Contribution of economic and nutritional context to overweight/obesity dynamics in Indian women from 1998 to 2016: a multilevel analysis of national survey data

Abstract

Background: Overweight/obesity increased dramatically among Indian women since 2000. We evaluated the independent contributions of economic and nutrition context to the changing distribution of overweight/obesity among women from 1998 to 2016 across India.

Methods: Individual-level data from 473 912 ever married Indian women aged 18-49 in the National Family Health Surveys (1998-1999, 2005-2006, 2015-2016) were merged with year-matched state-level economic and nutrition context indicators. Cross-classified generalised linear mixed models were estimated to quantify associations of contextual characteristics with overweight/obesity (body mass index ≥25 kg/m²) across survey rounds.

Results: Between 1998 and 2016, age-standardised prevalence of overweight/obesity increased from 13.9% to 27.5% nationally at an annual growth rate of 0.8%. After accounting for a woman's age, parity and social class, the adjusted OR (aOR) for overweight/obesity was 2.02 times higher for every unit of state log per capita gross domestic product (GDP) (95% credible interval (CrI) 2.00 to 2.03). Yet, the association of state GDP with overweight/obesity generally decreased over survey round. Women in states with higher per capita daily oil (aOR 1.02 per gram; 95% CrI 1.01 to 1.03) and sugar (aOR 1.05 per gram; 95% CrI 1.04 to 1.05) consumption were more likely to be overweight/obese, while women in states with higher cereal consumption were less likely to be overweight/obese (aOR 0.93 per 10 gram; 95% CrI 0.93 to 0.93).

Conclusions: Indicators of state economic development and nutrition transition were independently associated with a woman's likelihood of being overweight/obese. The impact of state wealth waned over survey round, suggesting that risks for overweight/obesity may be increasingly shaped by individual factors as economic development expands in India.

Keywords: epidemiology; nutrition; statistics & research methods.

Figure 1

Prevalence of overweight/obesity by economic and nutrition indicators. X-axis represents state wise (A) log (per capita GDP in rupees) (B) literacy rate (in %) (C) employed adults engaged in a sedentary occupation (in %) (D) median per capita oil intake (in grams) (E) median per capita sugar intake (in grams) (F) median per capita cereal intake (in grams). each point represents a state value for different rounds (white squares—NFHS-2, black circles—NFHS-3 and plus sign—NFHS-4). Sedentary occupation, and nutritional intakes were computed using survey estimation procedures as per respective NSSO rounds. The values corresponding to r_NFHS-2, r_NFHS-3 and r_NFHS-4 represent the Pearson’s correlation coefficient between the state economic and nutritional indicators and overweight/obesity prevalence by NFHS round. GDP, gross domestic product; NFHS, National Family Health Survey; NSSO, National Sample Survey Office.

Figure 2

Associations of state-level characteristics with overweight/obesity in Indian women in 1998–1999, 2005–20006, 2015–2016. Separate models were fit for each state-level variable. All values are aOR (95% CrI) estimated from logistic cross-classified random intercept model adjusted for woman’s age category, educational attainment, wealth index, parity and caste. aOR, adjusted OR, CrI, credible interval; GDP, gross domestic product.