News flash, Time magazine: EVERYBODY IS GOING TO DIE. NO ONE has a lower risk of death than anyone else.
I see this kind of story all the time, and the title is usually the same. “These foods reduce your risk of death.” “These habits reduce your risk of death.” “Living in these cities will reduce your risk of death.” Um, no… it most definitely WILL NOT.
Thanks for reading. I can’t figure out how to contact the author of the story, but I had to vent!
Well yeah, “We are (all) going to DIE” (as Indiana Jones once said), and it’s kind of silly to suggest any intervention is going to change that inevitability. Quantify it as “premature death” (and define that rigorously) and you may have something, but it doesn’t make for a quick and sexy headline.
The linked article describes a 45% lower risk of death from cardiovascular causes, but then I wonder what these people are going to wind up dying from - cancer? infection? being hit by buses in the crosswalk?
Gimme good evidence that an intervention results in a boost in productive longevity and I’m up for it.*
I once got into an argument online with a promoter of a resveratrol supplement who suggested strongly that his product would produce 'super-longevity". That sounds nice, but an extra 20 years spent drooling in an extended care facility isn’t all that appealing.
If Alzheimer’s runs in your family, go ahead and live a sedentary life and eat a lot of bacon! If your grandmother was doing crossword puzzles at 99 - you could take up running!
I think it’s just assumed that an intelligent reader will understand the context. Regular eating of vegetables won’t protect you from death by falling pianos, but it will (presumably) reduce your risk of death from certain causes that might include heart disease or bowel cancer.
Do we really expect or require medical precision from magazine articles? If the article on the benefits of running was scientifically precise, would you then decide to take up running?
Meh. I’ve worked on longevity experiments*, and statistics are all about risk of death in a given time period. For instance, the standard method of analyzing a survival data set is with the Kaplan-Meier estimator, which produces an estimate of the hazard function. This explicitly gives the risk of death at any given instant or window in time. Of course, the hazard function will tend to increase with time for most critters**, and the cumulative hazard will approach and eventually reach 1.
Anyways, it is entirely possible to have a data set where the hazard function of runners is always less than the hazard function of non-runners. That is, for all specific windows of time in the study, it’s possible that risk of death of non-runners is always higher than the risk of death for runners***. This is often simplified as the “hazard ratio” which is what is reported in the OP’s link.
Therefore, it’s not inaccurate to say “risk of death” is lower for runners, since the statement can be true for all specified time periods that anyone is interested in. It is hardly interesting to report that the cumulative risk at age 130 is exactly 1 for both runners and non-runners.
*of tiny model organisms. 47405 of 'em to be exact.
**But not all! There are some organisms that do not seem to have a set life span. Their risk of death, from disease and injury, is approximately constant for all time.
***That’s the “proportional hazards” assumption that’s baked into many of the statistical approaches. It’s not always true, but it’s often a good-enough assumption, and you can test for proportionality-of-hazards and use other methods if it’s not true.
Poll finds Obama is the worst president since World War II
The Poll actually could “find” only that Obama was considered the worst president since WWII, a poll can’t determine whether he actually was the worst or not.
Before he started running, Jim Fixx:
Was 240 lbs, lost 60 lbs before his book came out.
Smoked 2 packs a day.
His father had a heart attack at the age of 35, died from his second at 43
At autopsy, Fixx was found to have severe atherosclerosis, e arteries blocked over 70&, one blocked at 99%.
He was found to have had several small heart attacks before his fatal one.
Double meh. Understood by all but the densest that it means over some defined period of time.
You really are not familiar with the huge body of data about that? I am surprised. Indeed the bigger impact of healthy lifestyle choices is on years of productive non-disabled life than on years living.
Actually the NYT version of this was to headline another aspect:
The fact that just a little bit of exercise produced significant benefits was their emphasis.
The fact is pretty well established that the biggest benefit from exercise is gained in the move from inactive to any exercise. The open question that this adds to is trying to determine the dose response curve. How much exercise is how much better? And does it reach a point where more is worse with “with U-shaped mortality curves for distance, speed, and frequency”?
