A Bayesian Geo-Additive Modeling of Childhood Anemia in India

Background: The geographical differences that caused anaemia can be partially explained 2 by the variability in environmental factors, particularly nutrition and infections. The studies 3 failed to explain the non-linear effect of the continuous covariates on childhood anaemia. The 4 present paper aimed to investigate the risk factors of childhood anaemia in India with focus 5 on geographical spatial effect. Methods: Geo-additive logistic regression models were fitted to the data to understand fixed 8 as well as spatial effects of childhood anaemia. Logistic regression was fitted for the 9 categorical variable with outcomes (anaemia (Hb<11) and no anaemia ( Hb≥11) ). Continuous 10 covariates were modelled by the penalized spline and spatial effects were smoothed by the 11 two-dimensional spline. credible influence unobserved factors on childhood thereafter childhood thereafter increases. Conclusion: Strong evidence of residual spatial effect to childhood anaemia in India. 1 Government child health programme should gear up in treating childhood anaemia by 2 focusing on known measurable factors such as mother’s education, mother’s anaemia status, 3 family wealth status, child fever, stunting, underweight, and wasting which have been found 4 to be significant in this study, attention should also be given to effects of unknown or 5 unmeasured factors to childhood anaemia at the community level. Special attention to these 6 unmeasurable factors should be focused in the states of central and northern India which have 7 shown significant positive spatial effects. 8

Conclusion: Strong evidence of residual spatial effect to childhood anaemia in India. 1 Government child health programme should gear up in treating childhood anaemia by 2 focusing on known measurable factors such as mother's education, mother's anaemia status, 3 family wealth status, child fever, stunting, underweight, and wasting which have been found 4 to be significant in this study, attention should also be given to effects of unknown or  Keywords: Spatial effects, Geo-additive logistic regression, P-splines, Childhood anaemia. Anemia among children is still a major public health concern in both developed and 13 developing countries. Anemia is a condition in which the number and size of red blood cells 14 or haemoglobin concentration is lower than the established cut-off value (1). Haemoglobin is 15 essential to carry oxygen and if the body has abnormal or low red blood cells or not enough 16 haemoglobin level, there will be a reduced capacity of the blood to carry oxygen to the body 17 tissues. Globally, anemia affects 1.6 billion people, of which 47.4% were preschool-age 18 children (2). According to the World Health Organization (WHO), anemia is considered a 19 severe public health problem if the prevalence is 40 percent or more (2). In India,58.5% 20 percent of children between the age of 6 months to 5 years were anemic during 2015-2016 21 (3). Moreover, studies have acknowledged the high prevalence of anaemia in low and 22 middle-income countries (4), with 67.6% and 65.6% preschool-age children in Africa and 23 South-East Asia suffered from anaemia (2). 24 Iron is an essential element of haemoglobin, and iron deficiency is the most common cause of 1 anaemia. However, deficiency in micronutrient-rich diet, Vitamin A, and Vitamin B12 could 2 be the reason for iron deficiency (5). Also, disease like diarrhea (6), malaria (7), helminth 3 infection, and hookworms (5) increased the risk of anemia. In India, due to various socio-4 economic, cultural, and religious beliefs, dietary food habits also vary amongst the 5 population. Dietary pattern is an essential factor associated with iron intake and absorption. 6 For example, a vegetarian diet may increase the risk of anemia due to the lack of iron 7 fortification (8). Existing literature have also shown that socio-economic factors such as 8 lower maternal education, low economic status (9), and demographic factors such as age and 9 sex of a child (10) affect anaemia. Maternal health status during pregnancy had a significant 10 impact on the health and nutritional status of the child. Evidence from previous studies 11 reported that maternal anaemia, and child nutritional statuses such as wasting, stunting and 12 underweight increased the risk of anaemia (11,12). During the first 5 years of life, children 13 are most vulnerable to iron-deficiency anaemia because of the increased iron requirements 14 due to their rapid growth (13). Iron deficiency anaemia in children is a serious concern 15 because it may increase childhood morbidity, impaired growth development, and have long 16 term effects on cognitive development and school performance (13). 17 Accounting for geographical heterogeneity of anaemia and the possible cause of 18 heterogeneity is vital for the allocation of health resources to prevent and control anaemia. 19 According to Koissi & Högnäs, (2013) ignorance of geographical heterogeneity due to 20 unobserved characteristics could lead to biased estimation of the parameters (14).

