A detailed and brutal look at the tactics of buzzy AI compliance startup Delve
"Delve built a machine designed to make clients complicit without their knowledge, to manufacture plausible deniability while producing exactly the opposite."
https://substack.com/home/post/p-191342187
Hot take: HIPAA hinders patient care
AHA in 2010: " We can't make money just taking care of sick patients. We need elective, commercially insured patients. Shut down these damn doctor owned hospitals."
AHA in 2021:"We had record revenues in spite of the pandemic throttling high margin elective cases. Looks like we figured out how to make money caring for sick patients."
AHA in 2023:"Yeah so even though we had record revenues post pandemic, and our C suites got crazy production bonuses; we lost money on our balance sheets because our investment arms took a bath in the market. Can you guys at CMS maybe give us a pay bump to offset our losses?"
AHA in 2026:"We can't make money taking care of sick patients. We need elective, commercially insured patients. Don't repeal the ban on physician owned hospitals. Doctors are just greedy interlopers anyway.
Hey can I show you my new Maybach?????"
@GeBaiDC @DrDiGiorgio @DrBruggeman @DutchRojas @anish_koka @nickshirleyy @DrOz
If you want to know exactly how it lies about real journalists when they threaten it with too much truth (by "it" I mean "the science") here is a vivid case study.
Dear Kashmiri Hindus, I want to wish you happy Navreh but every time I think of you, I think of the hollowness of my wishes, I think of how we failed you, I think of how 700,000 of you still cannot return to your homes, and so please accept not my wishes but my apologies.
Sorry.
R to @anish_koka: VESALIUS was carefully designed with the 3, 4 point composite outcome and hierarchichal testing precisely because trial designers knew a lower risk population than 4S, FOURIER, ODYSSEY would have far fewer cardiovascular events. Fewer cardiovascular events means a trial has to be much larger to show a statistically significant benefit.
The size of clinical benefit is inversely proportional to the size of the trial needed to demonstrate statistical benefit. If an RCT has 100,000 patients in it, the size of the effect you are looking for is extremely small OR the trial designer was drunk/high.
The 4S trial which had patients that had survived a heart attack with high LDLs ? - the control arm had a 11.5% CV mortality. VESALIUS control arm had a 3.6% CV mortality.
Consequently the VESALIUS trial had 12,257 patients , the 4S trial had 4,444 patients.
Bottom line: VESALIUS is designed to be a positive trial.. and its hard to blame companies that are spending billions of dollars to run these trial.
It was an honor to testify in front of the @HouseCommerce subcommittee on health regarding healthcare affordability.
We discussed consolidation and the demise of independent physician practice.
My solutions include:
Repeal section 6001 of the ACA which banned physician owned hospitals
Reform Stark law
Implement site neutral payments
Reform 340B
Use FMAP to encourage states to be pro-competition (repeal CON, eliminate non competes)
PSA: Not Noninferior = Inferior.
Great post. I walked away in '93 and never looked back. GKS
NOT NOINFERIOR and INFERIOR
R to @anish_koka: Some background --
Why do hierarchical testing ? (any statisticians / trialists reading are welcome to correct/clarify)
The two word answer : Multiple Comparisons.
Every time you run a statistical test at p<0.05, you're accepting a 5% chance of a false positive, otherwise known as finding an effect that isn't real. Otherwise known as a Type 1 Error.
Basically, if you test 10 endpoints in the same trial, each at p<0.05, your chance of getting at least one spurious significant result by chance alone rises to about 40%. Test 20 things and it approaches 64%. This is the multiple comparisons problem.
Without any correction, a trial could throw everything at the wall and call whatever sticks "significant." This is commonly referred to as a fishing expedition.
Hierarchical testing is one way to keep the overall false positive rate at 5% across all the tests you plan to run.
Pretty simple logic : you only get to test the next hypothesis if the previous one passed. Because each test is conditional on the prior one succeeding, you never spend more than 5% alpha total across the entire chain. No penalty needed for multiple testing as long as you follow the sequence.
Back to VESALIUS ..
