At last, at long last, the end of these may be in sight for coronavirus.

No, really.

–China is probably through the worst. There are no major blow ups happening elsewhere. There are no signs of alarm of particularly dangerous new variants in the sequences from Chinese travellers. By the end of next month (really, by Valentine’s, but we’ll be conservative), if there is not some new variant lift off in other countries with more realistic reporting, the chances that the Chinese outbreak spawned something new and concerning go lim x –> 0.

Which is what we expected, right? We said a few updates ago when China was facing a choice between serious social unrest on one hand and just dropping their COVID Zero nonsense to take the COVID death wave, and chose the death wave, that while an elevated risk for a more dangerous variant popping out, the risk was still small because the starting point would be an omicron cousin. Further, as we have said over and over through the outbreak, the general trend of a pandemic or epidemic pathogen is to become more contagious but less severe over time. Also, in the end, there are only a finite number of ways that SARS-CoV-2 can mutate the spike protein, and other proteins, to evade the immune system AND still work as a functional virus.

As far as the recent doom headlines warning of XBB.1.5 starting to dominate new cases of COVID in the US, well, XBB.1.5 is indeed now the most common variant in the US, accounting for 60% of new cases. As a reminder, this is backwards looking data by about 2 weeks and/or modeling (depending on whose breakdown you are reading) based on the rate of change in variant detection two weeks ago. Either way, XBB.1.5 has filled in the gap left by the original omicron, which is all but extinguished in the US. Two more recent cousins of omicron are still circulating, but they are also getting long in the tooth for a SARS-CoV-2 variant. Like their OG omicron source, they are probably running short on new hosts they can successfully infect.

XBB.1.5 is NOT kicking off some massive new wave of COVID though. Cases remain quite flat in the US. Early indicators are level to down. If XBB.1.5 was going to be a thing, it would already be looking like a thing. It is not looking like a thing. Further, all available clinical evidence suggests it is not more virulent than other omicron strains. I do still hear from readers who (re)-caught COVID in the past month, often after some travel. You can still catch it. It can absolutely still knock you on your ass. Omicron and its descendants are still pretty good at sneaking around existing immunity, from either vaccines or prior infections. Fortunately, the treatments for acute COVID still work amazingly well, certainly from all the anecdotes I hear about these breakthrough cases that qualified for them.

In fact, the “triple pandemic” of DOOOOOOOOOM (which was mostly flu, and the billionty other cold symptom viruses) is largely OVERRRRRRRRRRR:

As you can see, you had a brisk flu season (equivalent to the 2019-2020 season, so right before COVID), but it has predictably run its course. The most curious observation is that the peak came unusually early this year.

–So the only other COVID news really worth discussing is around the vaccines. The FDA has asked Pfizer to collect more data about a possible safety signal for stroke with their vaccine and booster. Moderna does not appear to have the same risk. The possible stroke signal is already rare, if real, and if not a class effect (given how similar to the mRNA vaccines are) is probably a statistical mirage anyways.

The FDA will also be debating the necessity and recommendations for further boosters. Some of the FDA panel, as we reported last update, were already critical of the data to “support” the boosters (as we were at the time, since they got approved without much in the way of human data at all). So this may be a little more spicy than usual.

But… on the level readers… per the CDC data as I type this, only 16% of the population has gotten a bivalent booster. The CDC also has blood donor surveillance data for COVID antibodies that runs through the summer of 2022. By May of 2022, 94% of the US population has antibodies to COVID, be they from vaccine or prior infection or both. By now, that has likely edged higher (and for those who are still tracking such things, I still see no sign of prior infection in my own antibodies, and still have a titer from my vaccines). One of the other FDA panel member criticisms leading up to the meeting is that it’s not clear from the evidence that the bivalent booster protected better against the omicron variants, lasted for a few months at most, and there is some data out there to suggest that the antibody response to the booster was just renewed antibody production and titers for whatever variants you had already encountered, and thus had memory B- and T-cells ready to go when you re-introduced spike protein via the vaccine. Speaking of which, we are still apparently back to largely ignoring T-cells and focusing on antibodies, but I (finally) accepted I am screaming into the void reminding all of us of our immunology courses where T-cells are also a critical, and arguably more important, line of defense to COVID.

