Comparative Risk Assessment directly applied to GBD data

Previous versions of ITHIM have directly applied the PIF to deaths, ylls, ylds and hence dalys.

Applying them to YLDs is problematic as it does not reflect the longer term change in disease burden because not all the ylls lived will be in full health. It could be seen as indicative of where the benefits occur but not added to ylls to give change in dalys @nmaizlish. This is particularly important if we are working with health service costs Thus I see options are

• we only include ylls
• we calculate ylds but dont add them to ylls (prenting them for illustrative purposes)
• we adjust ylls to reflect likely health status (perhaps using average ylds per age group- as I understand it currently GBD gives incidence ylls and prevalence ylds so i think this should be doable)
• we adopt some other approach (mulitplier based on Joe Spadaro's approach but this will come with it's own challenges)
• we move whole scale to PMSLT (this would require user to produce the incidence, prevalence, case fatality data for the PMSLT)see https://github.com/ITHIM/ITHIM-R/issues/31

Applying PIFs to YLDs and YLLs also assumes that we can use the GBD ideal life expectancy