Social costs of morbidity impacts of air pollution.

Outdoor air pollution is a major determinant of health worldwide. The greatest public health effects are from increased mortality in adults. However, both PM and O3 also cause a wide range of other, less serious, health outcomes; and there are effects on mortality and morbidity of other pollutants also, e.g. nitrogen dioxide (NO2) and sulphur dioxide (SO2). These adverse health effects have economic consequences; OECD (2014) suggests that the social costs of the health impact of outdoor air pollution in OECD countries, China and India was approximately USD 1.7 trillion and USD 1.9 trillion, respectively, in 2010. However, the study highlights that though the social costs of premature mortality account for the majority of these totals, the social costs of morbidity remain poorly estimated. The objective of this paper is to inform the development of improved estimates of the social costs of human morbidity impacts resulting from outdoor air pollution in two components; namely to develop a core set of pollutant-health end-points to be covered when estimating the costs of morbidity, and to review current estimates of the cost of morbidity from air pollution. The paper recommends a core set of pollutant-health (morbidity) combinations, for application in OECD countries, China and India, that meets the criteria of being unbiased (i.e. not systematically underestimating or over-estimating the effects), credible (i.e. based on recognised expert reviews) and implementable (i.e. for which the necessary input data, including concentration-response functions (CRFs) and background rates of morbidity) are available or can be estimated. The review has identified five pollutant-health pairs for consideration. These are: 1. Respiratory Hospital Admissions (RHA) and Cardiovascular Hospital Admissions (CVHA) in relation to PM and to ozone. 2. Restricted Activity Days (RADs) and associated Work Loss Days (WLDs) in relation to PM and/or ozone. 3. Chronic Bronchitis in adults in relation to PM. 4. Acute Bronchitis in children aged 6-18 years, defined as “bronchitis in the past 12 months” resulting from PM10; and 5. Acute Lower Respiratory Illness (ALRI) in children aged below 5 Whilst the empirical evidence for these pollutant-health outcomes remains limited, they are judged to be sufficiently robust to be considered in policy analysis. The paper presents a list of unit values for the relevant health end-points. The unit values are comprised of three broad components: “resource costs” (which includes avertive expenditures, e.g. relocation to area of lower air pollution, staying inside, etc., and mitigating expenditures, e.g. the direct medical and non-medical costs associated with treatment for the health impact), “opportunity costs” (which includes costs related to loss of productivity and/or leisure time due to the health impact) as well as “disutility costs” (which includes pain, suffering, discomfort and anxiety linked to the illness). A pragmatic approach to calculating air pollution-related morbidity effects might be to assume that they are a (near) constant fraction of the total health impact. On preliminary investigation the ratio of mortality to morbidity effects in recent EU and US evaluations is very different. However, the apparent differences are strongly influenced by different approaches to valuation of mortality; the differences are much less when EU mortality valuations are based on VSL rather than VOLYs and, under some valuations, morbidity costs become less than 10% of the mortality ones. Given that the evidence points to underestimation of morbidity impacts in quantitative HIA, marking up mortality costs by 10% to 15% may give a broadly correct estimate, at low cost. However, since there are many real differences between countries and regions with regard to pollutant mix, valuation of health treatment costs, productivity losses and pain/suffering, as well as cultural factors, this approach should be complemented by the bottom-up approach that compiles CRFs and unit values for the given context, whenever possible.