All of which is a long way to say that with only 16% uptake of the bivalent booster already, I suspect that wherever the FDA committee comes down on seasonal COVID boosters, actual use of them will be pretty limited. And likely only to those with risk factors for severe COVID. Even that may eventually dwindle if we go a few years without major COVID activity and/or variant surveillance shows that the circulating variants just don’t pose the same mortality risk to high risk patients or “bed’s taken” pandemic threat to a broader population.

–Yes, I saw the Project Veritas hidden camera of the “senior Pfizer exec.” I’ve also seen the screenshots that Project Veritas posted on its Twitter account to verify the “Pfizer exec.” I can’t find him even as a contact of a contact on LinkedIn, so I am going to base all of this on what I have seen and on my 8 years in the pharmaceutical industry. This kid graduated medical school in 2018. He did a urology residency at Mass General, according to those screenshots. Which means he is -just- out of medical school. A “director” or equivalent title in a medical affairs type role is an entry level slot for an MD in pharmaceutical company. There do not appear to be any direct reports to him, and the screenshot of his organization strongly suggests his role is in pipeline management. Based on title and that of the others on his level (he’s one 6 directors, and not even one of the “senior directors” in his group) and his boss, his organization is likely doing nothing more than keeping tabs on the indications and molecules entering, as well as some evaluation of clinical trial success/failure odds and projected revenue to help the chief scientific officer, who his org kicks into, make decisions about what therapeutic areas to make investments in. I’ve worked with organizations like that. They talk us doing the research (occasionally), but have no oversight into what we are doing, and he is not running a lab.

“But the screenshots show he is just three runs from the CEO! He’s a director of mRNA research and development or something!” I hear you say, Hypothetical Reader.

Yeah, depending on the re-org of the month, I’ve been anywhere from 3 to 6 rungs from the CEO myself. I assume this is often true at Pfizer too. “Director” is almost without fail one of, if not the lowest, rung on the “management” track–not a research track.

So my initial take on the source here is that Project Veritas is trying to make this guy seem a bigger deal than I think he is. After all, I can find as a second degree contact on LinkedIn the “Vice President and Global Medical Lead for COVID-19 and mRNA Vaccines” at Pfizer (I don’t know the guy personally, for the record). I strongly suspect the VP has a much deeper understanding of Pfizer is doing on the science and their medical and sales strategies for the mRNA vaccines than the rando director in the pipeline management organization, even if the part of the pipeline that guy is responsible for includes mRNA vaccines.

And that’s before the more recent part of the video dropped, showing him flipping out when confronted by the leader of Project Veritas on camera, where our director said he was “making it up” to impress his date (possible) and “was not even a trained scientist.”

That second part I -absolutely- believe based on Project Veritas’ screenshots of his CV, his job title, and the screenshot of his part of the organization.

Regardless, our poor director has had a real bad time between the video confrontation and now all the news today. Whether he was making it up or telling the truth, either way, I strongly suspect he will be out of the industry entirely by the end of the year.

All of the HR out of the way, some of what our erstwhile director says Pfizer is doing, if true, is a big deal. So let’s cover that quickly.

The main claim of concern is the “directed evolution” of SARS-CoV-2, with the intent of mapping out the possible variants that could emerge, so that vaccines could already be in development against them, should they appear in the wild. If our director of pipeline is correct about what Pfizer scientists have discussed, or are possibly even doing, their plan is not to splice specific mutations into the virus, then see if the virus is made functional and/or worse by them. Instead, they appear to plan to just serially infect monkey populations with SARS-CoV-2, letting the virus naturally evolve to forms that can keep the serial infection going–presuming these are the mutations that would then be most likely to emerge in the wild. Then they isolate the viruses that “win” by continuing to infect new monkeys, and make vaccines to those. That strategy is plausible. Expensive–that’s a lot of monkey studies, and monkeys are not cheap–and slower than the alternative, but plausible.

