Join us to discuss some pertinent examples of bad research and problems with the FDA Approval process, especially in mental health and with psychiatric medications.
Ethan: Pharmaceutical research is done with the intention of FDA approval. It's really costly to develop a new drug. For example, I know that Roche, their shares tumbled somewhere between 5 and 10% several years ago when they were trying to roll out a new Alzheimer's medication. What some companies do is, instead of creating a new drug, they try to rebrand drugs, so that they can get a 5 year extension on the patent, if they get a new approval, new indication.
Michael: And that's the tricky part. So a lot of times you start with a very straightforward one. You have something that has a unmet clinical need, say schizophrenia, is one that people pick on because it tends to be easier to get these [00:01:00] indications for something like schizophrenia. But then you start getting your extension into more mainstream conditions, Bipolar depression, bipolar mania, major depressive disorder and now you don't just target psychiatrists, but you can get the PCPs to start prescribing. That's where they try to target some of the medications. If you can get to where you tell your average PCP that if you see somebody has these symptoms X, Y, and Z, we've got a medicine for that.
Ethan: So Seroquel, Quetiapine is the generic version, is an antipsychotic, originally FDA approved for schizophrenia, directly to your point. You get 20 years when you create a new drug; that it is your drug and nobody else can sell it, so that you can recoup the money that you put into developing it and researching it and everything. Towards the end of their patent, they get it approved for, what was it, bipolar.
Michael: You've got a couple, you've got bipolar depression, and maybe you've got unipolar depression. You've got as an adjunct for [00:02:00] major depressive disorder. In some countries, that's actually approved for generalized anxiety disorder.
Ethan: So 2010, AstraZeneca has to pay out 520 million due to fraud; hiring MDs as their thought leaders or their docs who were trying to push their product. They were trying to get a new indication for aggression and agitation in nursing homes. The doctor's supposed to oversee all the research and conduct the clinical trials and it was all just written by the pharmaceutical companies. The entire study, the papers about it, the efficacy, they were written by some ghostwriter and then they just throw the doctor's name on it, aka fraud.
Michael: The physician is good at spotting advertisements, like direct consumer ads, but this is for a journal. This should be science. You see the name on it and you're like, ‘That's that guy that spoke at the conference last year I went to, or that's a physician. He or she works at this big hospital. Okay, [00:03:00] there's a legitimacy to it. But in some cases, I don't know if it was that one or if it was (Paxil), they deposed one of the physicians and she was like, ‘Look, I don't know any of what you're talking about. I gave them kind of a boiler point statement about it and that was it.’
Ethan: Probably the most fucked up part in my mind is that they pay 520 million. Oh my gosh, that is so much money. Holy crap. 520. How did they survive? Like we really brought the hammer down on 'em. Seroquel, for that year, their revenue was 33 billion.
Michael: That was a medication that, at one point, was being used off-label for insomnia, for PTSD. There was a paper written that said, in a military theater, if you need a medication you give to soldiers in combat, use Seroquel, and they cited the package insert as the reference in this paper. The problem is the package insert didn't say any of that because all it could talk about was what it was approved for and it was approved for none of those things. They were using it for all these conditions and in the VA, it became in the top three medications period; (for) a [00:04:00] medication approved at that time for schizophrenia, bipolar disorder, maybe depression, to be top three expenditures for the entire VA network. You think of all the cholesterol medicines, all the blood pressure medicines, all the heart medications. Out of all of those, there's only two medications the VA spent more money on than Seroquel during that period of time. That's a lot of money, as you said, so you can afford to pay a fine.
Ethan: It's a public perception gain. You throw a number like 520 million out and you put it in the papers, on the Internet and it's, ‘Oh my gosh.’ But you don't pair it with, ‘It is 1% of their total revenue for that year.’ And it's not just AstraZeneca. It's not just Seroquel. You've got Valproate. Who is it? Abbott? Yes. Abbott. AbbVie. Initially this was an anti-epileptic or an anti-seizure drug, and then it became a “mood stabilizer,” or a drug thats used for bipolar disorder. Mood stabilizer is another term that was coined by the pharmaceutical industry, not by [00:05:00] physicians.
Michael: One of the best talks I ever went to was 2 docs arguing about what a mood stabilizer was. I learned so much from that talk. World Federation of Biological Societies of Psychiatry, I think, has a great paper that they go back and forth about a definition of it.
Ethan: So, Abbott pays 1.5 billion in 2012 for the exact same stuff. Ghostwriters, essentially fraud, right? Just signing docs names on things that they had nothing to do with. The studies on the SSRIs that got FDA approval, negative studies, those just get swept under the rug, and we're left with a couple of studies that show, somehow, statistical significance, in terms of crunching and likely manipulating the numbers, but it is not a clinically significant difference.
