Human bodies are really complicated. Doctors can always be better informed, but in the end they're still playing the statistics game, and we have pretty poor statistics on this stuff (especially with diet, which is almost always self-reported in any large study). For most people, beyond smoking, weight management, and maybe exercise and sleep, we've found basically nothing that will move your risk factors more than a handful of percentage points in either direction. You can easily negate those benefits by having a stressful few months in your job or in your marriage, and meanwhile you very well might die of cancer in the few years you bought yourself.

We're not going to "solve" the problem until we have a ton more data (maybe sensors in everyone along with the understanding of the markers to look for) so we can fit you to an optimal path based on a lot of observed variables. But even then we're just going to be playing percentages off each other, and contrary to the life-extender crowd out there, it may turn out that there is no 110-year optimal path for most of us, at least until we start reengineering the body at a more fundamental level.