Well, establishing a linear causal relation isn't always possible, at least not ethically. The recommendations have to balance average dietary intake, tolerable

Well, establishing a linear causal relation isn't always possible, at least not ethically. The recommendations have to balance average dietary intake, tolerable upper limits, and what would constitute an adequate intake that is both necessary and sufficient. For Vitamin C, for example, toxicity is probably in the 2-3 grams range, while just 10 mg/day is sufficient to overcome symptoms of deficiency. So what should the recommended daily allowance be? You don’t want to set it at the minimum level, to allow for individual variances and greater needs among particular groups. You also don’t want to set it artificially too high. Different agencies and countries draw that line differently. Japan sets it at 100 mg/day, while India at 40 mg/day. The Indian agencies have to be cognizant of the fact that given average Indian diets, setting the "healthy" level at 100 mg/day would artificially pathologize most Indians who do not suffer from Vit C deficiency symptoms. The situation is more complicated with recommendations around vitamin D. The most obvious and extreme abnormality associated with low D is bone demineralization. This is largely alleviated by levels above 12 ng/mL. Yet when you look a cohort of healthy young Caucasian coast guards, their vitamin D levels are 30 ng/mL. So what do you set as the recommendation: the level required to prevent symptoms, or the levels seen in (some) healthy people, but with no obvious health advantages? Most groups have tended to split the difference: set the minimum at 20 ng/mL, and the “optimal” at 30 ng/mL. Is there a health benefit with supplementing if the level is below 10 ng/mL? Quite likely. Is there a health benefit supplementing if your level is at 20 ng/mL? Probably not. The picture is even further complicated if you consider that the healthy cohort of young white coast guards in Hawaii is not representative of the population at large. Multiple studies have looked at healthy black and brown cohorts from tropical environments, and find most average around 20 ng/mL.