(P25-005-20) Contextual and Compositional Risk Factors of Overweight and Obesity in India over Two Decades: An Analysis of National Demographic and Health Surveys

Objectives: Identify contextual (state characteristics) and compositional (individual characteristics) drivers of changes in overweight in women from 1998-2016 across 26 states of India.

Methods: Nationally representative data on individual socio-demographics and objectively-measured anthropometry for non-pregnant and ever-married women 15-49y were obtained from the Indian National Family Health Surveys (NFHS-2, 1998-99, n=61,979; NFHS-3, 2005-06, n=66,694; and NFHS-4, 2015-16, n= 387,732). Individual-level data were merged with year-matched state-level nutritional and economic indicators obtained from multiple national databases. State indicators included per capita consumption of cereals, oils and sugar, per capita gross state domestic product (GDP), literacy rates, labour force participation rate, and population proportion engaged in sedentary employment. Cross-classified generalized linear mixed models (GLMM) with random effects classified each woman by state and round of survey. Models examined overweight/obesity trends as well as estimated the adjusted odds ratios and 95% Bayesian credible intervals for overweight/obesity associated with contextual and individual factors.

Results: From 1998 to 2016, the prevalence of overweight/obesity increased in all states of India. Higher household wealth (Quintile 5 vs 1: 4.18 [4.06-4.30]), education (post-secondary vs pre-school or less: 1.55 [1.51-1.60]), were associated with higher adjusted odds of overweight/obesity.  State average sugar (g; 1.04 [1.03-1.05]) and oil (g; 1.04 [1.03-1.06]) intake, log (GDP per capita) (2.18 [2.14-2.21]) and literacy rates (1.01 [1.01-1.01]) were positively and independently associated with higher odds of overweight. The impact of state characteristics on prevalence of overweight/obesity decreased across rounds.

Conclusions: While the role of state economic and nutritional context as a driver of overweight has waned over time, contextual indicators remain salient correlates of an individual adult’s likelihood of being overweight. Findings suggest that rising overweight in India must be understood and addressed from a socio-ecological lens that considers context alongside individual risks.

Funding Sources: None