2017 with a total of 12 variables from 34 provinces Indonesia. The significance level of the advanced regression analysis selected was 95% (α = 0.05).

The results obtained are presented in Table 3 which shows that from the 12 variables regression results of the second stage is based on the year 2015, 2016, and 2017, there are four variables selected as the best regression model as a prediction model of stunting in toddler, consist of incidence of malaria per 1000 people (X₃) in 2017 with p <α (0.000), the literacy rate of population aged ≥15 years (X₁₀) in 2015 with p <α (0.001), the households had access to proper sanitation services (X₁₃) in 2016 with p <α (0.001), and all methods of CPR to married couples aged 15-49 years (X₁₄) in 2016 with p <α (0.006).

The regression analysis model of the second stage is the best regression model selected as a prediction model of stunting in toddlers in Indonesia. The second stage regression model has a value of R² and R_adj respectively 0.731 and 0.694, which means that the diversity of the prevalence of stunting in toddlers in Indonesia could be explained by the regression model is by 73.1% or 69.4% and the rest can be explained by other factors. Based on the coefficient values of R2 and R adj, it means that the best regression model developed in this study is quite well used to estimate the prevalence of stunted toddlers in Indonesia. The regression equation, which becomes a prediction model for the responsiveness of stunting in toddlers in Indonesia stated as follows:

Ŷ = 105,894 – 0,498X₃ – 0,520X₁₀ – 0,286X₁₃ – 0,166X₁₄

This equation explains that if the incidence of malaria per 1000 people (X3), the literacy rate of population aged ≥15 years (X₁₀), households had access to proper sanitation services (X13), and all methods of CPR to married couple aged 15-49 years (X₁₄) are zero, then the estimated prevalence of stunted toddler (Y) is 105,894. When the incidence of malaria per 1000 people (X₃) lowered one unit, the prevalence of stunted toddler (Y) declined by 0,498. Furthermore, if the literacy rate of the population aged ≥15 years (X₁₀) increased by one unit, the prevalence of stunting (Y) declined by 0,520. Similarly, if households had access to proper sanitation services (X₁₃) increased by one unit, the prevalence of stunting (Y) declined by 0,286. As well as, all methods of CPR to married couples aged 15-49 years (X₁₄) increased by one unit, the prevalence of stunting (Y) declined by 0,166.

Discussion

The regression analysis model of the second stage is the best regression model selected as a prediction model of stunting in toddlers in Indonesia. The selected variables are the incidence of malaria per 1000 people, literacy rate of the population aged ≥15years, households had access to proper sanitation services, and all methods of CPR to married couples aged 15-49 years. The model obtained in this study compared with the research results by Wardani et al. (2020)¹² described corroborately the diversity of factors on the prevalence of stunting in toddlers with the predictor’s value of 73.1% or 69.4% compared to Wardani et al. (2020)¹², which only has 49.9% or 44.9% are equally derived from the value of R² and R_adj. Nevertheless, two of the four variables selected in this study have the same effect on the prevalence of stunted toddlers consisting of problems of malaria and access to proper sanitation services.

Malaria is an infectious disease caused by Plasmodium sp and transmitted through the bite of a female Anopheles mosquito¹⁵. Malaria is one of the problems set out in the Sustainable Development Goals (SDGs) targeted to be terminated in 2030. Malaria can occur due to the interaction among the environment, people, and mosquitoes ¹⁶. Malaria in this study is the first to suspect predictors of the incidence of stunted toddlers in Indonesia. Several regions in Indonesia are classified as malaria-endemic areas are Papua, West Papua, East Nusa Tenggara, Maluku, and North Maluku ¹². Although there is no relationship between stunting with malaria in malaria-endemic areas, the same study by Dal Bom et al. (2019) noted that malaria was significantly associated with the length for age z-score (LAZ) in the first year of life ¹⁷. Pinceli et al. (2018) stated that mothers who suffered from malaria during pregnancy have a difference in the length of the baby’s birth by 0.47 cm shorter than babies whose mothers did not have malaria¹⁸. It is corroborated by Schmidt et al. (2002) that birth length is a strong predictor of stunting¹⁹. Other than that, Natama et al. (2018) revealed that exposure to malaria in antenatal can interfere with the development of the fetus and the baby’s innate immunity, so that the levels of chemokines, cytokines, and growth factors will be lower²⁰. The existence of these circumstances leads to a higher risk of malaria infection in infants aged 6 to 12 months after birth, especially in malaria-endemic areas.

