I recently spoke at an event dedicated to “making clinical trials faster” in the US. Mine was the first talk of the day, so I wanted to set up a framework of thinking about clinical trials that would be useful for further discussion. The attempt was a resounding failure at the level of the talks that followed — more on the specifics of that below — but I still think it’s a good framework so I’ll turn my 5-minute talk into a few paragraphs here in the hopes that more people start looking at the problem from that perspective.

The problem with “making clinical trials faster” is that clinical trials are two very different things with very different purposes to the point that we should probably use a different phrase for one of them. And you can optimize for one or the other, but not both. I am a fan of the mental technique where you figure out the edges of a jigsaw puzzle first — identify the rare extremes — in order to find the more generalizable truth that is somewhere on the spectrum. Here we have the opposite problem: we live in and try to optimize for the middling hodgepodge whereas the two extremes would have been preferable.

These are the two extremes:

• Set 1: Clinical trials that get new drugs and medical devices onto the market, or find new indications for existing drugs.
• Set 2: clinical trials that help guide clinical care and answer practical questions about drugs and devices that are already in use

Set 1 trials are Phase 1, 2 and 3 trials regulated by the FDA with 20-page informed consent forms, lengthy approval processes, strict safety and data monitoring procedures, and costs up the wazoo. The costs are OK since they lead to “value creation” for the economy, of course, but they also contribute to rising cost of health care. And they are especially OK in cases where they are covered by the Sponsor, either a multi-billion-dollar “big pharma” conglomerate or a well-funded biotech. The pembrolizumab KEYNOTE trials are a typical examples but there are too many to count. KEYNOTE-001 in particular is a good example of “speeding up” these types of trials (it was a phase 1 study that lead to accelerated approvals for two indications).

Set 2 trials are “pragmatic” Phase 3 trials that are IND-exempt are they use drugs in their approved indications and should have short, 1-page consent forms, minimal regulatory oversight, and trial procedures integrated into the standard of care to the level that most if not all costs should be covered by medical insurance. And medical insurance companies should be particularly interested in these trials, as they help optimize care and remove ineffective drugs and devices from the market, if not literally then at least by the virtue of physicians not using them any more. As such they are “value-destructive”: imagine a potentially billion-dollar blockbuster drug being found to be no better than a $2-per-dose generic. The RECOVERY trial is the best example of this kind of a trial, leading to quick establishment of standard of care in severe Covid-19. RECOVERY was wholly done in the United Kingdom, and examples of Set 2 trials in the US of A are few and far between.

The outcome of making Set 1 trials faster is more new drugs and devices on the market, including the hypothetical miracle cures that the burdensome FDA approval process is keeping away from patients, leading to large invisible graveyards (for that hypothesis, consider me a skeptic). The outcome of making Set 2 trials faster is fewer ineffective drugs and devices in use. They are the yin and the yang of drug development, creating harmony when they are in sync: new drugs come in, bad ones go out. But at the same time they are diametrically opposed, so it should be no wonder that the tradeoffs required to make one or the other set faster are different.

Of course, clinical trial infrastructure in the US is wholly dedicated to Set 1 trials. This is what the FDA was set up to do, what IRBs are primed for, what the entire ecosystem knows how to do, from clinical research coordinators doing the administrative work at the trial site, to the pharmacovigilance staff monitoring safety, to the administrative and budget offices of universities managing the contracts. But the two sets share the same ecosystem, and all the money being pumped into it by Set 1 makes Set 2 trials nearly impossible: who wants to pay that kind of money for something that could potentially destroy the value of a drug?

And this is my fear for the “make clinical trials faster” project: it will have cheerleaders from both the Set 1 and Set 2 crowd. But there is only one set of tradeoffs to be made, and if I had to bet I would bet that the Set 1 tradeoffs will win if for nothing else then for the major lobbying potential of the pharmaceutical industry. The system is already out of balance, and it wouldn’t take much to tip it completely into disaster territory where so many new but marginally effective drugs overflow clinical practice and it becomes a competition for the best door-to-door salespeople to persuade doctors that their marginal drug is (marginally) better than the other guy’s marginal drug. And if you want to compare the two drugs head-to-head in a Set 2 trial, well, you are out of luck because the regulatory energy of activation and the cost of paying for a trial database and the research coordinator and the data manager is so high that you can’t even… But of course, this is not a hypothetical disaster scenario, this is just modern medicine.

So I gave my 5-minute talk and what followed was a near-perfect encapsulation of Set 1 and Set 2 people talking over each other’s heads. On one hand we should make informed consent forms shorter (Set 2). On the other, let’s use “real-world evidence” to find new indications for drugs and speed up their approval (Set 1). Let’s ease up the monitoring requirements for trials (Set 2). But also do human challenge trials (Set 1). And so on, and so forth.

