Introduction to the Environmental Burden of Disease concept – history, methods and selected results

Abstract

Abstract Background The Environmental Burden of Disease (EBD)-concept was introduced by the World Health Organization to quantify the impacts of environmental risk factors on population health in a comprehensive and comparable way. Since the first Global Burden of Disease (GBD)-study the EBD-methodology was used in numerous global, regional and national assessments. Methods In the EBD-concept data from different sources are joined using a standardized framework. To calculate the EBD several steps need to be taken into account. First, the exposure of the selected population towards an environmental risk factor is estimated. In a second step, exposure data and information from an exposure-response-function are combined by using the population attributable fraction formula. In a last step, the attributable fraction (percentage) is multiplied by the disease burden resulting from a selected health outcome to estimate the share of disease burden attributable to the environmental risk factor. Results Since the first GBD-study several stakeholders have used the EBD-concept to estimate the disease burden attributable to environmental risk factors. This lead to an increased number of available data sets, which - due to the varying assumptions used in the models - are not ad hoc comparable among each other. Generally, the results of EBD-assessments are presented using disability-adjusted life years (DALY) as the core measure of burden of disease assessments. However, e. g. due to lack of data, assessments also focus on the number of attributable deaths, illnesses or single components of the DALY such as the years of life lost due to premature mortality (YLL). Conclusions The EBD-concept allows to compare the impact of environmental risk factors on population health. Though the general concept is widely standardized, assumptions on model parameters can lead to varying results. Therefore, when communicating EBD-results it is necessary to be very transparent about data and model inputs.