locations. These observations lead to two directions for further analysis. First, what is the explanation for the declines driven by other factors? Some of this effect might be social policy working through various causal channels, and some is likely due to improvements in access to high-quality health care. This is particularly true for conditions such as selected cancers, ischaemic heart disease, cerebrovascular disease, chronic kidney diseases, HIV/AIDS, tuberculosis, and maternal mortality, for which health care is known to have large effects. Second, in view of the enormous potential of risk reduction to change health outcomes as documented in this and many other studies, why has progress on many risks been comparatively slow? For example, even though global tobacco consumption is declining in terms of rates, the pace of decline has been remarkably slow on average, despite more than 50 years of good evidence on the harms of tobacco. The relatively poor track record for global risk reduction might in part reflect the low rate of investment in risk reduction compared with curative health care. It might also reflect the continuing challenge of changing many risky behaviours. Relatively little funding for research on changing behaviours compared with new diagnostics and therapeutics