My gut instinct is to optimize the system for Set 2 trials and to untangle the FDA’s relationship for them. As much as some would like to shoehorn additional responsibility to an already overburdened federal agency, FDA has no place regulating clinical practice. Now, there is another set of federal institutions that have National Health in their name that could maybe lead the way for pragmatic trials, but they too mostly fund small Phase 1–2 studies of marginal drugs whose only distinction is that they came from an academic lab rather than the pharmaceutical industry (and this is because, of course, Phase 3 trials are too expensive to run on a federal budget).

But let’s leave that discussion for a different time.

1. It won’t be close. Most pollsters are hacks who commit even greater statistics crimes than physicians so their 50/50 is most likely to mean a landslide either way.
2. That link above is to Nate Silver’s Substack post, but please remember that he is also a hack who builds prediction models from the polling garbage he describes above while knowing it is garbage. That is even worse than what the pollsters are doing because shouldn’t he know better?
3. Worse yet are economists who excuse the pollster behavior: they see crimes being committed and think yep, that’s how it should be. This is a University of Michigan professor of economics and a senior fellow of some pretty serious Think Tanks who doesn’t realize that fiddling with your results after you’ve collected them in order to better align with the aggregate of other people’s results is scientifically unsound. I’d send all of his papers to Retraction Watch for a close inspection.
4. From 538 to the NYT to Nate, every poll aggregator has for months been fed back its own bullshit. Little wonder then that they all converged to a 50/50: complete ignorance.
5. Prediction “markets” are no better than equity markets in reflecting reality. Which is to say, they reflect the reality of vibes and wishful thinking, not the ground truth. They are best ignored.
6. Sátántangó (2019) by the Hungarian director Béla Tarr is a 7-hour masterpiece shot in black and white; perfect for watching on a crisp autumn evening like tonight’s, no other screens allowed.

Hell froze: I am about to link to hype-master Eric Topol in a non-judgmental way, because the article he is hyping is one that I co-authored. It’s titled “Engineering CAR-T therapies for autoimmune disease and beyond” and it came out yesterday in Science Translational Medicine as their one free article from the issue. I’ll stop there because it’s work-related and it’s good to have some boundaries.

Andrew Gelman recently wrote about Columbia surgery professor’s research missconduct. I haven’t looked into the details but it seems like the retracted papers were all about lab research with no true clinical relevancy. In that context, this part of the post stuck out:

Can you imagine, you come to this guy with cancer of the spleen and he might be pushing some unproven treatment supported by faked evidence? Scary.

I can’t tell whether this was supposed to be a joke or if Gelman truly believes that faking mouse experiments directly leads to using unproven treatments, but in case it’s the latter I have to say that the logic is stretched. Yes, the kind of person who has no qualms about fake data is probably not all that rigorous about the evidence for surgical procedures, but for all we know he could be a master surgeon with excellent technique and great outcomes who also happens to have been a bad judge of character and trusted a bad actor. I suspect it’s the latter: the kind of multi-tasking surgery “superstar” that the professor in question seems to be tends to spend a lot more time in the the operating room (or, for another kind of a superstar, the board room), than the lab.

Now, if he were a medical oncologist or any other kind of doctor that gives cancer treatment then maybe things would have been more dubious — that kind of research tends to jump to clinic too quickly and without merit. But unless you’re transplanting pig’s hearts and working on other large animals, the lab is so far removed from the operating room that it is extremely unlikely any such evidence could be used to back up actual surgical treatment.

Incidentally, that last link is to Siddharta Mukherjee’s abomination of an article titled “The Improvisational Oncologist” (subtitle: “In an era of rapidly proliferating, precisely targeted treatments, every cancer case has to be played by ear.") from the May 2016 edition of The New York Times Magazine (it’s a gift link so feel free to read it; caveat lector) and it describes actual scientific and medical malpractice of bringing half-baked — though, admittedly, not faked — ideas from the lab into clinic. Gelman didn’t comment on his blog back then, but he did praise Mukherjee the following year for a New York Times opinion piece “A Failure to Heal” (another gift link there) that is about — wait for it — clinical trials that show the treatment that you thought would work doesn’t. ⊕ These kinds of trials tend to be called “negative” but there’s nothing negative about them! They bring positive value to the world. Maybe our improvisational oncologist learn something in those 18 months that separate the two texts?

To be clear, what Mukherjee artfully called “improvisational oncology” was (lab) bench to (hospital) bedside medicine, which is distinct from bench to bedside research: the concept of bringing laboratory findings to clinical practice quickly, but still with some semblance of a clinical trial that includes a pre-specified protocol, informed consent and regulatory oversight. You know, all the stuff that decreases the odds of laboratory malfeasance endangering patient care. I say decreases the odds and not prevents them completely because we do have a case of a bad actor completely destroying an entire field of clinical research (Alzheimer’s disease). Can you imagine the damage that kind of shenanigans would do if we didn’t have clinical trials standing between the lab and the commercial drug market?