In short, the difference here is that instead of creating a bunch of Franken-viruses in the lab with genetic engineering, then put them in monkeys to see which ones work you put the virus into monkeys first and let the virus mutate itself into Franken-viruses that work better by constantly forcing it to evolve to infect new rounds of monkeys. Since only the “best” viruses will survive. After that, you figure out, on the back end, the mutations the virus developed to keep infecting monkeys in new rounds. Then you develop a vaccine to those mutations.

Yes, you are STILL creating, deliberately, via the selection pressure you put on the virus, more potent virus strains. To me, this is still gain of function research in its net effect, even if it’s not called that. We have discussed before why I think gain of function research like this is a terrible idea, and Pfizer really should not be doing it (if they are).

Hopefully, they are not. Again, I’m not sure how much insight this director has into what is actually happening in Pfizer research labs. There is a non-zero chance this is just a bad game of “telephone” from what the actual researchers have said, or him just popping off on what they could be doing. He certainly claims it’s the latter.

But he does highlight the evil seduction of this kind of research quite directly, with the argument that they could have vaccines under development already, and thus speed clinical testing and deployment to variants they developed in the lab if and when they appear in the wild–if they can figure out those mutations ahead of time via gain of function research. This has ALWAYS been the argument for gain of function research–we make it worse in the lab, so we can already have the treatment in hand IF it gets worse in the real world. He also discusses the main risks of this, when he expresses his opinion that this kind of research in Wuhan getting out is likely the source of SARS-CoV-2.

I want to stress there is absolutely no reason to suspect a director in the pipeline group hired just out of medical school has any special, smoking gun knowledge that this is what happened, despite efforts online to spin it that way. (Confirmation bias to a preferred narrative is just so intoxicating, isn’t it?)

When discussing how Pfizer views the vaccine as a “potential cash cow” and how developing new vaccine versions to whac-a-mole the new variants on into the future could be very lucrative for Pfizer for years… I mean, c’mon guys. Does that surprise anyone? Pfizer is a for profit pharma company. All for profit pharma companies see themselves as a business first and foremost. They are in the business of solving medical problems with chemistry. That is all they do. -Always- keep that in mind. If it isn’t a medical problem with a chemistry solution, it’s tough to get a for profit pharma to care. They will also tout a chemistry solution even when a chemistry solution may not be the BEST solution. Always keep that in mind too.

To the point of “Pfizer is a for profit pharma company and they view themselves as a business”, the mRNA vaccines, especially with government mandates and government money buying them up to hand them out, have been amazingly lucrative for Pfizer and Moderna. Easily their best sellers through the pandemic. They have every -business- reason in the world to project that continued success and convince themselves that with new variants will come new ways to tweak their vaccines, and thus continue to get revenue from these products. They certainly HOPE they are not done after the first series and some boosters. Their c-suite execs would not be doing their duty for shareholders if they were not at least thinking about the potential for additional COVID vaccines if additional COVID variants emerge that require new vaccine versions.

With a 16% bivalent booster acceptance rate, and new variants being less severe to the point of causing many to reconsider the risk/benefit ratio of continued boosters, I’m not sure I share their optimism on the business side of this. But I’m one guy, and that’s not financial advice (I have no positions long or short in any pharma company nor any intention to open positions long or short any time soon).

I think our poor director in the clips, at the prompting of the Project Vertias agent (who is egging him on for the financial boon of the vaccines to Pfizer), is just spouting off the internal Kool-Aid Pfizer is drinking on the future of new COVID boosters. This is not indication of some nefarious plan by Pfizer to create new variants, release them, all to make more money off of boosters on the back end. Is there a lot of money on the table if that happens? Enough that “bioterrorism for profit” is possible?

Stupid, unethical, dangerous and evil things have been done for less money, I suppose, so sure it’s possible.

Probable, though? I doubt it.

First, there are the main arguments against bioterrorism in general. Our erstwhile director mentioned that Pfizer got surprised by the emergence of omicron and delta variants–they did not predict those particular mutations. Did you catch that? Think through the implication of that vis a vis bioterrorism for profit. So let’s say Pfizer creates a new, dangerous strain of COVID, and a vaccine against it in a top secret, Bond-villain dungeon lair. Bond-villain style, they “accidentally” release the virus, making it look like it was totes natural evolution. The world demands a new vaccine quickly. Pfizer already has one, and releases it just fast enough to not be too suspicious. Pfizer is a hero, profits and yachts for all! Huzzah!