R to @anish_koka: After both co-primaries passed, the secondary endpoints were tested in this fixed sequential order:
1. MI + ischemic stroke + ischemia-driven revascularization — HR 0.79 ✓
2. CHD death + MI + ischemia-driven revascularization — HR 0.79 ✓
3. CV death + MI + ischemic stroke — HR 0.73 ✓
4. CHD death + MI — HR 0.73 ✓
5. MI alone — HR 0.64 ✓
6. Ischemia-driven revascularization — HR 0.79 ✓
7. CHD death alone — HR 0.89, p=0.39 ✗ — chain breaks here
R to @anish_koka: The design of the trial was interesting
-Two co-primary endpoints were tested simultaneously.
- 3-point MACE: Death from coronary heart disease + myocardial infarction + ischemic stroke
- 4-point MACE: 3-point MACE + ischemia-driven arterial revascularization
The choice to have two co-primaries rather than one is meaningful. 3-point MACE is the harder, more clinically meaningful endpoint — it excludes revascularization, which is partly operator and system-dependent. 4-point MACE is softer and easier to move because revascularization is more frequent and more influenced by physician decision-making than by hard events.
Having both as co-primaries gave the trial two shots at success and ensured that even if the harder 3-point endpoint had been borderline, the softer 4-point likely would have carried it through. (Both were positive)
R to @anish_koka: This tests PCSK9 inhibitors in a lower risk group of patients - No prior MI, but 50% had evidence of coronary disease or cerebrovascular disease (without a clinical hx of stroke or MI) .. so this is testing PCSK9inhibitors for primary prevention as opposed to FOURIER and ODYSSEY which were secondary prevention trials.
R to @anish_koka: The PCSK9 inhibitor class of medications have demonstrated benefit in the FOURIER and ODYSSEY trials that included patients who already had a heart attack or an ischemic stroke.
These were high risk patients, but they were already on statins with an LDL that was ~90mg/dl to start. The benefit with regards to mortality was pretty small - some of the lowest benefit seen of all of the trials done to date with lipid lowering despite the fact that the average LDL in the treated group was ~50.
On the heels of this comes the VESALIUS-CV trial
R to @anish_koka: As time progresses (see legend) the mortality reductions seen in the statin trials attenuates for a # of reasons
1. In an effort to broaden indications for use, lower risk populations are chosen
2. Subsequent trials have patients already on statins and consequently have much lower baseline LDLs to start with (FOURIER and IMPROVE-IT at bottom left) have starting LDL of ~90 and did not show a clinically significant decrease in mortality despite ~50% LDL lowering
R to @anish_koka: To date, lipid (LDL-targeted) RCTs have demonstrated reductions in cardiovascular events in a variety of settings. The great grandfather 4S trial in 1994 showed that simvastatin used in patients who had just had a heart attack, who on average had an LDL of 189, reduced cardiovascular mortality and overall mortality by 2.9% -- meaning you would need to treat ~35 people to prevent one fatal event.
I cannot speak more highly of @DrDiGiorgio's testimony before House E&C this morning: "The system effectively starves independent practices of revenue while burying them in paperwork, making selling to a hospital the only viable way out."
Full testimony: https://democrats-energycommerce.house.gov/sites/evo-subsites/democrats-energycommerce.house.gov/files/evo-media-document/digiorgio_testimony_he-hearing.2026.03.18.pdf
1) After I wrote a @bmj_latest investigation on virologists conspiring to label critics "conspiracy theorists" my reporting became a focus of NIH officials and researchers they funded.
"Really disappointed they gave space to this scheister," wrote Peter Daszak of my BMJ piece.
The inequality causes crime narrative is activist science. 43 studies. 1,341 estimates. Half the data never published.
Corrected effect: near zero.
Inequality doesn’t drive crime.
If this can happen, the presidency is not real. Neither is the quadrenial plebiscite, nor the idea that the people are in charge at all. It's all fake. https://www.cnn.com/2026/03/17/media/voice-of-america-staff-reinstated-trump?Date=20260317&Profile=CNN&utm_content=1773791351&utm_medium=social&utm_source=twitter
Berenson rarely misses.
The only good thing about all this is that the country is now aware a managerial elite who’s salary you pay controls all the major institutions and has run the country for some time.. you were simply never aware of it.
Yeah. To see former dean’s of public health now say consolidated health systems are the problem after writing tomes on the value of integrated clinical networks.. is really something.
This isn't rare. I have seen in academic centers, community centers, and everything in between!
Wish we could use CAC and reliably stop with medications. However, all to often we end up with further imaging and even invasive procedures, even among asymptomatic patients.