21
Geographical heterogeneity could be the effect of the unmeasured factors, which means that 22 the geographical differences of factors that caused anaemia can be partially explained by the 23 variability in environmental (15). Malaria which caused anaemia are known to be associated 24 with altitude and weather conditions such as temperature and rainfall (16). Similarly, soil-25 transmitted helminth infection, which caused anaemia is influence by the distance to water 1 bodies, surface temperature, index of vegetation and rainfall (17). There are a number of 2 studies using different statistical models such as multilevel and spatial mixed model to 3 determine the effect of geographical heterogeneity on childhood anaemia in India (9,10); 4 however, all these studies have overlooked the advantage of using the bivariate spline in 5 modelling geographical heterogeneity. Specifically, the above model failed to explain the 6 non-linear effect of the continuous covariates on childhood anaemia. Thus, the pioneering contribution of this study would be to explore the spatial variation of 8 anaemia among children aged 6 to 59 months using the spatial mixed model by assuming the 9 flexible approach of bivariate splines. More, specifically we want to explore the spatial 10 effects on childhood anaemia which arise due to the unmeasured factors. Identifying the 11 spatial clustering of anaemia and its associated risk factors may help improving the allocation 12 of health resources.

14
Study area and data 15 The study used the fourth round of the Indian National Family and Health Survey (2015-16 2016) which adopted a multi-stage stratified cluster sampling design (18). A total of 699686 17 eligible women between 15-49 years of age completed the interview. The whole data for the 18 present study use child as the unit of analysis, rather than the mother itself. Information was 19 available on 259627 children born in the last five years preceding the survey. The present 20 study excluded the two union territories, Andaman & Nicobar and Lakswdeep as their 21 borders are not connected to the map of India as this will create problem in the estimation of 22 spatial effects. Children with missing haemoglobin level were dropped from the analysis. 1 With this criterion the final analytical sample size of children was 208707. 2 The covariates in the present study were selected based on previous study (15). The outcome 3 variable used in the analysis was based on the categorization of haemoglobin level of children 4 adjusted for altitude giving a binary variable where children whose haemoglobin level was 5 less than 11Hb was categorised as being anaemic otherwise not anaemic. Mother educational 6 level, household wealth index, child cough, child fever, received vitamin A, mother anaemia 7 status, child stunting, wasting, underweight, child birth weight, child birth order, family size, 8 child age, mother age. Duration of breast feeding, child age, and mother age were treated as 9 continuous variables. However, the standard -2SD cut off values of z-scores categorization of 10 height for age, weight for height, and weight for age were used to characterize stunting, 11 wasting and underweight respectively. 12 Statistical analysis 13 Multiple logistic regression model was employed to select potential covariates for childhood 14 anaemia prior to spatial analysis. A significance level of 20% was set for the selection of 15 potential covariates to allow for selection of more variables to be used in the further analysis 16 of spatial modelling. 17 Geo-additive logistic regression models were fitted to the data to understand fixed as well as 18 spatial effects of childhood anaemia. If is the probability that child j from location i being 19 anaemic, then child anaemic status which is binary is distributed as . The 20 following models were fitted to estimate fixed and spatial effects. Bayesian approach was adopted to estimate the parameters and the estimated posterior odds 10 ratio (OR) can be interpreted as the odds ratio from the logistic regression models. The 11 models were fitted using the freely available package bamlss (21) in R (R Core Team, 2020).

12
A total of 40,000 MCMC iterations and 10,000 number of burn in samples were used in the 13 analysis. Convergence of models were checked through autocorrelations and sampling paths.

