Even if you are drug-free now, you could have a procedure in couple years and become addicted to opioids during recovery. Many people who are suffering from the opioid epidemic were prescribed these drugs.
But that's even more reason to split out the stats. It would be very good to know that the average life expectancy for opioid users is much lower before I consider taking them.
Life expectancy if you get the surgery and medicate with pot: 80 years.
Life expectancy if you get the surgery and use the opioids: 60 years.
That seems immeasurably more useful than a national average alone which only can be used for political talking points.
That's not a matter of splitting out stats, because the information gathered doesn't support that. That kind of research does get done, but it's expensive, time consuming, and only really possible when alternatives have been fairly widely applied for enough years to have good data to support meaningful analysis of expected lifetime (and may still, even then, be misleading I'd treatment protols for either drug have evolved significantly over the period.)
So you won't find them for all the options of interest, and you emphatically won't find them for emerging options (including pot, which hasn't been even quasi-legal long enough to support good data.)
Sure, though (and this is usually reflected in research that it includes it now), lifespan alone for alternatives isn't the only or necessarily most important piece of comparative information; relative risks of other conditions even if they are non-fatal and other quality-of-life concerns are important, too. For aggregate purposes when we aren't comparing path risks but looking at overall status, the patallel is merely statistics on the incidence of various conditions and effects with quality-of-life impacts, which is currently done separately from life expectancy.
Life expectancy if you get the surgery and medicate with pot: 80 years.
Life expectancy if you get the surgery and use the opioids: 60 years.
That seems immeasurably more useful than a national average alone which only can be used for political talking points.