Well, unless that new strain of COVID has another of those pesky unexpected mutations and goes right around the vaccine. Then kills -you-, Evil Bond Villain Pfizer Exec, before you ever get to buy that yacht, let alone party on it.

Yep, sad bikini-clad trombone indeed.

That whole “life-finds-a-way-and-your-genetically-resurrected-T-Rex-or-virus-comes-back-and-eats-you-first” is a MAJOR risk of bioterrorism, biological weapons (the “Dumbest of Weapons of Mass Destruction”), AND bioterrorism for profit.

Secondly, even if the virus doesn’t mutate to turn around and nuke them, there is equally existential risk to our Evil Bond Villain Pfizer Exec if it is ever discovered they hatched this nefarious plot. And way too many people would HAVE to know to be able to keep that secret long, let alone execute on that plan before the evil of this plan got out.

Hell, to show you just how difficult it would be to keep the top secret evil plan top secret, they didn’t even get off square one before you had the just-out-of-residency-director-of-pipeline spouting off about the whole thing in a restaurant to cameras. To “impress a date.” And now the internet knows!

Yep, sad Congressional-hearings trombone indeed.

Again, maybe my optimism and trust in fellow man bias, but I think when you map out the incentives and follow them, the reasons NOT to do bioterrorism for profit are just as, if not more, compelling than the reasons to do bioterrorism for profit.

To be VERY CLEAR, there is no indication in this “gotcha'” Project Veritas video that Pfizer is planning this. At worst, they are considering, if not doing, gain of function research on SARS-CoV-2 to speed booster design. Sorry, sorry. “Directed evolution.” But we repeat ourselves. While necessary steps to bioterrorism for profit, you can just as easily keep the vaccine anticipating new COVID mutations on the shelf, without accidental or deliberate release of a new COVID strain. That gain of function, even via “directed evolution,” makes accidental or deliberate release of new COVID strains more -possible- though is the reason Pfizer should knock that shit off if they are doing it.

The strategy that our director laid out is so expensive, and comes with no small degree of exactly this terrible publicity risk should some dumbass director of pipeline blah blah blah it to someone, that I’m not convinced, just by Project Veritas “gotcha’!”, that Pfizer is actually doing this.

But someone should ask them. And inspect the labs. And the notebooks. Just in case they are that dumb.

Lastly, if you did not already know that the revolving door between government regulators and the industries they regulate (pharma included) exists, and yes, presents potential conflict of interest… well, now you do. That said, there are definitely some regulatory heroes who never walk that road and fight the good fight for their careers to keep the industries they regulate as honest as possible.

–I could very easily have done another section this week on vaccine conspiranoia, ranging from the vocal minority who apparently really believe that the Bills safety died on the field and has now been replaced with a body double to keep the narrative. Assuming Devar Hamlin did, indeed, have commodio cortis, and not some genetic cause for his arrhythmia, he’ll be cleared to play again–and if that happens, I, for one, am going to be fascinated by the cognitive dissonance this will create in the “body double” movement. Finding a body double that can ALSO play football at a high enough level to be in the NFL is improbable to say the least. I’ve also seen any number of chart crimes from sudden experts in epidemiology. Some have mistaken Google search trends showing an increasing in the number of people typing “myocarditis” and “pericarditis” into the Google search bar for increases in the actual number of myocarditis and pericarditis cases. Others have found graphs of overall mortality in several countries before and after a mass vaccination campaign started in mid-2021 (when the vaccines typically became available in those countries). The overall mortality shows a spike right after the mass vaccination campaign was announced–often within a few short months. Yet, this is NOT the smoking gun claimed by those presenting these chart crimes. For, as you know from reading our updates, confounding variables abound. Namely, all of the countries I have seen online for this “argument” were getting slammed by the emerging delta wave right around the time vaccines became available–and delta was a relative killer, particularly for the pregnant and those with underlying risk factors for severe COVID. But most importantly, just because a mass vaccination campaign was announced in that country on day X doesn’t mean everyone in that country got vaccinated on day X. In many of them, it was a very gradual thing, because all of you are old enough to remember the developed world getting slammed for buying up most of the available vaccine stocks in 2021, so they were only trickling out for mass vaccination in these countries. Thus, the increase in overall death rate and COVID death rates that the new wave of conspiranoia has been putting all over the Twitter this past week in these countries is best explained by the delta wave getting them JUST as they were starting to roll out vaccine, but before much of their population got it. And thus most of the “post mass vaccination campaign” COVID deaths are in unvaccinated people–which our new epidemiologists always seem to forget to chart in a breakdown of death rate of vaccinated versus unvaccinated people in those countries and especially in those periods they are drawing attention to.