The second predictor of stunted toddlers in Indonesia in this study is the literacy rate of the population aged ≥15 years. A mother’s level of education and intelligence is one of the factors that affect the nutritional status of toddlers. The level of education, high intelligence, and good stimulation of mothers at home can act as protective factors²¹. It can reduce the detrimental effects of low birth weight (LBW) or malnutrition in early childhood on development. Conversely, the same nutritional condition leads to more severe effects on the development of the child if the mother was illiterate. Research results by Uliyanti (2017) showed that mothers who had a high nutritional knowledge only 27.5%, the rest had moderate nutritional knowledge by 56.9% and 15.7% had low nutritional knowledge at South Matan Hilir Subdistrict ²². Low maternal nutrition knowledge is allegedly closely associated with Mother’s level of education and there are still people around who are illiterate, so the access and the opportunity to gain knowledge of nutrition are limited. In addition, the findings by Torlesse et al. (2016) states that mothers who do not complete primary education likely to increase the incidence of stunting 3.3 times greater chance than mothers who have a high level of education²³.

The third predictor of child stunting in Indonesia in this study is households that have access to proper sanitation services. One of the determinants of stunting in toddlers is the problem of washing hands before preparing or feeding children ^12,24. Besides, there are still people who do open defecation, especially children living in rural and suburban areas. Research by Fregonese et al. (2016) states that children living in a contaminated environment with inadequate sanitation have experienced a 40% risk of stunting and significantly higher in rural (43%) and suburban areas (27%) compared to those living in urban areas (5%)²⁵. The highest incidence rate of stunted toddler is who lives in the rural area because most people still do open defecation ²⁶. Families with low environmental sanitation can increase 8.5 times greater risk of children experiencing stunting ²⁷. It was corroborated by research by Rah et al. (2015) that behavior related to lack of access to proper sanitation services is associated with a reduced risk of stunting by 15% ²⁸. Lack of adequate sanitation facilities can increase disease originating from fecal bacterial contaminants that can infect the intestines of children when inserting a finger into the mouth. Conditions of intestinal infections such as diarrhea and worm infection can affect children’s nutritional status by reducing appetite and interfering with the absorption of nutrients causing the children to suffer from malnutrition and impaired growth^29,30. Families with ownership of water facilities and latrines had a lower prevalence of diarrhea and stunting. Therefore, an increase in household access to adequate sanitation services is required to reduce the prevalence of stunting in toddlers ³¹.

The fourth predictors of the occurrence of stunted toddlers in Indonesia in this study are all methods of CPR (Contraception Prevalence Rate) in married couples aged 15-49 years. CPR especially for women of childbearing age (WCA) is an indicator of family planning on the nutritional status of children. CPR is related to the literacy rate of women, the poverty rate and anemia in women. The use of contraception does not directly reduce the prevalence of stunting but has an indirect effect on the biological and reproductive mother and child. During pregnancy and breastfeeding, maternal nutritional intake dwindles, if it occurs in the teenage, it will impact the children’s linear growth process. Ideally, reproductive health is around the age of 21 years. Birth in adolescence can increase the risk of a baby born with stunting. This does not affect the risk of children but also has an impact on maternal risk for reproductive rudimentary that can lead to death. The use of contraception can reduce the risk of death for pregnant women due to the interval of pregnancy. Contraceptive use can improve the nutritional status of mothers and children to the maximum before the next pregnancy³². Mothers with long pregnancy intervals can restore their nutritional intake and body weight ³³.

The responsive prediction model of stunting obtained can be used to help set priorities and public service strategies, especially for decision-makers in formulating, implementing and evaluating stunting prevention policies in Indonesia to support the acceleration of stunting management in Indonesia.

Limitations of the Study

This research is research that uses secondary data. Model is a method that can reduce information to stakeholders but needs attention and careful interpretation because Indonesia is a country with a wide area as well as the diversity of each province. Limited data is also an obstacle to the necessary mapping with small-scale coverage at the district and sub-district levels.

Conclusion         

The best regression model selected as a prediction model of stunting in toddlers n Indonesia is Y = 105,894 – 0,498₃– 0,520₁₀– 0,286₁₃– 0,166₁₄. The regression model can explain the diversity of the prevalence of stunted toddlers in Indonesia by 73.1% or 69.4% of the coefficient value of R² and R_adj, and the rest can be explained by other factors. Related to these results, it is advisable to consider the fulfillment of access to household sanitation facilities, especially in the region of residence and vulnerable groups such as malaria-endemic areas, increased literacy especially the mother, and increased use of all methods of contraception to the married couple 15-49 years.

Acknowledgment

Thank you to the Education Fund Management Institution (LPDP) of the Ministry of Finance of the Republic of Indonesia for providing a full scholarship.

