The same week Alexey Guzey proposed abolishing the NIH, two more essays popped up:

• Distinguishing the real from invented problems with the NIH, from Sasha Gusev, himself an NIH-funded geneticist
• The NIH Needs Reform: Here are 10 Sensible Suggestions from Joseph Marine, a cardiologist

A few things came to mind:

• You can clearly see the difference in backgrounds. Gusev’s essay is an “insider-y”, show-me-what-we’re-doing-wrong approach. Marine’s is outward-looking-in, dealing more with perception of the NIH.
• Marine makes several immediately actionable proposals which are the policy equivalent of politicians kissing babies but since these days any politician seen kissing a baby would be called a creep (or worse) I suspect that even those modest proposals would become divisive.
• Neither states conclusively what the NIH is for. Is its mission to give out grants? Advance biomedical science? Or help people live longer and/or better? I’d say it’s that last one and that everything else is means to that end.
• So if we see the NIH as a grant-giving machine, I guess we could give it a passing grade. The awardees certainly seem happy! But in the last 30 years we saw several massive public health crisis, from the obesity epidemic to the opioid murders to the bungled response to the pandemic so from that standpoint at least there is room for improvement.
• Gusev’s essay does not at all consider the opportunity cost of the current system. He lists length of grant proposals as an “invented problem” and unironically writes (emphasis mine):

The last NIH proposal I submitted was about ~150 pages which might indeed seem daunting. But only ~12 pages of that was dedicated to science and will be the focus of study section reviewers (and I can also assure you that I wish I had more than 12 pages to work with). The remainder was some combination of budgets, resumes for all of the personnel involved, descriptions of the data and resources, and contractual language between the NIH and my institution. Nearly all of it was handled by experienced grants administrators in my department who can put these documents together in a matter of hours.

• You don’t have to be an expert in probability and statistics to see how this “invented problem” leads to a winner-takes-all system, the winner being a handful of investigators in a handful of institutions. This is an incredible systemic risk that can lead to billions of dollars wasted and set back an entire field of study by decades.
• I want to read something from Adam Mastroianni on the topic of NIH reform.

I am at ACR Convergence all weekend, but here are some quick shots:

• How it went
• Epistemic hygiene
• Maybe Bluesky has “won” (ᔥDave Winer)
• I hated Dubai until I learned about it

Here is a good list of pragmatic trials (on X) that a national institute serious about the health of the public may want to sponsor and/or run. There are many such open questions in oncology as well, and don’t get me started on diagnosing every deep venous thrombosis for hereditary thrombophilia.

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.

A Wall Street Journal article on physician work-life balance prompted lots of online chatter, including people remembering their parents' dedication to the calling. But times have changed. The choice now isn’t between spending time with family and patients, it’s between spending it with family and corporations. If practicing medicine were more meaningful, there would be less of a retreat to family life by people who self-selected for delayed gratification and frank masochism.

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.

Speaking of blogs of old, Joel Topf’s Precious Bodily Fluids has been online since 2007. As most, he went from writing several times per week to every few weeks to not even every month as life moved to Twitter but he just published a new post that includes Neal Stephenson’s treatise on the Hole Hawg and for that alone is worth a shout out.

Halloween is nigh. This year, our eldest decided to dress up as a plague doctor, and looking through costume options reminded me of one of the worst pieces of doomscrolling churnalism that proliferated after covid. It is titled Get Ready for the Forever Plague, by one Andrew Nikiforuk, “an award-winning journalist whose books and articles focus on epidemics, the energy industry, nature and more”. ⊕ Of course, back in March 2020 he was just “an award-winning journalist who has been writing about the energy industry for two decades”. So it goes.

Such is the nature of echo chambers that he continues to write, putting out articles like this month’s As COVID Surges, the High Price of Viral Denial. At first glance they are meticulously sourced, a hyperlink to a peer-review journal underlining each claim:

COVID can even whittle away your intelligence. A recent New England Journal of Medicine study looked at the memory, planning and spatial reasoning of nearly 113,000 people who had previously had COVID. Almost all had significant deficits “in memory and executive task performance” regardless of the variant.

Alas, the linked NEJM article says no such thing. In fact:

Participants with resolved persistent symptoms after Covid-19 had objectively measured cognitive function similar to that in participants with shorter-duration symptoms, although short-duration Covid-19 was still associated with small cognitive deficits after recovery. Longer-term persistence of cognitive deficits and any clinical implications remain uncertain.

And as for the “regardless of the variant” claim:

The largest deficits in global cognitive scores were observed in the group of participants with SARS-CoV-2 infection during periods in which the original virus or the alpha variant was predominant as compared with those infected with later variants.

Crucially, the control group was people with no documented covid infection; we have no idea how covid-19 compares to other coronavirus infections, other viral infections in general, and even any illness requiring hospitalization. Staying in the ICU takes a toll regardless of what put you there, and last I checked covid has been putting fewer and fewer people in the hospital, let alone the intensive care unit.

This is a common theme for most covid-19-related research. Here, again, is Nikiforuk’s latest article:

No COVID infection is completely benign because each infection plays a role in deregulating the immune system. Even a mild infection, as one recent study noted, can increase “autoantibodies associated with rheumatic autoimmune diseases and diabetes in most individuals, regardless of vaccination status prior to infection.”

Two things here. One, autoantibodies associated with a disease do not imply a disease: I myself have had high titer for antibodies associated with Sjogren’s syndrome for more than a decade without ever having symptoms of the disease (how I found out about those antibodies is a story for another day). Two, note that the study compared autoantibody levels of three groups of people: those with long covid and persistent neurologic and fatigue symptoms (“neuro-PASC”), covid convalescents, and healthy controls with no known exposure. Ideally it would have included people with non-covid “neuro-PASC” and/or convalescents of other, non-covid viral infections. But at the very least it should have mentioned prior similar research in other viral diseases and put the findings in context of other viruses and hypothesis for autoimmunity. Presented like this, SARS-CoV-2 is a celestial bugaboo unchained from other parts of reality — no wonder that the lab leak hypothesis is so tempting!

Because there are two things that could be happening here. Either a humanity-ending event occurred somewhere near the end of 2021 and we are living a somewhat prolonged but inevitable decline in which so many people will have symptoms of long covid that civilization as we know it will end (queue “the Forever Plague”). Or maybe, just maybe, we experienced a once-in-a generation spread of a new virus — new to us but something humanity has had to deal with throughout its existence — at a time when we have the means to analyze its genome, our genome, its proteins, ⊕ Kudos to the Nature group of journals for their SEO. our proteins, the cells it infects, our cells responding to the infection, the microbiome, the food, the water, the air, the animals and yes, even art. And all that without the context of other viruses and other pandemics.

Last week’s EconTalk with Marty Makary featured several topics relevant to zombie medicine. One was a zombie’s return to the world of the living, with hormone replacement therapy for women not being as bad as we thought, particularly for preventing hot flashes in early menopause (before age 60). The other was the emergence of a new zombie: removing ovaries to prevent ovarian cancer when it is now thought that most ovarian cancers arise from the Fallopian tubes, not the ovaries themselves. I wouldn’t call it a full blown zombie just yet as there is an ongoing randomized controlled trial comparing the two approaches and who knows, its results may kill the old practice outright.