When you break them down by deaths in the vaccinated versus unvaccinated, as we have covered many times in this update, the death rate from COVID is ~10x higher in the unvaccinated with no previous COVID infection. This has been consistent in the literature and available data.

The vaccines unequivocally reduce the risk of death from COVID, and reduce the risk of severe COVID (the bed’s taken risk, which was the main pandemic threat). As the virus becomes LESS virulent though, so as omicron and its descendants become the most common variants, these benefits will APPEAR less pronounced. And like we said last time, it is very fair to begin to consider the benefit of the vaccine versus the diminished risk from these less virulent strains. That much I think we can all agree on.

Marginally better arguments have focused on the rise of overall death rates. Overall death rates were higher in 2020, trending up as the year went along (combination of COVID and knock on effects from lockdowns). They were markedly higher in 2021, particularly around delta waves, and again, as knock on effects from lockdowns. This was consistent throughout much of the world. The overall death rate is DOWN again, back to 2020 levels in the US. This is elevated relative to 2018 and 2019, however. The more skeptical among us have seized on this as a smoking gun for poor COVID vaccine safety. There is no difference that I have seen between the all cause mortality between the vaccinated and unvaccinated–and that would be more compelling. Presumably if it exists, it would be getting flogged harder. I suspect it doesn’t exist.

Instead, I suspect anecdotal stories like the one I saw today, about the Seattle Medical Examiner’s Office running out of morgue space this week because the rate of deaths from fentanyl overdose are double what they were last year. In fact, the rate of fentanyl overdose deaths in the US shows no signs of abating at all. Those deaths will cluster in the “young, healthy” age ranges in which the conspiranoia like to cite all cause mortality stats to imply the mRNA vaccines have caused long term damage to the health of those who received them. So too will suicides, car accidents. Auto accident fatalities increased by 10.5% in 2021 over the slightly higher rate already in 2020, according to the NHTSA. I have not seen data yet for 2022, but your anecdotal observation of bad and impatient driving? Yeah, it’s there in the stats too. I can only find data through 2021, but I doubt it will surprise you either to learn that rates of suicide have increased during the pandemic years. In short, there are other possible drivers for the rise in all cause mortality that have NOTHING to do with COVID vaccines.

Perhaps more to do with the pandemic malaise, that zeitgeist we have discussed–and so many of you readers have mentioned as feeling and seeing as well.

No one has come out of these years unscathed.

More socioeconomicky, but also going to throw some more infectious disease at you by the end

–I am sure I have said it before, but if I haven’t somehow, one of the most persistent medical AND social effects of this pandemic will be vaccination, as a whole, has been set back by several decades, if not a century.

Not just for “Devar Hamlin is now a body double because the vaccine (no one is sure if he even got or when) killed him” conspiranoia.

No, the uptake of the bivalent booster is 16.5%. That means a lot of people “tut tutting” at all those misinformed anti-vaxxer crazies aren’t exactly running out for that booster either. Again, I doubt any of you will even care what the FDA committee decides on future booster strategies. I expect the uptake to be abysmal.

