It also makes me wonder that if some cancers are rare enough to average to 0 months (rounded down) that could still work out to a 1/500 chance of living to 80 vs 40. Long odds it matters, but big difference if it does.
They were testing for some of the most frequent cancers (breast, colorectal, lung cancer (in smokers), prostate), and measuring differences in outcomes down to the day.
