The obesity wars and the education of a researcher: A personal account (2021) | Hacker News Reader

The obesity wars and the education of a researcher: A personal account (2021)

sciencedirect.com

21 pointsblending4 comments

> The 3 most recent nationally representative data sets from NHANES.

See also:

> "The NHANES dietary data are physiologically implausible and inadmissible as scientific evidence" - https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5573021/

There seems to be an entire industry around extracting nutrition research from these deeply flawed self-reported surveys ... on the grounds that it's the best data we have. Which is like looking for your car keys under the street light because that's where the light is.

So while I sympathize with this story and think it shines light on the process of institutional character assassination, it's about replacing one set of non-probative data with another.

Did the paper here use the calorie data from the nhanes dataset? My (perhaps incorrect) understanding was that they used bmi which was measured rather than self reported?

I presume you relate this as a critique of self-reported data rather than a critique of the paper in TFA?

The article specifically mentions how all the datasets their study used were NOT self reported, but measured. It's one of the biggest arguments for why their study was more trustworthy and why the assumptions/accusations were spurious and baseless

I'm having trouble finding the supposed issue with the meta analysis[0] cited in this article which finds that by excluding smokers there actualy is a corelation between obesity and mortality.

The auhtor claims this is wrong because "238 of the 239 data sets used by the GBMC were previously known to the senior author" but by the authors own admition "By repeating the MEDLINE search, we identified six eligible data sets that were not included by the GBMC, with two shown as unable to obtain data and the remainder apparently overlooked" and these papers were seemingly published after the meta analyists had created their dataset.

That meta-alalysis, by the author's admition, a "highly visible publication" which was "authored by a writing committee of 61 distinguished epidemiologists and included over 500 collaborators" has over 2,000 citations, yet the author's three responses to it in the OP are self citations. Certainly others should find the mistake?

Additionaly, not considered by the OP is healthcare cushioning the danger of obesity:

>People with obesity experienced a statistically significant twofold increase in average direct healthcare costs per year (EUR 5,934), compared with controls (EUR 2,788) and had statistically significantly higher indirect costs compared to controls[1]

If costs were kept the same between obese and normal weight, mortality would almost certainly increase even further.

[0] https://www.sciencedirect.com/science/article/pii/S014067361...

[1] https://www.tandfonline.com/doi/full/10.1080/14737167.2022.2...

Reply on news.ycombinator.com