What has set the cause of vaccination, as a medical strategy for all things viral/bacterial pathogenic, is the worst vaccination campaign in history. We said it at the time. We stand by it now. One of the most lasting, catastrophic failures of institutions and leaderships in this pandemic was to allow the politicization of vaccination. The schismogenesis, where politics is everything, and the political narrative trumps everything, and everything must be evaluated through a political lens (and a narrow one, at that, between two approved tribes), was allowed to wreak its destruction on vaccination. On -vaccination-, fer chrissakes. And even the institutions in charge of medical science were not immune, delaying release of data, overstating claims, making no effort to meet the patients where they were at and answer their reasonable questions in a time of crisis. To the point that at least the CDC has acknowledged its process broke and promised re-arrangement, and the FDA has started to become more critical of some of its processes.

But it’s too late. You have a -sizeable- die-hard population now that will not trust vaccines. Hopefully, they limit that to just mRNA vaccines (best of luck to any pharma company trying to expand the use of that technology for awhile; your sales projections are likely too rosy). But I doubt it. Which means that for some infectious diseases, you are likely in the coming years and decades to not reach a critical herd immunity threshold, and epidemics of vaccine controllable infectious diseases seem poised to increase.

Note I say “epidemic” not pandemic. These will be “pop up” outbreaks like you occasionally hear about for measles, mumps and the like, in places that either could not afford or achieve full childhood vaccination programs or had too many people in the community who were believers in the science of Jenny McCarthy. But I suspect they will be more common. Even worse, they will “prove” to this new die-hard population that vaccines don’t work as they do.

If there is a pandemic flu in our lifetime (not a huge risk, but again, it’s flu that scares me the most), God help us.

–Speaking of schismogenesis and the bonfire of the credibilities, came across something called the Edelman Trust Barometer. You can read their current report here. Business leaders and scientists come out relatively well, but globally, trust in most institutions has fallen (European Union and WHO showing gains in trust, interestingly). In fact, business leaders and business in general is held to be more reliable sources of information than the government or the media. Of interest is that the US is in the “severely polarized” range, suggesting differences between political groups are wide and potentially insurmountable. A huge range of countries, mostly developed, are in a range that is basically “at risk for severe polarization.” I do have some doubts about the results for “least polarized” countries which report strong belief in their government and little disagreement–most of those are authoritarian governments and call me cynical, unduly skeptical perhaps, but I think that folks in those places are less likely to believe a survey is really anonymous and their answers won’t come back to haunt them in police action or persecution of some kind. They do a case study on the US near the end that is worth perusing. Chief findings are that distrust in institutions in the US correlates strongly to social class, with higher income people MORE likely to express trust in institutions and low income least likely. Most troubling were the results of “those expressing a strong opinion” (which I read as those “most tribally aligned”) when asked if they would help someone of the opposite tribe or live next to them or have them as a co-worker. Only 20% would live next to someone they disagreed with politically, or work with them.

Most sad–only 30% would help someone from the other tribe if they were in need.

Guess it is no surprise that of the 27 countries surveyed, the global average was for 65% to say the lack of civility is the worst they have ever seen, and 62% say that the social fabric of their nation had grown too weak for unity and common purpose.

The sole beacon of optimism? If you dig down deep in that data, the percent reporting “trust in their fellow citizens” had increased.

Now, I will caveat all of this by saying it aligns with my anecdotal experience, and my impressions of the zeitgeist. So clearly it’s right and genius and not at all at risk of confirmation bias.

But if it is correct, that means schismogenesis has a significant lead. And we have a lot of work to do in ourselves and with each other–active and conscious work–to move beyond schismogenesis’ gravity and towards each other again.

Or the pessimists in this survey will be right, and we won’t make it.

–A few really good reader submissions this week worth bringing to the broader audience. First, ChatGPT was given the three written portions of the United States Medical Licensure Exam. This is the same test we have to pass to become licensed doctors (and the first two parts are usually required for graduation from medical school).

ChatGPT, the natural language AI program, was able to “comfortably pass.” (apparently it has also done this for bar exams, engineering exams etc.)

Am I about to be eaten by AI? Should I learn to code?

