An association study can not determine causation. It is a step towards more trials to determine causation. There can be confounding factors.

"In the first study, researchers found that having a flu-like illness increased the odds of having a stroke by nearly 40 percent over the next 15 days. This increased risk remained up to one year."

Forty percent of what though? (I did not look up the research paper). A forty percent increase in a 1 in 100 case is a lot. In a one in a thousand not much. In one in ten thousand not even a blip. They provided no hard numbers of the amount of occurrences. This is kind of like statin drug advertisements where they claim 50% reduction in heart related deaths where two in ten thousand people die instead of three. Whuppty doo. Thousands and thousands of people have to take the drug to improve the outcome of one person. Or put another way taking the drug ten years statistically shows you will live 3 days longer from no heart attack (all cause mortality is another whole conversation). This is how statistics can be used to create headlines (and sales) but are pretty much meaningless.

"There are many proposed mechanisms behind the flu-stroke link, but no definitive reason has been described to explain the association. Researchers suspect it could be due to inflammation caused by the infection."  From the article.

Interesting article. Thanks for the link. Headlines create clicks and research funding for more studies. But often there is more to the story than is revealed.

John