A proper logician would understand the inherently irrational/unreasonable/illogical fault plaguing most "clinical" studies and thus, the indisputably flawed "evidence" resulting from them. "Clinical" studies are often performed in controlled, "clinical" settings with predetermined variables, and many variables are discounted for ease. Real-life settings are undeniably different from such clinical settings. Additionally, they ignore empirical evidence from real life if it doesn't gel with their "laboratory"-derived "facts", when it ought to be done the other way around.
For example, for decades now, the scam about saturated fat being a malignant evil and plant-based/vegan diets being a panacea has been pervading mainstream discourse. Yet, the original four Blue Zones: Okinawa, Ikaria, Sardinia, Costa Rica, where people have longevity and long healthspans have traditional diets rich in saturated fat, low in seed oils, high in red meat, low in simple sugars.. In fact, in Okinawa, the Blue Zone with the longest lifespan of them all, they even had lower seafood consumption than the rest of Japan. Even more tellingly, after 2000, Westernization changed Okinawan diets, and their lifespans decreased. On top of that, they discard the French Paradox, Israeli Paradox, and refuse to consider the unhealthiness of the largest population of a historically vegetarian ethnicity in India.
In this scenario, a bunch of people have testified to using split dosing because it helps them with greater appetite suppression, and staving off food noise. Yet, some people claim a lack of clinical evidence as reason enough to dismiss these assertions. That same flawed principle is used in other contexts too, where it is equally wrong because it fails to account for a very important fact:
ABSCENCE OF ANY EVIDENCE (EVEN CLINICAL) IS NOT EVIDENCE OF ABSCENCE.
I disagree with a great deal of this.
Medical research apart from purely economic concerns, which are not irrelevant, is about finding the best possible ways of determining true outcomes with reliable information out of noisy messy data. This is very hard to do. And there are lots of different ways to try to do this, each which produces different types and qualities of data, some of which are more reliable than others. And it really matters, peoples lives depend on getting it right.
In general the highest possible quality of data that exists in medicine is a placebo ( or alternate drug ) controlled , double blind, prospective clinical trial. Where neither the researcher or the subjects knows if they are getting the drug or placebo. This matters as placebo responses can be astonishing, for example placebo morphine in acute severe pain works, not as well as actual morphine, but far better than anyone would guess. And researcher bias matters as well. You can go as far as triple blinding so the data is randomised so even the data analysis is blind but this is not very common. The patients in the two groups need to be selected to match as closely as possible. And then the drugs are given and you find out what happens, analyse the results and determine if your new drug or other treatment worked or not and apply statistical methods to determine how likely it is to be a true effect or a chance effect. The larger the number of patients and the longer the time usually result in higher quality data. The amount of thought , time , and evolution of this process is significant, gradually over time more and more flaws in the process have been found and removed, and in general modern studies have fewer methodological flaws
The other method is to gather up all those controlled trials and add up the results to make an effectively much bigger study or a meta analysis with systematic review. If the trials are similar enough then the quality of result from these can be even higher than the individual trials, but recently paper mills have been cranking these papers out by the millions which have highly variable quality, making it hard to determine which ones are the good ones.
Any drug that gets approved for human use in modern times has gone through these trials, and this process is as good as humans can get to absolute truth, where the data is noisy. So in general the results of this type of trial is as reliable as it is possible to get, and is considered accurate enough to bet peoples lives on. As that is usually exactly what is at stake.
So what is known from these types of trials is the gold standard of research, the results can be relied on , assuming they are interpreted by people with sufficient skill and knowledge. The problem is that these trials are narrow and have to be so they can control as many variables as possible to make the information trustable. But often they do not answer all the questions or the results might not apply outside that narrow context, and often clinical decisions need to be made anyway without having that information. And sometimes they are later proven to be wrong, a study that has very good statistical evidence of only a 1% chance that the results are from chance is still going to be wrong 1% of the time. But this is as good as it gets.
There is a huge number of other methods of obtaining information relating to health outside of that type of trial, with varying degrees of trust that can be obtained from the results.
Population based studies do provide information, but they usually do not provide evidence that qualifies as proof of causation of illness or that a particular treatment will work. There are many examples of prospective controlled clinical trials that showed results different to those expected from population based studies. The current best quality information that exists on diet and health is that the mediteranean diet or similar variants, are associated with the lowest health risks. I am not expert enough to explain the processes used in these more recent studies.
Empirical research is in general at the other end of that scale in terms of trustability, as in can it be proven to be true or as close as we can get to that. An observers experience of a treatment of a patient is a valid experience and is true for them as is the patients experience of that treatment. The problem is whether that data can be reliably applied to others. Common sense would say yes. But there are confounding variables here that the clinical trials try so hard to avoid. Placebo responses to start, either from the patients end or the observers end. The patient or disease process itself may be responding in this case in an atypical fashion. Were I an expert in the philosophy of science I am sure I could add many more here.
In general empirical evidence is not useless, and all doctors use it and rely on it, and repeated evidence over time does make it more reliable, but medical therapies based on it do include things like blood letting that were used for thousands of years, so it can most certainly provide incorrect or untrue information. And in general is not considered to be good enough quality evidence to bet peoples lives on, unless there are no better options.
