Cabbages, cucumbers and coronavirus

[Bird on a Fire]

Here's a good one.

"Preprint article" (ie, an unreviewed pdf) on medRxiv postulating a link between different vegetables and covid mortality - based on national per capita consumption.

Background Many foods have an antioxidant activity, and nutrition may mitigate COVID-19. To test the potential role of vegetables in COVID-19 mortality in Europe, we performed an ecological study. Methods The European Food Safety Authority (EFSA) Comprehensive European Food Consumption Database was used to study the country consumption of Brassica vegetables (broccoli, cauliflower, head cabbage (white, red and savoy cabbage), leafy brassica) and to compare them with spinach, cucumber, courgette, lettuce and tomato. The COVID-19 mortality per number of inhabitants was obtained from the Johns Hopkins Coronavirus Resource Center. EuroStat data were used for potential confounders at the country level including Gross Domestic Product (GDP) (2019), population density (2018), percentage of people over 64 years (2019), unemployment rate (2019) and percentage of obesity (2014, to avoid missing values). Mortality counts were analyzed with quasi-Poisson regression models to model the death rate while accounting for over-dispersion. Results Of all the variables considered, including confounders, only head cabbage and cucumber reached statistical significance with the COVID-19 death rate per country. For each g/day increase in the average national consumption of some of the vegetables (head cabbage and cucumber), the mortality risk for COVID-19 decreased by a factor of 11, down to 13.6 %. Lettuce consumption increased COVID-19 mortality. The adjustment did not change the point estimate and the results were still significant. Discussion The negative ecological association between COVID-19 mortality and the consumption of cabbage and cucumber supports the a priori hypothesis previously reported. The hypothesis needs to be tested in individual studies performed in countries where the consumption of vegetables is common.

A third of their statistical tests come out significant. I wonder how many vegetables they started off with before selecting them for the paper?

[Martin Y]

Did they control for green jelly bean consumption?

[Gfamily]

Did they control for green jelly bean consumption?

That man has spawned so many memes. There will be PhDs about his impact in the future.

[TheScientificHippy]

I propose that the shape of the veg and it's ability to be efficiently washed could be a contributor to transmission of Covid 19.

[bob sterman]

WTAF? "Lettuce consumption increased COVID-19 mortality"?

Why on earth would anyone think national consumption statistics tell you much about what the demographic groups infected and dying from COVID-19 actually eat?

In any case - the different national dietary surveys don't even cover the same age groups!!! E.g. the surveys in Cyrpus and Spain go up to 75, Romania goes up to 92, and Beligum only goes up to 64.

Why am I bothering to look this up? Aaaargh - what a pile of total......

[Sciolus]

Tell me they aren't doing correlations on national-level stats? Why yes, they even say they are in the title.

How did they find twelve people to put their names to this? Maybe they promised that it would end up in the textbooks...

[sTeamTraen]

Tell me they aren't doing correlations on national-level stats? Why yes, they even say they are in the title.

I don't think there's anything wrong with that, as long as you don't use it to draw conclusions about individual behaviour. (It can actually make sense when evaluating something like government policies.)

Of course, a lot of researchers who go on these types of fishing expedition really want the reader of their paper (or, at least, the reader of the news outlet that picks it up) to make such inferences, but I think the authors have done a better-than-typical job of this when they point out the possible "ecological fallacy" issues at the bottom of page 5.