UNDERNUTRITION PREVALENCE AMONG CHILDREN UNDER TWO YEARS OLD IN INDONESIA DURING ECONOMIC CRISIS AND ITS RELATED FACTORS

The objective of this study was to assess undernutrition prevalence among children under-two years in Indonesia and its related factors. This ecological study was conducted to 437 districts using data extracted from nationally representative surveys, which were National Socioeconomic Survey (Susenas) and National Basic Health Research (Riskesdas) year 2007. Overall, prevalence of High Stunting (HS), High Wasting (HW), and High Underweight (HU) was 46.9%, 59.5%, and 17.2%, respectively. Median prevalence of inadequate intake of energy, protein, iron, and zinc were 33.3%, 11.1%, 81.8%, and 55.6%, respectively. Median prevalence of Acute Respiratory Infection (ARI) was 15.9%, while the median prevalence of diarrhea was 13.0%. Median proportion of low education of the mother was 46.9%. Prevalence of HS was associated with household income, low education proportion of mother, and proportion of food expenditure. Prevalence of HW was associated with prevalence of inadequate energy intake and proportion of food expenditure. Prevalence of HU was associated with proportion of low education of mother and proportion of food expenditure.


INTRODUCTION
Global economic crisis as the combination of food, fuel, and financial crisis in 2007-2008 caused increase in global food price.The agricultural production potential (more land and irrigation, higher yields) which were failed to develop with the growing needs made biofuel feedstock prices (in relation to the price of petroleum) arose to uneconomic energy production level (Tyner et al. 2007).The three commodities including all non fuel, metals and oil continued to undergo price increase (Radetski 2006).In term of consumption, this food crisis and high food price was caused by increased food demand as the impact of rapid population growth and high economic growth, especially in China and India (Alexandratos 2008).The rate of price index of wheat, corn, and rice was increased from 16.7% in 2006 became 50.6% in 2008 (Rusastra et al. 2010).Moreover, according to Brinkman et al. (2010), between 2000 and its peak in 2008, the FAO cereal price index was more than tripled.
The food price increase can have direct effect on food security or on insufficient utilization, availability, and accessibility to adequate and safe diets (Darnton-Hill et al. 2010).The study results from Ahmed et al. (2010) showed that the crisis caused increased in poverty at macro level by 80 percent, from 22 to 40 percent, while in micro level, there was 4 percent increase in poverty and 1.3 percent reduced food consumption.
Households have reduced the quantity and quality of food consumed in the face of higher food price.The reduction was seen in the consumption of high-quality foods (Block et al. 2004).Households reduce energy intake in response to increased food price and reduced income.The pathway is by reducing the size and frequency of meals in their consumption.When household reduce the size and number of meals, they will likely suffer deficiencies in macronutrient, including energy and protein (Brinkman et al. 2010).
Inadequate food intake makes the impact of price increase is worse in children as one of vulnerable groups since malnutrition in child is associated with inadequate food intake (Block et al. 2004) Study on undernutrition prevalence among children under-two years using national data is still limited, especially in Indonesia.Therefore, this study is important to be conducted.

Design, Place, and Time
This ecological study was conducted from February to July 2011.The data used in this analysis were collected from nationally representative surveys, which were National Socioeconomic Survey (Susenas) and National Basic Health Research (Riskesdas) year 2007.
Susenas is household survey made to collect socioeconomic data widely in national scope.The sampling of Susenas was multistage cluster sampling.It is charged with a mission to gather socio-economic data among others, concerning education, demography, health/nutrition, housing/environment, criminality, socio-culture activities, household consumption/expenditure, domestic travel, and community opinion regarding their own welfare.The data collection is conducted in February-March 2007.The Susenas collects data on two groups of consumption, i.e. food and non-food consumption.There are 215 items of household food consumption appeared in the questionnaire in terms of quantities and values.For non-food consumption items, which consists of 104 items, only data on value is collected, except for some consumption items such as electricity, water, fuel, and charcoal.The interview is conducted to head of household or other household member who is familiar with the household's affairs.