Conflict of Interest

The author declares that there is no conflict of interest.

References

1. Vonaesch P, Tondeur L, Breurec S, Bata P, Nguyen LBL, Frank T, et al. Factors associated with stunting in healthy children aged 5 years and less living in Bangui (RCA). PLOS ONE. 2017 Aug 10;12(8):e0182363.
   CrossRef
2. Yuliastini S, Sudiarti T, Sartika RAD. Factors Related to Stunting among Children Age 6-59 Months in Babakan Madang Sub-District, West Java, Indonesia. Current Research in Nutrition and Food Science Journal. 2020 Aug 25;8(2):454–61.
   CrossRef
3. Wondimagegn ZT. Magnitude and Determinants of Stunting Among Children in Africa: A Systematic Review. Current Research in Nutrition and Food Science Journal. 2014 Aug 28;2(2):88–93.
   CrossRef
4. Adekanmbi VT, Kayode GA, Uthman OA. Individual and contextual factors associated with childhood stunting in Nigeria: a multilevel analysis. Matern Child Nutr. 2013 Apr;9(2):244–59.
   CrossRef
5. Adair LS, Fall CH, Osmond C, Stein AD, Martorell R, Ramirez-Zea M, et al. Associations of linear growth and relative weight gain during early life with adult health and human capital in countries of low and middle income: findings from five birth cohort studies. The Lancet. 2013 Aug;382(9891):525–34.
   CrossRef
6. Black RE, Victora CG, Walker SP, Bhutta ZA, Christian P, de Onis M, et al. Maternal and child undernutrition and overweight in low-income and middle-income countries. The Lancet. 2013 Aug;382(9890):427–51.
   CrossRef
7. Kemenkes RI. Riset Kesehatan Dasar tahun 2013. Jakarta: Badan Litbangkes Kementrian Kesehatan RI; 2013.
8. Aridiyah FO, Rohmawati N, Ririanty M. The Factors Affecting Stunting on Toddlers in Rural and Urban Areas. Pustaka Kesehatan. 2015 Jan 17;3(1):163–70.
9. de Onis M, Dewey KG, Borghi E, Onyango AW, Blössner M, Daelmans B, et al. The World Health Organization’s global target for reducing childhood stunting by 2025: rationale and proposed actions. Matern Child Nutr. 2013 Sep 18;9(Suppl 2):6–26.
   CrossRef
10. Kemenkes RI. Riset Kesehatan Dasar (Riskesdas). Jakarta: Badan Penelitian dan Pengembangan Kesehatan Kementerian Kesehatan RI; 2018.
11. Ruaida N. Gerakan 1000 Hari Pertama Kehidupan Mencegah Terjadinya Stunting (Gizi Pendek) Di Indonesia. GLOBAL HEALTH SCIENCE. 2018 Jun 30;3(2):139–51.
12. Wardani Z, Sukandar D, Baliwati YF, Riyadi H. Akses Sanitasi, Merokok dan Annual Parasite Incidence Malaria sebagai Prediktor Stunting Baduta di Indonesia. MKMI. 2020 Mar 31;16(1):127–39.
    CrossRef
13. Kullu VM, Yasnani, Lestari H. Faktor-Faktor Yang Berhubungan Dengan Kejadian Stunting Pada Balita Usia 24-59 Bulan Di Desa Wawatu Kecamatan Moramo Utara Kabupaten Konawe Selatan Tahun 2017. JIMKESMAS. 2018;3(2):1–11.
    CrossRef
14. Laksono AD, Megatsari H. Determinan Balita Stunting di Jawa Timur: Analisis Data Pemantauan Status Gizi 2017. Amerta Nutrition. 2020 Jun 18;4(2):109–15.
    CrossRef
15. Zekar L, Sharman T. Plasmodium Falciparum Malaria. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2022 [cited 2022 Feb 3]. Available from: http://www.ncbi.nlm.nih.gov/books/NBK555962/
16. Selvia D. Keluar Rumah pada Malam Hari dan Penggunaan Kelambu Berinsektisida dengan Penyakit Malaria di Desa Lempasing. JIKA. 2019 Dec 31;1(2):89–95.
    CrossRef
17. Dal Bom JP, Mazzucchetti L, Malta MB, Ladeia-Andrade S, de Castro MC, Cardoso MA, et al. Early determinants of linear growth and weight attained in the first year of life in a malaria-endemic region. Boeuf P, editor. PLoS ONE. 2019 Aug 6;14(8):e0220513.
    CrossRef
18. Pincelli A, Neves PAR, Lourenço BH, Corder RM, Malta MB, Sampaio-Silva J, et al. The Hidden Burden of Plasmodium vivax Malaria in Pregnancy in the Amazon: An Observational Study in Northwestern Brazil. The American Journal of Tropical Medicine and Hygiene. 2018 Jul 5;99(1):73–83.
