Serious Data Issues at Play in a Randomized Trial of Vitamin D for Covid-19

A major trial of vitamin D treatment for Covid-19 has a number of flaws that make the results unreliable

Gideon M-K; Health Nerd


Pictured: Probably fish oil, but that’s what you get when you search “supplement” on stock photo websites. Photo by Leohoho on Unsplash

One of the bigger questions throughout the pandemic has been whether vitamin D supplements are a useful tool in preventing and treating Covid-19. Since early 2020, people have been constantly wondering whether taking some extra vitamin D would help them fight off the dreaded spicy cough.

Stock photos for “spicy cough”, on the other hand, deliver amazingly well. Photo by Fernando Andrade on Unsplash

There have, at this point, been dozens of studies published attempting to answer the question of whether vitamin D works for Covid-19. Of these, one of the most influential has been a medium-sized randomized controlled trial from a group in Hyderabad. The study found that there was a massive improvement in people given vitamin D compared to a control group in terms of their inflammation caused by Covid-19.

This paper has been read by more than 100,000 people, cited academically dozens of times, appears in numerous media reports, and generally has had a big impact on the question of whether vitamin D supplements are effective as a treatment for Covid-19. Unfortunately, it is also filled with serious deficiencies and errors that make it entirely useless as a piece of scientific evidence.

Let’s dig in.

Scientific Shenanigans

The study itself is described as a randomized trial where patients recruited from a hospital in Secunderabad, India, were either given high-dose vitamin D treatment or just received usual treatment for Covid-19. The authors then compared inflammatory markers taken before and after the treatment, and found an absolutely massive benefit for those people who’d received vitamin D.

Searching for “vitamin D”, of course, brings you many photos of oranges because algorithms are wonderfully bad at differentiating complex but important differences. Photo by Andrei Castanha on Unsplash

At face value, the results are impressive. People who were given vitamin D saw massive, potentially life-saving improvements in terms of their Covid-19…