Dealing with uncertainties in environmental burden of disease assessment

In environmental health research, focus has shifted from relatively simple to more complex issues. Empirical single agent – single effect studies have been supplemented by research on risks of complex environmental exposures in varying economic, cultural and political settings. Environmental health impact assessment has become a valuable tool for decision support. These types of assessments increasingly use so-called environmental burden of disease (eBoD) measures to express health impacts. The eBoD can be viewed as the gap – caused by environmental factors – between current health status and an alternative situation in which environmental exposures are reduced or eliminated. Burden of disease estimates enable comparison of divergent environmental health problems. This in turn enables policy makers to set priorities. However, scientists often have to make many assumptions when assessing the eBoD. Knowledge and data are often incomplete, and diverging perceptions exist about what the most important aspects of a problem are. Assessments are often highly interdisciplinary, complex and multifaceted, and the uncertainty about results can be significant [1]. This may affect decision making based on these assessments.A 2005 comparison of 17 eBoD studies published between 1996 and 2005 (internal RIVM/MNP publication by Knol et al.) showed that there are significant differences between eBoD estimates that concern – at first sight – similar issues. Smith et al. [2], for example, estimate the fraction of the total global disease burden attributable to the environment to be 25–33%, whereas Melse and de Hollander [2,3] estimate this to be 7.5 to 11% (for OECD countries only: 2–5%). Such differences can sometimes not be fully explained by reading the assessment reports. Methods, assumptions and input data are often insufficiently explained, which hampers interpretation and comparability of results. Fox-Rushby and Hanson [4] show that 9 out of 16 papers on burden of disease publis