20
Descriptive results 21 Table 1 provides prevalence of childhood anaemia according to region and states in India.   evident that children from rural, mother with low education, household of poor economic 4 condition show higher prevalence of anaemia than respective counterparts. There is a clear 5 significant difference in childhood anaemia between levels of place of residence, mother's 6 education and household wealth. But the sex of child does not show any significance 7 difference in childhood anaemia. Children with fever shows a tendency of higher prevalence 8 of anaemia. It can also be seen that consumption of vitamin A supplement during childhood 9 is helpful to reduce prevalence of anaemia. Under nutrition of children also shows an increase 10 in prevalence of anaemia. At 5% level of significance the categorical variables, place of 11 residence, mother's education, household economic status, child fever, vitamin A, stunting, 12 wasting, underweight and, mother's anaemic status are all associated with childhood anaemic 13 without controlling for other covariates. The categorical variables child birth order, 14 household size, child birth weight show a non-significant effect on childhood anaemia at 20% 15 level of significance in the preliminary analysis. Therefore, only categorical variables listed 16 in Table 2 are included in the spatial logistic regression model in Table 4.   Table 4 shows fixed effects to childhood anaemia. Place of residence, mother's education, 9 poorest, rich, richest categories of household wealth, fever, cough, child under nutrition and that children whose mothers are anaemic have higher risk of being anaemic than those whose 18 mothers are not anaemic.  2 Another reason behind the geo-additive modelling is the ability to incorporate non-linear 3 effects of continuous variables. In the present study, we incorporated non-linear effects of age 4 of child, mother's age and, duration of breast feeding. 5 Child age has non-linear effect to childhood anaemia (Fig 1). It is evident from  Mother age also has a non-linear effect to childhood anaemia (Fig 2). The functional 7 relationship between childhood anaemia and mother age depicts almost a U shape. This    (Fig 5). With respect to 80% posterior credible interval more states show significant spatial 2 effects (Fig 6). Most of states in northern and central regions show significant positive spatial 3 effects with respect to 95% credible interval. However, almost all states in north-eastern 4 region of India show significant negative spatial effects with regard to the 80% credible 5 interval (Fig 6). states and union territories (UTs,) anaemia is a matter of concern, whereas for states like 19 Haryana, Jharkhand, and Madhya Pradesh it is of extremely serious concern. These three 20 states need to revisit existing programs targeting to address the child health in general and 21 anaemia in particular.

23
Anaemia has a close link with the food habit. Food habit is closely associated with culture 24 and the nature. Geographical settings decide the nature of food supply and the micronutrients.  The prevalence of anaemia among children in rural areas is comparatively higher than their 6 counter part in India. Rural mass in India might be less aware about the balanced diet which 7 has potentials to improve the hemoglobin count. Because as high as one third of rural 8 population in India are illiterate. Ignorance of food items relating to iron content food staff 9 may also add to the problem of anaemia in rural areas. This indicates that mass media 10 campaign to address anaemia should emphasize pictorial and or audio-visual means, rather 11 than on the written leaflets. A distinct negative relationship between wealth quintile and child 12 anaemia is quite evident. This is indicative of the fact that economically poorer households 13 may not be able to afford to procure nutritious food. This calls for better public distribution 14 system which provides subsidized food in India. The system need to keep an eye on 15 regularity, quantity, quality, etc. 16 17 Uneducated mothers are less equipped with knowledge of hygiene and proper knowledge of 18 child care. Unhealthy feeding habit can lead to various types of food related health problems. 19 Feeding practice is closely associated with diahhroeal disease and studies exhibit that there is definitely are less educated and relatively old mothers might take child rearing for granted, as 5 they may already have older children. Other study also indicates U-shape relationship 6 between mother's age and the childhood anaemia (15) and others (10,25) found children born 7 to young mothers are more likely to be anaemic. In India usually the educated and rich women, due to various reasons, do not practice 10 exclusive breast feeding. Exclusive breast feeding in India is usually practiced among the less 11 educated and poor women, as a result a positive association between exclusive breast feeding 12 and childhood anaemia is observed. However, this finding contradicts studies conducted 13 elsewhere (26).

15
Conclusions 16 There is strong evidence of residual spatial effect to childhood anaemia in India. Government 17 child health programme should gear up in treating childhood anaemia by focusing on known 18 measurable factors such as mother's education, mother's anaemia status, family wealth status, 19 child fever, stunting, underweight, and wasting which have been found to be significant in 20 this study, attention should also be given to effects of unknown or unmeasured factors to 21 childhood anaemia at the community level. Special attention to these unmeasurable factors    Non linear effect of mother age to childhood anaemia. Lower and Upper lines indicate 95% con dence interval Figure 3 Non linear effect of duration of breast feeding to childhood anaemia. Lower and Upper lines indicate 95% con dence interval. Residual spatial effect to childhood anaemia. Colour ranges from black to white representing low to high risk of childhood anaemia. Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area o bbnhjr of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors. The 95% credible intervals map for prevalence of anaemia. White: negative effect; light black: insigni cant effect; black: positive effect. Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area o bbnhjr of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors. The 80% credible intervals map for prevalence of anaemia. White: negative effect; light black: insigni cant effect; black: positive effect. Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area o bbnhjr of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.