Particulate Matter Related Disease Burden – A Plethora of Estimates – Which Ones to Accept and Which Ones to Trust?

Abstract

Background/Aim: Environmental burden of disease estimates are increasingly used to inform society and politics about harmful effects arising from ambient particulate matter (PM2.5) pollution. Institutes and international bodies, such as WHO, EEA and IHME, provide estimates on the number of premature deaths and years of life lost, related to air pollution, which are not in agreement. Using estimates for Germany, the aim of the study was to systematically assess the differences in model assumptions leading to diverging estimates. Methods: Necessary information was extracted from publications and methodological appendices. The assumptions were systematically compiled and evaluated. As reference years for the estimates were not matching, we selected estimates from 2010-2015. The evaluation focused on exposure estimates, health data, exposure-response-functions and the theoretical minimum risk exposure level (TMREL). Results: Comparisons for Germany indicate considerable differences in the number of premature deaths attributable to air pollution. For 2012, WHO and EEA report 26,160 and 59,500 deaths, respectively. IHME’s estimates for 2010 and 2015 are in between those with 40,993 and 43,117 deaths, respectively. The most salient difference is the choice of TMREL. EEA estimates are based on a counterfactual level of 0µg/m³, assuming no safe level for PM2.5-exposure, and thus are considerably higher than the other estimates, that use a uniform distribution for the TMREL (WHO: 5.9-8.7µg/m³; IHME: 2.4-5.9µg/m³). Conclusions: Policy makers often do not have the time and expertise to consider full details of methodologies used for estimating disease burdens. IHME’s estimates provide a high level of transparency, as they follow the GATHER guidelines. The estimates from the WHO and EEA do not provide a comparable level of transparency and EEA does not even provide uncertainty intervals around their estimates. However, a formal validation of environmental burden of disease estimates is not possible, because e.g. death statistics cannot capture a person’s risk-profile.