RCTs are great, but they are almost always underpowered when studying nutritional deficiencies. For the cited study, the cancer incidence rate was 1.2% in 5 ye

RCTs are great, but they are almost always underpowered when studying nutritional deficiencies. For the cited study, the cancer incidence rate was 1.2% in 5 years among 12500 recipients. That's barely 150 patients. Getting statistically significant effects with such low sample sizes is very difficult. We will have to live with correlational instead of causal studies for nutrition. If you are interested, smoking causes lung cancer hypothesis was based on correlation- which was how cigarette manufacturers attacked the hypothesis. Proving a causal connection between smoking and cancer is much more difficult. https://towardsdatascience.com/correlation-does-not-imply-causation-92e4832a6713 https://towardsdatascience.com/correlation-does-not-imply-ca...