Michael: And that's a huge thing. You can have a massive number of patients enrolled in a study, find a small difference statistically just based upon, if I give you a dart board and a hundred darts, one of 'em is probably gonna stick. But again, does that mean [00:06:00] anything? And that's where you have to look at other things like effect size, and should you have to publish every single study, or at least acknowledge it, even if the study results are negative. And believe it or not, YOU DON’T HAVE TO. You can run 30 studies and if 27 of them are not so good, you don't have to publish those things. Technically you do, but it's not enforced. If you did a study that shows that your drug causes brain cancer, you don't have to publish it. If you do 40 studies and then you magically find that the drug is effective in one of them, you can publish the one. There's like a meme where Mario’s jumping over something, like all these dead bodies underneath; that's exactly what it is; there's all these dead bodies of all the other failed studies you didn't have to publish, so you just magically found 10 milligrams work because you studied five and it failed, you studied two and it failed, you studied 20, and you don't have to publish any of those. You just have to say, ‘You know what, yeah, we did a study, did two studies, it was super effective in both of them, we think it should be on the market.’ The FDA doesn't enforce it. So Ben Goldacre says every study that started [00:07:00] on clinical trials should be published and it should be enforced with fines. I was proposing that Iloperidone (Fanapt) might be good for PTSD. You know what? It got studied. Never finished. No idea. Maybe it worked, maybe it didn't. We will never know what happened in this site.
Ethan: Probably didn't.
Michael: Probably didn't if it didn't get published.
Rob: Is there like a certain sample size that you have to have?
Michael: Just enough so you can do your calculations yourself and to achieve a 80% power in a study, we need X number. But that's tricky. Sometimes a study is underpowered and it won't show a difference, and that's where you can pool the studies together. And so some companies will run a bunch of identical studies that are smaller and then you take some of them and you can either publish 'em on their own if it's a good study. If they're close enough, you say, all right, I'm gonna put 'em all together. And then the number of patients is bigger, and now it's powered to detect a difference instead of the small ones. So you can tweak the data.
Ethan: They can also remove patients from studies who aren't fitting some criteria.
Michael: At the discretion of the [00:08:00] clinician researcher. There's this little word at the end of the exclusion criteria.
Ethan: …who is being paid by the company to conduct the study? I'm trying to think of a good real world example of a situation that is similar. It's like me working with an attending physician in my residency and they're like, ‘Oh no, you really should prescribe Wellbutrin (Bupropion) to this patient. I really think you should prescribe Wellbutrin, but what do you want to do? I'll let you make the decision.’ But if I say anything other than Wellbutrin, I'm gonna be scolded or I'm gonna lose my position. Same thing with these docs, right? You have this kind of under-the-radar pressure put on you to operate in the best interest of the company that's paying you. You're not gonna be employed very long if you're not working with them to try to get their drug approved.
Michael: Yeah, like some of the times they'll say this other person that we paid, they were involved in the design of this study, but they weren't involved in the writing of it. And sometimes literally the drug company itself is [00:09:00] involved in not just designing this study, but even how you write the results of it.
Ethan: I didn't realize this until recently. The SSRIs, or the things like Zoloft (Sertraline). The studies done on those medicines are using, or some of them, something called the Hamilton Depression Index. Patients report their symptoms on a scale and whether you're having suicidal ideation or whether you're having a change in your appetite, both of those get the same number of points assigned to them. It's a 52 point scale. And medicines like Sertraline, in the two positive studies taking out all the negative studies, they changed the Hamilton Depression Rating Scale score by a factor of, I think it was 1.8. So we're talking about possible 52-item score. The worst depression you could ever have is a 52. You take this drug for 8 weeks and all of a sudden, you go from a 52 out of [00:10:00] 52 to a 50 out of 52. And we call that effective.
Michael: Yeah. And then the other interesting thing, there's even different rating skills. If I recall, even the Hamilton D tends to be the one that I think has more questions about how I feel, more somatic concerns. You can even use a different rating scale than another one in studies to tease out, Am I gonna get a better result out of this one than that one? And you create some studies, oh, I have a primary outcome, then I have 30 secondary outcomes. There are gonna be patients who do well on SSRIs and there's data showing that. But not everyone is going to have that robust response. And you have to go at it with that understanding. And that's where, the data shows, your trial showing that, a third of patients will respond from the first trial. But again, a number of people won't. You have to set up the expectations correctly. And one of the issues, I guess to go with, is who provides the money for these studies? Is it a company looking to recoup the cost in terms of the sales of it? What happens when a medication goes generic is you [00:11:00] lose a lot of data on it. An example you've alluded to before is lithium, a medication with wonderful efficacy, both anecdotal and clinically in use for bipolar, but since this is a medication that is at this current point cheaper than dirt, and essentially contains what's in dirt, nobody is barking up the tree to get it approved (for major depression)… The same thing with Prazosin, is another one that nobody's trying to get Prazosin FDA approved for nightmares because it's been generic for a very long time, so nobody's gonna recoup the cost. If you spend 10 million designing a study for it, and you're one of 20 manufacturers, who's gonna run the study if I know that if I run the study, your generic company is gonna get the same amount of profit as mine? The only difference is I just pay 10 million to do the study.
Ethan: Yeah, lithium drives me crazy because it's been totally demonized as a medicine. It pops up in drug interactions in the EMR, the Electronic Medical Record, systems that I never learned about throughout medical school as a possible side effect or possible interaction. It is the most [00:12:00] effective treatment for bipolar disorder. It can allow people with bipolar to lead pretty much normal lives. In about a third of those people, they're called super responders to lithium. It is a natural substance. It has potential efficacy in dementia, in preventing herpes recurrence, CTE, is an antiviral in a lot of different viruses, but nobody makes any money off of it.