Developing Methods for Estimating the Impacts of Climate Change on Malnutrition

Abstract

ISEE-0403 Background: Climate change is likely to influence future crop productivity globally, with greatest vulnerability in developing nations. To date, assessments of the climate-change attributable impacts on “health” have focused on food availability and access to estimate ‘millions at risk of hunger' and have neglected malnutrition as an outcome. We developed a new model for estimating climate change-attributable burdens of malnutrition for the QUEST-GSI project (global-scale impacts of climate change: an integrated multi-sectoral assessment), part of the NERC-funded QUEST research programme on global change. Methods: GLAM global crop model (Walker Institute) estimates of future cereals yields were used to drive a food availability model (Southampton University) to provide estimates of future undernourishment at the country level. As malnutrition has multiple causes, stunting in children aged under 5 was estimated as a function of undernourishment (food-causes) and a GDP-derived development indicator (non-food causes) of stunting. A Bayesian approach for estimating the parameters of the health impact model was used by combining prior expert knowledge of the parameter values with current health outcome data observations. Results: The model outcome measures are proportion of population with stunting for five world regions (Sub-Saharan Africa and South Asia) under 4 SRES emissions scenarios. The proportion stunted was converted to DALYs following methods developed for the WHO global burden of disease assessment. Malnutrition burdens were estimated under three scenarios of improvements in agricultural technology. The main sources of uncertainty were described and quantified, where possible. Conclusion: Although the models indicate the burden of malnutrition is likely to increase under climate change, such estimates are very sensitive to model assumptions, and incorporate a large range of uncertainty from both the climate and socio-economic inputs and assumptions.