“Comfortably passing” is good enough to get called doctor, it’s true.  But it is worth pointing out that ChatGPT could only take the written portions—and those written parts are multiple choice.  The USMLE, at least until the pandemic, included an observed structural clinical exam, where you have to do observed histories and physical exams on trained and paid actors mimicking real patients, and then decide on your diagnostic and treatment plan.  That’s the real practice of medicine—getting the right history and symptoms out of a patient, establishing rapport so the treatment plan fits the patient and is one they can and will follow.

Ask me if the system is set up to foster that though.  Actually, don’t.  Pretty maudlin about the entire field of late.

Anyways, as it turns out, since interrupting that “live” portion of the USMLE during the pandemic, the sponsoring bodies for the USMLE decided to discontinue that part of the test (the Step 2 CS portion). My medical school included a mini-version of it where our attendings pretended to be patients–I suspect others do as well. If medical school accreditation does not already formalize this kind of evaluation as a requirement, I bet they will soon.

Regardless, I doubt ChatGPT was assessed for the truly practical part of the job that an OSCE attempts to evaluate.  That ChatGPT can call down from the great repository of all human knowledge that is the internet enough to pass a multiple choice medical knowledge test, though, is no surprise. But that’s not, and never really has been, a good way to evaluate a doctor. It’s just a way to make sure you can know and recall the basic package.

After all, every “patient” on the USMLE written portions walks in with a multiple choice list of conditions they could have or treatments you can try. I never had a patient walk in with 5 convenient options to choose from, one of which HAS to be right. Meanwhile two are clearly wrong, so it’s really just parsing among these three that are kind of close…

Real patients are more open ended. Real diseases often don’t read the textbook either, in quite the way they do to hit the “keywords” on multiple choice exams.

That said, specialized AI for express, limited purposes will definitely play a useful role in healthcare going forward. Even ChatGPT may play a role. I can see ChatGPT or similar as a screener for your visit, asking you the basic questions and getting basic follow up, so I can be more efficient during the visit. Or it could be a way for me to double check a differential or more rapidly search if it seems to me you may have something quite rare.  A generalized AI that can do the entire job of being a physician, and do it well, I am less certain about—especially seeing AI development and limits up close in the “day job” these days.

Also worth remembering Watson, IBM’s supercomputer AI, crushing all the best Jeopardy champions over a decade ago. Those games weren’t close. AI passed humans in chess long ago, and “Go” more recently. Feel free to take on the “Mittens” bot on chess.com if you have a few moments to be ruthlessly crushed by an AI that isn’t even the best at chess. The best chess player in the known universe is the AI program “Stockfish” and Stockfish’s real-time analysis of positions is used to analyze who is winning games between grandmasters as part of the commentary on live tournaments now.  I have seen a few “super grandmasters” (top 10 human chess players in the world) fight “Mittens” to a draw—but they can’t actually win against even this lesser chess AI without Stockfish helping them in key moments or starting a pawn or piece up.  But games like these are what computers are good at, while we tend to mistake being good at chess or Jeopardy! as a sign of generalizable intelligence. While parts of trivia knowledge, and some principles you can learn from chess are metaphors for other problems and aspects of life, the key is that the rules and goal (how to win) in Jeopardy! and chess are known and fixed and importantly never change during the game. Not all problems in life are like that.

Watson was partnered to large Ivy League health care systems shortly after to great fanfare, certain to revolutionize health care for it too could do ChatGPT’s trick in terms of summoning and contextualizing human knowledge in health care scenarios. Had demonstrated that in Jeopardy! no less.

Watson was pulled from healthcare work by IBM a few years ago to far less fanfare, as results never quite matched the promise.

More recently, there were a plethora of pandemic era “AI predicts new variant/new combos/new risk model for progression/potential new treatment” papers.  With the pandemic winding down, none are in common use.  The current treatments and vaccines came from quite conventional, empiric vaccine, antibody and drug development.

It’s not just healthcare where AI has struggled either. You can check out this Axios article to see where CNET and Bankrate stopped publishing AI generated articles they had been posting on their websites last week because the AI was including inaccuracies. But it also mentions other outlets that have found successful use cases for AI in journalism, albeit in more specific, smaller tasks like image sourcing. Writing an entire article de novo is a more general task, and ChatGPT only looks like a genius at that because it is good at splicing together what already exists on the topic on the internet, then transforming it into a more general readable format. The less that already exists on the topic where ChatGPT can find examples of it, the more ChatGPT will predictably struggle.

