One of the points made by Ioannidis is that we’re far too easily swayed to believe what sounds plausible when no one has yet asked “Could this be for some other reason altogether that we aren’t testing for?” and then tests for those reasons. That flaw applies here: studying just one community automatically implies that the results are invalid, simply because there’s no evaluation of other communities that share various attributes that could be hidden factors for the benefit.
EDIT: Per one article, that study “would try to cut through the stories and establish the facts about Ikaria’s longevity”. It has succeeded in establishing that, factually, Ikaria’s longevity is a thing relative to other populations. It has not established which of the properties of Ikaria are responsible. (For example, perhaps the island is slightly radioactive, and old people thrive in a slightly higher radioactivity.) Extrapolating that their diet is responsible is an instance of the A/X error described in the above-linked article.
> Conclusion. Modifiable risk factors, such as physical activity, diet, smoking cessation and mid-day naps, might depict the “secrets” of the long-livers; these findings suggest that the interaction of environmental, behavioral together with clinical characteristics may determine longevity. This concept must be further explored in order to understand how these factors relate and which are the most important in shaping prolonged life.
It could be that the quantity red meat is solely responsible for their longevity, or has nothing to do with it, or plays some role with the many other confounding variables.