Total Subject and Sampling Procedure
Total number of children under two years in Riskesdas was 31 569 children from 440 districts in 33 provinces in Indonesia.Since the data was in unit of analysis of district, 3 districts from Riskesdas were taken out to match the Susenas.The z-score was compared to WHO standard for exclusion criteria (HAZ -6 and +6, WHZ -5 and +5, WAZ -6 and +5).Finally, 25 922 children from 437 districts were taken as the subject.
Unit of analysis used in this study was districts.The justification of using district as subject unit was the spatial heterogeneity in socio-economic parameters: poor people tend to be clustered in specific places (MOA et al. 2005).Therefore, analysis on district level was useful to capture the variation which further used as determination in formulating effective policy by decision makers.Subject in this study was 437 districts.

Data Processing and Analysis
All data have been collated from 2 secondary sources of the Central Bureau of Statistics (BPS) and MOH, which were Susenas and Riskesdas year 2007, respectively.Total monthly HH, food, and strategic food expenditure in Rp/cap/mo were obtained from Susenas, while Riskedas provide data on weight, height, birth weight, intake of energy, protein, iron, and zinc, occurrence of Acute Respiratory Infection (ARI) and diarrhea, and mother's education.Stunting, wasting, and underweight was determined by using cutoff <2 SD Z-scores.High stunting was prevalence of 30% upper.High wasting was obtained by cut off of 15% upper, while high underweight was categorized by cut off of 20% upper (WHO 2010).
Data on birth weight of children in gram was transformed into prevalence of Low Birth Weight (LBW) (<2 500 g) per district.When data on birth weight was not available for particular district, provincial data where the respective district existed was used.Nutrient intake of children under-two was compared to nutrient requirements from complementary foods with average breastmilk intake in developing countries obtained from Dewey et al. (2003).A particular nutrient intake was considered adequate if the percentage to RDA ≥77%.Data on individuals was then transformed into prevalence of inadequacy for each nutrient per district.
Data on health status was obtained from prevalence of ARI and diarrhea per district.The data was analyzed in continuous form.Mother's education was considered low based on elementary school (graduated) and less.When data on mother's education in particular district was not available, provincial data where the respective district existed was used.
Total household expenditure was used as proxy for income.Further, income was categorized as low (40% lowest income), medium (40% higher income), and high (20% highest income).Proportion of food expenditure in percentage was obtained by dividing food expenditure by total HH expenditure, both in Rp/cap/mo.Proportion of food expenditure was used also to determine household food security with >60% was considered as food insecure.In each household, expenditures on different food items, including rice, corn, cassava, sweet potatoes, wheat flour, soybean, beef, egg, chicken, milk, fish, sugar, cooking oil, formula milk, and manufactured complementary food (manufactured CF) in the previous week was collected.Expenditure on food was grouped into 7 categories: 1) rice; 2) non-rice staple, the sum of expenditures on corn, cassava, sweet potatoes, and wheat flour; 3) plant protein; 4) animal protein, the sum of expenditures on beef, egg, chicken, milk, and fish; 5) sugar and cooking oil; 6) formula milk; and 7) manufactured CF.
Chi-square tests and bivariate logistic regression were used to examine the relationship between nutritional status and its various underlying indicators.Separate multivariate logistic regression models were performed to examine relationship between food expenditure categories and high stunting, high wasting, and high underweight prevalence.
Variables that are significantly associated with high stunting, high wasting, and high underweight prevalence in the previous test were included in the final multivariate models.P-value <0.05 was considered significant.Data analyses were conducted using SPSS version 16.
This study was conducted after receiving permission from Indonesian Center for Agriculture Socio Economic and Policy Studies (ICASEPS), MOA, to use Susenas data year 2007.Permission from National Institute of Health Research and Development (NIHRD), MOH, was also obtained to use Riskesdas data year 2007.

