Testing Testing

In the world of Drugs, in the USofA the FDA is the gatekeeper. Until the Ascension of Hate filled, Petty, Paranoid, Demented, Dictator Don, that has meant Gold Standard placebo controlled, Phase III clinical trials. Now, it is not widely understood that FDA seldom interferes with the conduct of clinical trials by drug companies. In almost all cases, if the Sponsor (aka, drug company) proposes to do something that endangers patients (or is discovered in the course of a trial), then FDA will step in. Otherwise, FDA is happy to let bidnezz piss away money on compounds/trials that are doomed from the gitgo.

Again, in general, final phase (III) trials are almost always placebo controlled, blinded exercises. During the trial, no one involved with the trial (who has some skin in the game) knows what the patients are getting. There's a sub-culture in the drug biz, the Contract Research Organization, which mitigates between the FDA and Sponsors. IOW, while a Sponsor pays for, and usually, designs the trial, it's the CRO which carries out the trial separate from the Sponsor. Often enough, there's also (by various names) a Data Monitoring authority (again separate, but reporting to the Sponsor) which has continuous access to data. The DM may find issues, and will recommend to the Sponsor to modify/halt the trial; usually for futility, but on rare occasions, wonderful results. But the DM can only recommend. The Sponsor is supposed to notify FDA of any Adverse Events as they happen. But that can, at times, by contentious.

Which brings us to this X post from Feuerstein:

The validity of external control arms like the one used by Uniqure is still hotly debated, even as regulators, including the FDA, allow them under certain circumstances. It's been lost to memory, but recall that the AMT-130 study started with a true, sham-surgery control arm that showed no benefit for the gene therapy. It was only after the company reconfigured the study and started using natural history comparators was a benefit observed.

For those who find that gibberish, here's what it means. From the beginning, the Sponsor established a placeco controlled, blinded trial. For some part of that period of the trial, it became known that the therapy fucked the dog. The Sponsor: "what the fuck do we do now!!!!" The answer: 'Will no one rid me of this turbulent control group?' And the Sponsor did. Viola!!! Success!!

Feuerstein's post was, well addendum, to the report that the trial results reported recently were stat sig. That's good. Drug companies, with more aggression recently, have been trying to scrap placebo controled, blinded, PII trials (more so in some areas than others; that's another episode) where the Compound/Therapy is compared to some placebo equivalent (so far as patients and practitioners know) to comparing to some data set representing historical experience/drug/therapy. All the patients get the Therapy. Their response is compared to this historical record. Of course, it does matter what that historical record consists of.

And, it's also worth mentioning: the trial was merely a PI/II trial of a gene therapy; such 'drugs' have a different profile from 'normal' drugs. This is not normally grounds for approval. Huntington's Disease is particularly troublesome.

Feuerstein was just reminding, mostly retail plungers (not the Pros, who mostly know better), to be vewy, vewy cawful believing this result.