These numbers are real and it is important that as economists we try to relate economics to the real world. I have used the WTO quote before and will probably do so again.
This new working paper by Grasso, Maera, Chiabai and Markandya provides a good survey of the literature.
The Health Effects of Climate Change: A Survey of Recent Quantitative Research [PDF]
In recent years there has been a large scientific and public debate on climate change and its direct as well as indirect effects on human health. According to World Health Organization (WHO, 2006), some 2.5 million people die every year from non-infectious diseases directly attributable to environmental factors such as air pollution, stressful conditions in the workplace, exposure to chemicals such as lead, and exposure to environmental tobacco smoke. Changes in climatic conditions and climate variability can also affect human health both directly and indirectly, via changes in biological and ecological processes that influence the transmission of several infectious diseases (WHO, 2003). In the past fifteen years a large amount of research on the effects of climate changes on human health has addressed two fundamental questions (WHO, 2003). First, can historical data be of some help in revealing how short-run or long-run climate variations affect the occurrence of infectious diseases? Second, is it possible to build more accurate statistical models which are capable of predicting the future effects of different climate conditions on the transmissibility of particularly dangerous infectious diseases? The primary goal of this paper is to review the most relevant contributions which have directly tackled those questions, both with respect to the effects of climate changes on the diffusion of non-infectious and infectious diseases. Specific attention will be drawn on the methodological aspects of each study, which will be classified according to the type of statistical model considered. Additional aspects such as characteristics of the dependent and independent variables, number and type of countries investigated, data frequency, temporal period spanned by the analysis, and robustness of the empirical findings are examined.