COI statement: I am involved in a [course about clinical trials][6 and think they are the best thing that has happened to medicine since a cloth merchant wanted to take a closer look at some garments so there is some bias involved, but then again say what you’ll do and do as you say is both a major tenet of clinical trialists and good general practice.

Adam Mastroianni nails it:

When people revile a degree from Harvard’s Extension School and revere a degree from Harvard College, they’re saying that the value of an education doesn’t come from the fact that you got educated. It comes from the fact that you got picked.

Alas, it is a paid post, and the more I encounter those the less I think of Substack and — this is not rational, I know — people who use it to blog. Because that is all Substack is: a blog with an optional, easy to implement paywall. I am not a fan.

The Nobel Prize in physiology or medicine went to Victor Ambros and Gary Ruvkun, two American scientists for their discovery of micro RNA:

The pair began studying gene regulation while they were postdoctoral fellows at the Massachusetts Institute of Technology in the lab of H. Robert Horvitz, who won his own Nobel Prize in 2002.

And so the Nobel family tree grows.

From the Institute For Progress-supported newsletter, Macroscience:

Last year, IFP brought together some of our closest friends and collaborators to put together a podcast series that would serve as a beginner-friendly introduction to metascience.

The result? “Metascience 101” – a nine-episode set of interviews that doubles as a crash course in the debates, issues, and ideas driving the modern metascience movement. We investigate why building a genuine “science of science” matters, and how research in metascience is translating into real-world policy changes.

So far so good. First guests?

Journalist Dylan Matthews sits down with economist Heidi Williams and IFP co-founder Caleb Watney to set the scene.

Bah-rump. Episode two?

OpenPhil CEO Alexander Berger interviews economist Matt Clancy and Stripe co-founder Patrick Collison to talk about whether science itself is slowing down, one of the key motivating concerns in metascience.

A journalist, an entrepreneur, two economists and a policy wonk gather around the fireplace to talk science. What seems to be missing is actual scientists. Stop me if you’ve heard this one before.

And if your retort is that few if any scientists have metascience as there full-time field of study, well, are any of the above doing it full time? I am sure the discussions will be brilliant — I will write up updates as I listen along — but the start looks a helluva lot like an echo chamber. Hope I’m wrong!

(↬Tyler Cowen, because who else. He will also be a guest in a future episode.)

Here are a few unrelated articles that crossed my inbox this morning:

• “Studios start at $1,927”
• "$1,134: The amount I paid for my round-trip cross-country Amtrak ride"
• “A six-bedroom apartment costs about $50,000, and the maximum income for a family of four is $160,000."

The latest episode of The Talk Show was with Taegan Goddard, who all the way back in 1999 founded the blog Political Wire which is apparently a continuous intravenous drip for people interested in US politics. Now, I’ve had other preocupations back then and not being an American citizen still have little to no interest, so this blog wasn’t even on my radar until listening to the episode. But now I wonder: are there any more relevant blogs I’ve missed out on, about medicine and biotechnology in particular?

ChatGPT’s first pass was mediocre. I’ll save you the verbalist padding, but here are its suggestions in response to my prompt: “Is there a website/blog like politicalwire.com or daringfireball.net but for biotechnology?”

• Endpoints News
• Fierce Biotech
• STAT News
• In the Pipeline
• Xconomy

It’s a 20% hit rate: only Derek Lowe’s In the Pipeline comes close to what I asked for. The others are all medium to big news outlets that yes, focus on biotech, but that’s not what I asked for. The second try, after I asked for more like Lowe’s, was a tad better:

• Timmerman Report
• The Niche
• The Biotech Strategy Blog
• Leonid Schneider’s For Better Science
• Science Translational Medicine Blog

That’s more like it! 80% now, and if I were feeling generous I’d give it a full 100% since In the Pipeline is, in fact, a Sci Trans Med blog. But then I asked for too much, and it hallucinated 3 more, two of which were hallucinations (BioPunk and BiotechBits, which were at least plausible names) and one was a sub-blog of Endpoints that also didn’t exist.

So, now I have two new blogs to follow (Timmerman Report and The Niche; Biotech Strategy is behind a paywall and I’ve already been following the others), and an ever-increasing urge to update the Blogroll, which has been under construction for the past five months with no end in sight.

I had Linus Lee’s blog The Sephist filed under “Paused and Defunct” for a while now, but he is back at it. Although most of the subject matter is out of my wheelhouse this mental model of Motivation as a function of Exploration (or was it the other way around) rang true — certainly truer to the scientific method than what my 6th-grader has been hearing at school.