    CrossRef
19. Schmidt MK, Muslimatun S, West CE, Schultink W, Gross R, Hautvast JGAJ. Nutritional Status and Linear Growth of Indonesian Infants in West Java Are Determined More by Prenatal Environment than by Postnatal Factors. The Journal of Nutrition. 2002 Aug 1;132(8):2202–7.
    CrossRef
20. Natama HM, Moncunill G, Rovira-Vallbona E, Sanz H, Sorgho H, Aguilar R, et al. Modulation of innate immune responses at birth by prenatal malaria exposure and association with malaria risk during the first year of life. BMC Med. 2018 Dec;16(1):198.
    CrossRef
21. Committee on Supporting the Parents of Young Children, Board on Children, Youth, and Families, Division of Behavioral and Social Sciences and Education, National Academies of Sciences, Engineering, and Medicine. Parenting Matters: Supporting Parents of Children Ages 0-8 [Internet]. Gadsden VL, Ford M, Breiner H, editors. Washington, D.C.: National Academies Press; 2016 [cited 2022 Feb 3]. Available from: https://www.nap.edu/catalog/21868
22. Uliyanti U, Tamtomo DG, Anantanyu S. Faktor Yang Berhubungan dengan Kejadian Stunting Pada Balita Usia 24-59 Bulan. Jurnal Vokasi Kesehatan. 2017 Jul 31;3(2):67–77.
    CrossRef
23. Torlesse H, Cronin AA, Sebayang SK, Nandy R. Determinants of stunting in Indonesian children: evidence from a cross-sectional survey indicate a prominent role for the water, sanitation and hygiene sector in stunting reduction. BMC Public Health. 2016 Jul 29;16(1):669.
    CrossRef
24. Nasrul, Fahmi H, Thaha A, Suriah. Faktor Resiko Stunting Usia 6 – 23 Bulan di Kecamatan Bontoramba Kabupaten Jeneponto. Media Kesehatan Masyarakat Indonesia. 2016;11(3):139–46.
25. Fregonese F, Siekmans K, Kouanda S, Druetz T, Ly A, Diabaté S, et al. Impact of contaminated household environment on stunting in children aged 12–59 months in Burkina Faso. J Epidemiol Community Health. 2017 Apr;71(4):356–63.
    CrossRef
26. Chakravarty I, Bhattacharya A, Das SK. Water, sanitation and hygiene: The unfinished agenda in the World Health Organization South-East Asia Region. WHO South-East Asia Journal of Public Health. 2017 Jul 1;6(2):22–6.
    CrossRef
27. Kusumawati E, Rahardjo S, Sari HP. Model Pengendalian Faktor Risiko Stunting pada Anak Bawah Tiga Tahun. Kesmas: Jurnal Kesehatan Masyarakat Nasional (National Public Health Journal). 2015 Apr 1;9(3):249–56.
    CrossRef
28. Rah JH, Cronin AA, Badgaiyan B, Aguayo VM, Coates S, Ahmed S. Household sanitation and personal hygiene practices are associated with child stunting in rural India: a cross-sectional analysis of surveys. BMJ Open. 2015 Feb 12;5(2):e005180–e005180.
    CrossRef
29. Prendergast AJ, Humphrey JH. The stunting syndrome in developing countries. Paediatr Int Child Health. 2014 Apr;34(4):250–65.
    CrossRef
30. Owino V, Ahmed T, Freemark M, Kelly P, Loy A, Manary M, et al. Environmental Enteric Dysfunction and Growth Failure/Stunting in Global Child Health. Pediatrics. 2016 Dec 1;138(6):e20160641.
    CrossRef
31. van der Hoek W, Feenstra SG, Konradsen F. Availability of irrigation water for domestic use in Pakistan: its impact on the prevalence of diarrhea and nutritional status of children. J Health Popul Nutr. 2002 Mar;20(1):77–84.
32. Zapata LB, Murtaza S, Whiteman MK, Jamieson DJ, Robbins CL, Marchbanks PA, et al. Contraceptive counseling and postpartum contraceptive use. Am J Obstet Gynecol. 2015 Feb;212(2):171.e1-171.e8.
    CrossRef
33. Goli S, Rammohan A, Singh D. The Effect of Early Marriages and Early Childbearing on Women’s Nutritional Status in India. Matern Child Health J. 2015 Aug;19(8):1864–80.
    CrossRef