Community-Level Sanitation and Child Malnutrition Puzzles in Rural India

Nancy Luke, Pennsylvania State University
Hongwei Xu, University of Michigan
Marco Faytong-Haro, Pennsylvania State University
Winsley Rose, Christian Medical College

Health researchers face an intriguing puzzle: Despite a high level of development, Indian children are among the most severely malnourished worldwide. One hypothesis is that large portions of the Indian population defecate in “open space” outdoors, exposing children to bacteria that can cause diarrhea, infection, and poor nutritional status. We test this hypothesis using new data from the South India Community Health Study in Tamil Nadu undertaken in 400 villages. Combining measures of toilet facility usage and GPS coordinates of households with anthropometric data from 2500 children ages 0-10, we examine the relationship between neighborhood-level measures of open defecation and child nutritional status. We also test a “gendered hygiene hypothesis,” which argues that girls have better nutritional status than boys due to cultural restrictions in their everyday movements. Hence, we expect the negative association between community measures of open defecation and nutritional status will be less strong for girls.

Health researchers face an intriguing puzzle: Despite a higher level of economic development, Indian children are among the most severely malnourished worldwide. In 2005/6, 48% of Indian children under age 5 were stunted and 20% were wasted (IIPS and Macro 2007). Not only do the immediate effects of malnutrition impact child health and development, but poor health and disease in childhood can have lasting effects on human capital in later life.

One hypothesis for these high rates of child malnutrition is related to local sanitation practices. Across India, large portions of the population continue to defecate in “open space” outdoors instead of using communal or household toilet facilities. In these open fields or spaces, children are exposed to or ingest fecal bacteria that can cause diarrhea, infections, malnutrition, and death.

Another puzzle is why girls have better nutritional status than boys in India, given that parents generally invest more resources in boys, and girls have poorer health outcomes for other conditions. Clough (2010) has postulated a “gendered hygiene hypothesis,” which recognizes that cultural standards of hygiene and cleanliness tend be higher for girls than boys, and thus girls are less exposed to allergens, bacteria, and other microorganisms. She uses this reasoning to explain why girls have higher rates of asthma, allergies, and other immune diseases than boys in many industrialized countries, as exposure is believed to foster immune response in these cases.

We contend that a similar argument could be applied to sanitation practices and gender differences in child malnutrition in India. Cultural restrictions in Indian girls’ everyday movements, including their play activities due to safety concerns and their greater likelihood of engaging in domestic chores at home, could limit their exposure to human waste pathogens in the wider neighborhood compared to boys. In this case, girls will be less impacted by neighborhood exposure than boys and have better nutritional status outcomes. Therefore, we expect the negative association between community measures of open defecation and nutritional status to be less strong for girls.

We test these hypotheses using unique new data from the South India Community Health Study (SICHS). SICHS covers 400 villages in rural Vellore district, Tamil Nadu. The SICHS project began with a census of all households in the study area. The study team visited each of the 300,000 households and collected basic demographic and socioeconomic information and GPS coordinates. With respect to sanitation, the census collected information on toilet facilities used by each household. We find that 73.7% households have members who defecate in open space.

We propose several measures of neighborhood sanitation:

  1. Proportion of households using open defecation within the neighborhood
  2. Density of households using open defecation within the spatial boundaries of the neighborhood
  3. Proportion of households using open defecation within the neighborhood weighted by each household’s spatial proximity to other households in the neighborhood. This is an egocentric measure accounting for nearness of neighbors and its values vary across individual households within the same neighborhood.

The first measure serves as a benchmark for our study and allows comparison with other studies where GPS coordinates of individual households are not available. The second measure also takes into account variation in the geographic size of a neighborhood. The risk of exposure is probably larger in a small neighborhood than in a large neighborhood given the same proportion of households using open defecation within a neighborhood. The third measure recognizes the potential within-neighborhood heterogeneity in environmental exposure at a fine-grained scale and is likely to capture individuals’ neighborhood experience more accurately. The egocentric measure implicitly incorporates the aspect of spatial density through distance-based weighting.

The SICHS census data were used as a sampling frame for a household survey completed in 2016. The sample targeted 5000 ever-married men between the ages of 25 and 60, and the study team undertook assessments with selected men, their wives, and their children. We use anthropometric information on height and weight for children ages 0-10 to construct Z-scores for stunting, wasting, and underweight. In the SICHS sample, 26.7% of boys and 22.6% of girls were underweight and 19.0% of both boys and girls were stunted. For children ages 0-5, 13.8% of boys and 12.9% of girls were wasted.

We use a multilevel random intercept logistic model with children (level 1) nested within neighborhoods (level 2) to estimate the probability of children ages 0-10 being stunted, wasted, or underweight. The key independent variables are the three measures of neighborhood-level sanitation. We will fit a separate model for each of the three measures. To test the second hypothesis, we will add a cross-level interaction between child gender and neighborhood sanitation.

Clough, Sharyn. 2010. Gender and the hygiene hypothesis. Social Science & Medicine 72(4):486-493.

International Institute Population Sciences (IIPS) & Macro International. 2007. National Family Health Survey (NFHS-3), India, 2005-06. Mumbai: IIPS.

Presented in Session 1179: Health, Wellbeing, and Morbidity