Comment by mannykannot
1 year ago
The one thing that this paper does is demolish the claim that people living in these blue regions are living much longer than average.
1 year ago
The one thing that this paper does is demolish the claim that people living in these blue regions are living much longer than average.
I think a lot of commenters either didn't read the abstract or assumed from its tone that it was supportive of the idea of blue zones.
Yes, I happen to be interested in this topic and had previously read this paper. So it's interesting to see how many of the commentors (even on HN) didn't even read the first few paragraphs of the paper or the article. I guess I must do that with articles I'm not previously familiar with. :-)
keeping in mind that they live in countries with higher life expectancy than most countries anyway. Indeed they may not even be outliers within those countries.
Except, it misses the point and doesn't really do that while being persuasive.
According to the Blue Zone researchers, some of the Blue Zones are disappearing because the generations that came after the oldest live differently and much shorter. By differently, their eating, body movement, and other characteristics are different. Looking at the whole population doesn't segment for differences between generation. So, nuance is lost.
In some areas, like the Blue Zone in the US other research is finding the people who live there are healthier than the surrounding populations. Then you have to ask, what area do you average over for your measurement and statistics?
> According to the Blue Zone researchers, some of the Blue Zones are disappearing because the generations that came after the oldest live differently and much shorter.
Of course they would say that. But if these zones are simultaneously recording births better and reducing welfare fraud, and if 80%+ of the centenarians either had no birth certificate or were actually dead, I'm going to need more than "but they're also changing lifestyles" as an explanation.
We're talking here about unusually long life, not just "he's still going strong at 85" long. No one here is arguing that people who are active and eat right don't have a longer healthspan, but that's a concept that's provable without the so-called Blue Zones.
I listened to an interview with the author during the week, in short as soon as you start getting reliable recording of births and clamp down on old age welfare fraud the phenomenon disappears.
"Then you have to ask, what area do you average over for your measurement and statistics?"
This is a big thing that I didn't seen in the paper this article is based on. It seemed like the author was comparing adjusted numbers from the blue zone with unadjusted numbers from non-blue zones. Without comprehensive investigation of error rates and even different error mechanisms by locale, it seems like a poor comparison to make. Comparing life expectancies is better than comparing outlier centarian numbers, but you are right that it depends on what other areas we are using as the baseline or average (and I take it a step farther by saying it depends on what error adjustments need to be made to both data sets).
The whole blue zone idea is a bit misapplied though. These population studies find new variables to look at. Then you have targeted studies to investigate thos variables. Discrediting the centarian numbers doesn't discredit the findings on stuff like a mederteranian diet having better health outcomes than the standard western diet, etc.
Kinda depends on how much you value inductive vs. deductive reasoning, but the authors make the deductive case that:
- There's strong incentives to misreport in these areas (the compelling example from Sardinia was that the person is alive for the purposes of pension fraud, but really dead)
- People who are incentivized to report people being older than they are will do so
And the inductive case relies on data, which is presumed to be totally flawed because of the misaligned incentives.
It is very much to the point, addressing the specific claims and methodology of a specific (and apparently somewhat influential) study.
You are, of course, free to speculate that there are other issues related to longevity than those considered in the study in question, but even if these suppositions are correct, in no way would this justify saying the paper being discussed here misses the point. The point is that the blue zones study is too flawed to support any definite position, which includes both its own conclusions and the more nuanced issues about which you speculate.
> It is very much to the point, addressing the specific claims and methodology of a specific (and apparently somewhat influential) study.
Except, the author doesn't discredit specific claims of the Blue Zones. For example, the Blue Zones might take an area and state there is a higher rate of centurions who are healthy and capable. The counter to that might be the average life span in the region isn't an outlier. In one case you're looking at a targeted subgroup and the other your looking at the population as a whole. One observations doesn't disprove another.
This is just one example. It's why I call the work misleading.
1 reply →
Or now better records are kept the incorrect data dies off
There was a study I read about in Barrons that was noting that places a western diet goes the health care costs then start going up. Other studies have found that a western diet leads to more unhealthy outcomes (increased disease and earlier death).
I state this to point out that there are other variables at plan than just changes in record keeping.
3 replies →
> the generations that came after the oldest live differently and much shorter.
This presupposes that the previous generation was in fact living longer, which the linked study showed is not the case at all.
Not exactly. It establishes that error rates are high in those areas, demolishing the centarian numbers. It doesn't give much investigation into the averages at all. Where it does, it seems to compare adjusted numbers of one data set with unadjusted numbers of another. If you really want to get into the averages, you'd have to determine error rates and adjustments for each specific area, probably by jurisdiction or record keeper, and then compare them. The problem is, nobody is going through that process for the entire world so we just use the face value numbers until we want disprove a specific area and then compare the adjusted numbers against unadjusted numbers. The data is too massive to rigorously investigate. But this whole effort is moot. What tangible benefit comes from disproving blue zone data? These population level studies aren't meant to provide answers. They're meant to provide new variables. Each of the blue zone longevity recommendations have their own studies to either prove (food stuff) or disprove (drinking wine daily) them.
So yeah, it's great the errors in the data have been called out it's a bit surprising that the author interviewed is so angry in the article. I guess it's fitting that he got the Ig nobel, since this correction doesn't have any applicable impact to end result, which were additonal studies investigating the individual suggestions/variables, such as specific dietary practices.
If the error rates are high, there is no reliable signal that these areas are different, so how the hell can looking at their "new variables" help?
Go look up the studies that came out of it.
It would be different if these were new studies, but this is all in the past. This new finding of unreliability doesn't have any impact, hence the Ig nobel instead of the real nobel.
9 replies →
There are basically zero studies which prove anything about particular foodstuffs. It's all observational studies with small effect sizes and multiple uncontrolled confounding variables: junk science.
We know we need certain essential nutrients to prevent deficiencies, an energy intake surplus causes weight gain, and a few substances like trans fat are problematic. Beyond that, people seem to be making claims and recommendations not backed by hard evidence and frequently confuse correlation with causation.
We aren't talking about unequivocal proof. If someone asks what they can do to increase longevity, it's perfectly reasonable to tell them about studies that show strong correlations and mention the way the confounding factors play a role.
You might be interested to look into some of the twin studies that put twins on similar exercise regimens and differening diets. They seem to be the strongest evidence possible for this sort of thing. Hardly what I would call junk science.
5 replies →
That's my point.
I think a single "Welp" at the end of the comment is not communicating that clearly enough.