To examine whether changes in food prices are associated with changes in obesity prevalence among women in developing countries, and assess effect modification by individual socioeconomic status (SES).
Longitudinal study of country-level food price inflation temporally and geographically linked to anthropometric data on non-pregnant adult women (n = 295,984) in 31 low-income and middle-income countries over the 2000–2014 time period, using separate multivariable multilevel growth models of five SES indicators. Post-estimation analysis computed the relationship between food price inflation and predicted mean probabilities of being obese, by SES.
Rising food price inflation was strongly associated with women’s obesity prevalence, and SES consistently modified the relationship. Regardless of indicator used, higher food price inflation was positively associated with obesity among women in top SES categories, but was flat or negative among women in low SES categories, averaging over time. The SES differences were widest across educational strata and were most pronounced when food price inflation was highest. Overall, for every 1-unit increase in food price inflation, predicted mean obesity prevalence was between 0.02 and 0.06 percentage points greater in women of high SES compared to low SES women.
There is a strong link between food price inflation and obesity in adult women in developing countries which is clearly modified by individuals’ SES. Greater food price inflation was associated with greater obesity prevalence only among women in higher SES groups, who may be net food buyers most at risk of obesity in low-income and middle-income countries.
This is the first longitudinal study specifically examining the impact of rising food prices on health across different social groups and countries.
Our study uses a robust multilevel model for change, with national food price inflation data linked to individual-level anthropometric and socioeconomic data between 2000 and 2014.
We found food price inflation was positively linked to obesity in women of high SES, but no associations in other groups, and potential further modification by country income group.
Study results are threatened by residual confounding from unavailable country and individual factors.
Access optionsAccess options
Subscribe to Journal
Get full journal access for 1 year
only $60.33 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
The authors appreciate statistical support and expertise from the UCLA IDRE Statistical Consulting Group. We also thank the anonymous reviewers for this feedback to improve the manuscript.
Conceived and designed the experiments: AIC, AD, RS, NP. Performed the experiments: AIC. Analyzed the data: AIC, AD, RS, NP. Wrote the first draft: AIC. Contributed to the final paper: AIC, AD, RS, NP. All authors approved the final version.
The authors received no specific funding for this project. AIC acknowledges previous salary support from the CHIR Postdoctoral Fellowship Award (MFE-135520).