Similarly, where AI has had success in healthcare is limited, specific functions that are highly repetitive. These limited function AI provide decision support, such as warning of sepsis by checking key vitals and labs automatically in the background, starting treatment protocols where the disease is common and the treatment cookie cutter, avoiding drug-drug interactions, and reminders for hospice vs last full measure discussions with poor prognosis cancer patients. You also have AI that can improve image analysis, with many successes in radiology and yes, to our own modest extent, pathology. Limited, task specific AI can focus radiologists and pathologists to areas of likely interest on the images our specialties use so much.

The future of AI is man plus machine, not machine replacing man.  Let the computers do what they are good at—well defined tasks, which are repetitive and predictable with well defined outcomes.  Let man focus on “wicked problems,” where AI applications might provide -part- of the solution, but human wisdom and intuition is necessary to decide when, where and how.

–Also received an article about noma. Noma is an infectious disease you have almost certainly never heard of, as it has the terrible misfortune of being a disease primarily of poverty or extreme immunosuppression. Don’t worry–there will be no noma pandemic. Places with high incidence of malnutrition have noma. These are not places that get a lot of news coverage though, and since it’s not a disease of the developed world in any significant extent, it’s almost completely forgotten. Fortunately, noma is still rare even in the areas where conditions are ripe for it.

And yes, it happened in concentration camps too. Anywhere you get extreme malnutrition, there’s a risk.

The exact pathology of noma is not entirely well understood. Because again, no one really studies it. What appears to be going on is that severe malnutrition weakens the immune system. The normal bacterial flora of your mouth becomes less diverse, and more homogenous. Then it starts eating the mouth and face from the inside out. Literally. For those who survive (and there is very high mortality from this), the disfiguring scars are permanent and hideous.

Do NOT Google images of noma.

Of course, having told you NOT to do this, you went ahead and did it, didn’t you?

Curiosity and cats, right?

Well, you’ve learned an important lesson then. When a pathologist, especially, tells you something medical is “hideous” and “do NOT Google images”, you should consider that advice…

Surgery can correct some of that, but only if the person is lucky, and even then, there is often only so much reconstructive plastic surgery can do.

Frankly, that noma even exists is further condemnation of a world that can produce enough food for everyone, but has never solved distributing food to everyone so that no one goes hungry or malnourished.  With the likely food inflation resuming this year, continued war in the Ukraine which seems likely to impact food again this summer, AND persistent supply bottlenecks on fertilizer, there is a non-zero chance that noma cases will rise. You won’t hear about it though, because the people impacted will largely have the misfortune of being in the wrong place, earning too little money for anyone to pay attention but the medical volunteers who go over to try to heal the survivors as best as can be.  We’re lucky even then that noma is uncommon to rare, but at the same time–noma is utterly preventable with proper nutrition, and better care for the severely immunosuppressed.

–Before we all get too bummed though, take a moment to remember our own personal history and that of humanity. How many crises and potential world ending political events/natural disasters etc. have the news screamed at you about so far? And yet, here we all are! Somehow, some way. Never forget that you are the sons and daughters of the people who had to put up with their own forms of political crises, their own intertribal conflicts, malnutrition, food crises, climate crises, and various forms of natural disaster–without modern medicine or modern energy or transportation.

While fighting saber toothed tigers.

With sharpened sticks and stones.

We may have 99 problems, but they are ALL first world problems compared to what some of our ancestors were dealing with.

They got through. Built better. So can we. So will we.

–Finally, your chances of catching coronavirus remain equivalent to the chances that I will start spacing these out even more as COVID drifts into contagious, but less virulent strains, all of which respond well to the acute treatments. And there is just less and less to say as the pandemic part of things drifts inexorably into the rearview.

–For the OGs, this means the Ramble will be back later this year. A little less random, a little more focused, based on feedback from the socioeconomic sections in particular over these years. More to come on that as it gets closer.

<Paladin>