I wrote the passage below the line as a follow-up for a now deleted comment on scurvy. The comment covered much more than European history of citrus as prevention and cure, I was just adding some detail on the British/European story. In direct context for the question, it's worth noting that Lind reviewed two centuries worth of (predominantly European) literature and even then, it took nearly 50 years for his conclusions to be implemented. (When I say "predominantly European" I mean by authorship; I don't know how much of this literature took inspiration from the practices of other cultures.)

It's probably worth adding a note about why it often takes a very long time to demonstrate cause and effect in health-related contexts. You note that the Somali "superstition" was dismissed as a coincidence of timing when it was actually correct. But there are situations where a coincidence of timing does create a powerful illusion of causality where there is none. Notable modern day examples include the (non-existent) link between MMR and autism, because the MMR jab is commonly given at around the same age as it becomes possible to diagnose autism; and the apparent effectiveness of homeopathy when people tend to reach for new treatments at the time when their symptoms are at their worst and so will likely experience an improvement shortly afterwards regardless of treatment. These coincidences of timing create a powerful illusion of causality which, as in the case of malaria, can be difficult to test. They will sometimes be absolutely correct and sometimes, not so much.

While Lind's 'fair test' of treatments for scurvy took place almost three centuries ago, and was itself based on 200 years worth of literature on scurvy, it was a long slow road to introducing more scientific methods into modern medicine. Ignaz Semmelweiss (hand washing and childbirth), Florence Nightingale (nutrition and hygiene) and John Snow (cholera and water pumps) were all in the mid-nineteenth century, with their findings (and the new germ theory) broadly accepted by the end of that century. The first randomised controlled trial (an elaboration of Lind's 'fair test') was conducted in 1948 (by the MRC, in tuberculosis). We have, to a very great extent, been playing catchup ever since. There will be many traditional medicines or beliefs which turn out to be absolutely correct (eg 1,000-Year-Old Onion And Garlic Remedy Kills Antibiotic-Resistant Biofilms in The Lab) and others which will turn out to be based on illusion or outright quackery. The difficult part is determining which is which.

---

Just adding a reference for the story of how the British discovered citrus as a treatment for scurvy: Documenting the evidence: the case of scurvy. Importantly, it involves both one of the first controlled tests of treatments and also one of the first reviews of the literature:

In 1753, James Lind, a Scottish naval surgeon and medical graduate of Edinburgh University, published a 450-page, three-part Treatise of the scurvy (1). At that time, scurvy was killing thousands of people every year and was responsible for many more deaths of sailors in the Royal Navy than enemy action. Believing that one of the reasons there was so much confusion about the diagnosis, prevention and cure of scurvy was that “no physician conversant with this disease at sea had undertaken to throw light upon the subject”, Lind set about filling this gap, with a clearly stated commitment to base his work on “observable facts” rather than on the theories that dominated medical decision-making at that time.

Lind’s Treatise is a classic for two main reasons: it is one of the earliest accounts so far identified of a prospectively organized controlled clinical trial, comparing six commonly used treatments for scurvy, and it is a systematic review of what had previously been published on the diagnosis, prognosis, prevention and treatment of scurvy.

That's not to say that Lind's work was immediately definitive, any more than the previous work he cited, which spanned two centuries (of predominantly European literature). In part because his conclusions were not particularly clear and in part because changing medical practice is slooooow, even now:

The least satisfactory feature of Lind’s Treatise is that he leaves his readers confused about his recommendations. Some passages suggest that he is very clear about the implications of his review, for example when he writes: “Some new preservative against the scurvy might in this treatise have been recommended; several indeed might have been proposed, and with great show of probability of their success; and their novelty might perhaps have procured them a favourable reception in the world. But these (citrus) fruits have this peculiar advantage above anything that can be proposed for trial, that their experienced virtues have stood the test of nearly 200 years.” The number of times each of the six purported treatments compared in the experiment is mentioned in his book should leave little doubt about his preferences: 117 mentions of orange(s) or lemon(s), 29 of vinegar, 29 of vitriol, 19 of seawater, 16 of cyder/cider, and 2 of nutmeg. In spite of these apparently clear indications of Lind’s conclusions, however, his readers are left wondering whether he regards fruit and vegetables as relatively more important than fresh air — one of the other factors that he cites as crucially important in preventing and treating scurvy. This is probably one of the reasons that it took so long for oranges and lemons to be widely recognized as antiscorbutics.

Although Lind’s Treatise was published in three editions in English (1753, 1757 and 1772), two in French (1756 and

1783. and one each in Italian (1766) and German (1775), it was not until a year after Lind’s death in 1794 that the Admiralty issued a general order sanctioning the provision of lemon juice in the navy on a far more generous scale than previously. The effects of the Admiralty’s order were dramatic: within two years scurvy had more or less disappeared from the Royal Navy. Today there are a variety of reasons why research evidence has little or no impact on policy and practice. It seems likely that similar problems also existed in the years after the publication of Lind’s classic work.