About 46 million Americans smoke, or about 15%. I’m assuming that proportion is probably similar in other developed countries. So if smokers died at the same rate as non-smokers, about 15% of deaths would be smokers’ deaths. If one in eight, or 12.5% of all American deaths will be from smoking, then that doesn’t leave much else for smokers to die of - only 2.5 out of 15 smokers’ deaths could be from other causes. That seems wildly improbable.
A few years ago, I remember reading that “alcohol related fatalities” included the deaths of all persons who died in an auto accident if anybody in any involved car had been drinking, even if that person was just a passenger in a vehicle not responsible for the accident. It inflated the tragic dimensions of the drunk driving problem, presumably generating more political support and aiding in fund-raising.
I wonder if the same thing is going on with smoking death statistics. The only way this 1 in 8 figure could make sense is if all deaths of smokers who died from diseases in which there is a known association between smoking and higher incidence of the disease (for instance, cardiovascular disease) were being labeled as smoking related.
This would be a huge distortion. If, say, 20 out of a 100 non-smokers died from cardiovascular disease (CVD) and 25 out of 100 smokers died of it, the accurate expression of the smoking related deaths would be 5 - the excess deaths among smokers from the disease over the expected number if they had been non-smokers. If instead you claimed that there were 25 smoking related deaths, you could argue that smoking is known to cause CVD and 25 smokers died of CVD, so there were 25 smoking related deaths, but the real effect of smoking on deaths would have been inflated by 500%.
Since billions of dollars of governmental research, activism and bureaucracy funding now follow along with these studies, it’s hard not to be skeptical.