RESULTS AND DISCUSSION
Overall, prevalence of high stunting, high wasting, and high underweight was 46.9%, 59.5%, and 17.2%, respectively.This showed that malnuttrition problem is still prevalent in Indonesia.Median prevalence of inadequate intake of energy, protein, iron, and zinc among children under-two were 33.3%, 11.1%, 81.8%, and 55.6%, respectively.Prevalence of energy, iron, and zinc were ranged between 0 to 100%, while it was ranged from 0 to 75% for protein.There was significant association of household income with prevalence of inadequate iron and zinc intake.Higher household income was associated with lower prevalence of inadequate iron and zinc intake.With high income, household is capable to find more micronutrient-rich-food, such as animal and plant protein, fruit and vegetable.
Overall, median prevalence of ARI among children under-two was 15.9%, ranged from 0 to 71.4%.Median prevalence of diarrhea was 13.0% with range between 0 and 50%.Higher household income was significantly associated with lower diarrhea prevalence.Higher income gave opportunity to provide high quality of nutrient source.Micronutrients, such as iron and zinc, are responsible to prevent child from diarrhea.
Median low education proportion of the mother was 46.9%, ranged from 6.1% to 100%.This study found no association between income class and low education proportion of the mother.Higher income does not always give opportunity to the mother to have higher education, since the income is used mainly to fulfill need for food.
Prevalence of LBW was not associated with prevalence of high stunting, wasting, and underweight of children under-two (Table 1), while there was significant association between prevalence of inadequate energy intake and prevalence of high wasting among children under-two (Table 2).
Bivariate analysis found no association of prevalence of ARI in the last one month with prevalence of high stunting, wasting, and underweight among children under-two.Similar with ARI, there was no association of prevalence of diarrhea in the last one month with prevalence of high stunting, wasting, and underweight among children undertwo (Table 3).Wasting indicates current malnutrition.Therefore, the result of this study showed that the malnutrition was caused more by inadequate food intake which happen in recent time rather than the disease (Babatunde et al. 2011).
Bivariate analysis showed that household income was associated with prevalence of high stunting among children under-two (Table 4).District with low income were at higher risk of having high stunting.Stunting is an indicator of past growth failure.Therefore, this study showed that the crisis, which happened for quite long time, caused reduction of real income which was responsible for inadequate intake as the cause of stunting (Olivieri et al. 2007).
Low education proportion of the mother was associated with risk of having high prevalence of high stunting and high underweight (Table 4).This was in line with Torlesse et al. (2003), Campbell et al. (2010), andSari et al. (2010)   health and nutrition which is important to practice good caring related to feeding and disease management.
There was significant association of proportion of household food expenditure with prevalence of high stunting, wasting, and underweight among children under-two (Table 5).Higher proportion of household food expenditure was associated with risk of having high prevalence of high stunting, high wasting, and high underweight.The households are food insecure, so that they may have problem in fulfilling the need for food (Melgar-Quinonez et al. 2006).

CONCLUSIONS
Overall, prevalence of high stunting, high wasting, and high underweight was 46.9%, 59.5%, and 17.2%, respectively.Prevalence of high stunting was associated with income, proportion of low education of mother, and proportion of food expenditure.Prevalence of high wasting was associated with prevalence of inadequate energy intake and proportion of food expenditure.Prevalence of high underweight was associated with proportion of low education of mother and proportion of food expenditure.This study recommended that enhancing income generation and education are essential to reduce undernutrition prevalence. .
Total household in Core Susenas 2007 was 274 184.Number of household included in Module Susenas 2007 was 64 880.This study analyzed 14 767 households having children less than two years included in Susenas 2007.The households came from 437 districts in 33 provinces in Indonesia.Riskesdas 2007 is a descriptive cross sectional survey carried by National Institute of Health Research and Development, MOH.The subject of household and household member in Riskesdas 2007 was designed identically with the list of household and household sample in Susenas 2007.The data collection was started from July 2007-September 2008.To obtain data on individual intake of children under-two, Riskesdas conducted single 24-hour recall through interview with head of household or housewives who can give information.Riskesdas 2007 collected information from 258 366 household sampled.

Table 1 .
Association of LBW Prevalence with Undernutrition Prevalence among Children 0-23 Month Old

Table 2 .
Association of Nutrient Inadequacy Prevalence with Undernutrition Prevalence among Children 0-23 Month Old

Table 3 .
Association of Disease Prevalence with Undernutrition Prevalence among Children 0-23 Month Old * Energy, protein, iron and zinc were in term of inadequate intake prevalence ** Logistic regression, p<0.05 *** ARI and diarrhea were in term of prevalence in the last 1 month HS: High Stunting; HW: High Wasting; HU: High Underweight

Table 4 .
Association of Household Income and Low Education Proportion of the Mother with Undernutrition Prevalence among Children 0-23 Month Old * Low income was the reference category ** Logistic regression, p<0.05 *** Low proportion of food expenditure was the reference category HS: High Stunting; HW: High Wasting; HU: High Underweight

Table 5 .
Association of Household Food Expenditure with Undernutrition Prevalence among Children 0-23 Month Old