Antidepressants - Going Backwards: RANT: The Problem with RCTs and Evidence-Based Medicine

Ethan: One of the themes of this podcast is that we, as in the medical establishment, have to take back control over medicine and what is considered good medical care, what are considered good treatments, good diagnostic criteria. And right now, we are in shackles. We are at the whims of the pharmaceutical companies and we have very little control over the messages that are put out. This is about the only way that I can think to get the proper messages out there.

Ken: Let me speak a little bit to that as well. Science is about mechanisms and causality. One of the famous researchers in this field who got the Turing Award, which is the computer and IT equivalent of the Nobel Prize, Judea Pearl, he made the statement that I thought just expressed this wonderfully; ‘Science is nothing without mechanism and causality,’ and RCTs tell you nothing about mechanism and causality. You can argue that it would be possible to design an RCT that might throw some light on mechanism and causality, but basically RCTs and by extension, so-called evidence-based medicine. But evidence-based medicine isn't evidence-based itself. Evidence-based medicine has never demonstrated that the results of using evidence-based medicine actually produced better outcomes. That's the fundamental idea. And yet it's never been tested scientifically. It's a joke, isn't it? It's just window dressing.

The model of randomized controlled trials as being essentially the only way of getting approval for a drug, that all started in the early 1960s. The American political system decided that they needed to improve the way drugs were dealt with and regulated. And they decided that they would review all the drugs that were available and make them meet some kind of standard to be approved. And the FDA was created and given some teeth. Especially in the early stages, the FDA probably had far more responsibility and work than the personnel and the expertise available to them were capable of meeting. So it's important to understand that a lot of what happened in the early part of the FDA business in the ‘60s was, I would go first to say somewhat amateurish, and I'll give you the example. I may get the name slightly wrong, but Lasagna, L-A-S-A-G-N-A, I imagine that's an Italian name, and Walter Modell, they were the 2 people who were responsible for starting the FDA off on this road of saying that you had to have two randomized controlled trials, which showed effectiveness, to get over the finishing line for a drug to be approved by the FDA. And they latched onto the work that had been done by Sir Anthony Hill, Austin Bradford Hill, the famous statistician in the UK, who played an enormous part with Richard Doll in the whole smoking-cancer debate. They built on that role of RCTs with the golden standard. But, within 3 or 4 years, Louis Lasagna was on record as saying already, randomized control trials had gained too much influence and people were beginning to deride clinical experience and the opinions of experienced researchers and experimenters and say if it wasn't a randomized controlled trial, it wasn't real.

And I find that so interesting cuz it's just such an illustration of how people completely fail to learn from history. I'd have to say that possibly the quotation I find myself remembering most frequently, the famous Spanish-American philosopher, Santiana, said, ‘History, far from consisting in change, depends on retentiveness. Those who cannot remember the past are condemned to repeat it.’ And I just think that sums up what we've all done over and over again so well. So many of these famous people who were in on the ground floor of creating this monster of the hegemony of the RCT were themselves saying within a few years that the pendulum had already swung too far. Heavens, in fact, Hill himself actually said, ‘Not only has it swung too far, it has come completely off the hinges.’ Wouldn’t you think that especially English chaps in those days, tended not to say things that were controversial. To think that somebody like him said that way back in the sixties, early sixties, it's extraordinary, isn't it? Even then he was saying the pendulum had swung so far in favor of RCTs that it'd already come off the hinges. Since then, it's swung further and further so that now, if you express any opinion now that isn't based on an RCT, you're a pariah. You are sent to Coventry, aren't you?

Ethan: Oh I'm sure that'll happen to me with my opinions that will be shot down and, I'll say this sarcastically, you'll be glad to hear that the FDA now no longer requires two positive trials; they only require one positive trial for conditional approval, such as the new monoclonal antibody for Alzheimer's disease that removes beta amyloid, but has pretty shitty efficacy in terms of actually improving outcomes, has a pretty bad side effect profile, the newer version causes brain bleeds in 15%. But they now allow those companies to get approval; they're already based on a biomarker. They're not based on actual clinical outcomes.

Ken: Proxy outcomes are very dangerous. Short-term proxy outcomes. It's disastrous. But of course, that's again, driven to a substantial extent, by money and the pharmaceutical companies. Let's wind up by going back to what we were saying before about RCTs. And cause-effect relationships, mechanism, and causality and stuff like that. Cuz you remember I alluded to the fact that right back in the sixties people weren't taking sufficient notice of clinical experience and they were relying on RCTs too much. And of course when you give examples, like patients with psychotic depression who got better within a couple of weeks of taking Tranylcypromine and such cases, people say, ‘oh, oh, oh, Well, you know, a series of anecdotes doesn't constitute evidence.’ That is actually from a point of view of philosophy and logic incorrect because I was starting to write the outline of something relating to this not very long ago. And as a provocative title, I called it ‘Strangling the Bloody Cockerel’. Now, those without a classical education may not understand that, but of course the whole business of post-hoc ergo proptor hoc in philosophy and logic after that, therefore because of that, goes right back to Aristotle. And the analogy they used to justify that was the cock crows every morning and the sun comes up, therefore the sun rising is caused by the Cock crowing. Post Hoc Ergo Proctor Hoc, classic logical mistake. That's not actually true because Judea Pearl's answer to that in terms of methodology would be what he calls the do operator. If you want to unravel mechanisms and causality, you change what's supposed to be the cause. Now, if the direction of causality is from the cockerel to the sun, then strangling the bloody cockerel will not stop the sun rising, and you have learned something about mechanism and causality. This is what's lacking in RCTs. RCTs are divorced from science in the sense they essentially say nothing about mechanism and causality. And without mechanism and causality, science is muted essentially. What you can do if you are treating somebody with Tranylcypromine for psychotic depression, people say, ‘oh, well you admit them to hospital and you've got all these placebo effects. They've been referred to a special specialist, and now they're going into a special unit and they've got lots of nurses waning around them and blah, blah, blah.’ Of course that is irrelevant because if you do that, you start 'em on Tranylcypromine the day they come into hospital, the time period between starting the intervention and them starting to improve is gonna be the same, irrespective of whether you start 'em on Tranylcypromine the first day they come in or three weeks later, like in the typical teaching hospital where they engage in mental gymnastics, that's not usually the word I use to describe it, but I'll leave listeners to imagine the words I usually use. They engage in mental gymnastics for weeks on end before they decide what the diagnosis is and what the treatment is, so that's the do operator, strangling the cockerel. That helps you to establish cause-effect relationships.

And the point of all of that is that helps you to understand why properly recorded and carried out clinical practice and clinical experience is actually a powerful scientific technique that in many situations is superior to an RCT. An RCT by its very nature simply cannot investigate many of these things. Obviously, the huge general problem in terms of all the drugs that are being tested now is all the patients who we'd be most concerned with severe depression, who are suicidal, might have comorbidities, blah, blah, blah, are the ones that are excluded from the trials. Need I say more?