RT @txsportsdoc: With the new @nickshirleyy Medicare/Medical fraud video dropping in Los Angeles CA it’s even worse than MN. CA has highest…
Med Students yesterday:
"After 10 years, including a gap year and 2 years in a non ACGME accredited fellowship, I'm pleased to announce I've matched in Neurosurgery. Happy to serve."
Van Nuys Home Health Care operators:
"Lulz. Let me show you my new Maybach"
CMS UPIC Contractor Directory - Interactive Map | It would actually make a whole lot more sense if @DrOzCMS
was leaning on his fraud prevention contractors and Hospice falls right into UPIC territory, considered high risk http://cms.gov/data-research/…
Makes more sense versus trying to emulate juvenile Nick Shirely with X videos..besides Oz is there for the PE opportunites, his son was a PE investor, and they created ZorroRX..Oz is buddying up with PE Bill Frist - they post this stuff themselves as giant narcissists
ZorroRX is a health benefits company co-founded by Dr. Mehmet Oz and his son, Oliver Oz. The company, which connects patients to the federal 340B Drug Pricing Program, was established in 2024 to help hospitals and employers reduce costs. Oliver Oz serves as the Chief Operating Officer
And then there's the CMS antifraud contractor, UNH IDIQ contractor..Optumserve, hiding in plain sight for those who choose to read..and not be groomed into alternative perceptions..
@DutchRojas
@DrDiGiorgio
@mass_marion
@anish_koka
@real_doc_speaks
This is the shell game I speak of.
The geniuses who make healthcare rules disallowed physicians from running hospitals when they passed Obamacare because of “misaligned incentives”, but “anyone” can open a home hospice / healthcare agency in California.
Enter the Armenian mafia.
Monty python wld like a word.
There’s an industry with an interest in putting you on meds evidenced by dropping target LDLs as new therapeutics to lower LDL hit the market.
There is also an industry with an interest in selling you a path to get off of meds.
Things to consider for a 40 year old with an elevated LDL and CAC of 0. Depending on your other risk factors that is not necessarily a get out jail free card.
Crude cardiac mortality (not adjusted for clinical RFs, so lumping as an example bmi 30 smokers with iron man non smokers) for a 40 yo is 1/2000 so we are not hunting for zebra’s in Alaska.
Numerically there are still a lot of events in calcium zero folks — perhaps there is a phenotype of lean mass hyperresponders ( metabolically healthy except for LDL) that have low rates of cv events.. definitely deserving of more study. Would need long term follow up, and wld be nice to have a better surrogate of longer term cv outcomes.
Also Need much better clinical validation of current tools that say they measure plaque well.
The program was championed by Sanjay Gandhi, youngest son of Prime Minister Indira Gandhi , Scion of the Nehru dynasty that ruled India post British Independence.
The issue is that it is rarely straight to cath!
I frequently see CAC ➡️ stress (it's in AUC!) ➡️ false positive since most nuclear stress in US is done without attenuation correction ➡️ cath ➡️ PCI
Can also happen with CCTA as interpretation is harder with calcium blooming
Either way, asymptomatic patient has both financial and potentially medical and/or procedural toxicity along the way
The only reason to do this is so universities can hide their racist admissions policies in defiance of the Supreme Court decision.
Ever wonder why these costs are so high? What policies are driving it?
Tune in to the US House hearings Wednesday, 3/18. Link below:
I’m honored to testify before the @HouseCommerce Health Subcommittee on March 18 as Congress examines the background, scale, and consequences of rising health care costs.
Health care affordability is one of the defining policy failures of our time.
Frame this one. Frame it.
None of this "Free Palestine” activism in Hollywood is sincere. They do it because it’s an “in thing” and wins approval from the young crowd.
Don’t believe me? A reminder from recent history helps. In the 1990s and early 2000s, “Free Tibet” had an even bigger celebrity endorsement machine. Hollywood stars, musicians, concerts, award speeches, support for Tibet and the Dalai Lama was loud across the entertainment industry.
And now, Free Tibet is nowhere discussed. Not because the issue was resolved. It went silent because China became an indispensable economic power and market for films, music, and brands. Angering China carried real financial consequences, movies banned, concerts cancelled, market access lost. So the moral fervor disappeared.
That should tell us something about celebrity activism. Today they say “Free Palestine” because there are no financial consequences. The day that changes, it will fade like “Free Tibet.” What looks like moral conviction is just aligning with whatever cause is trending.
CIA Chief of Station in Beirut William F. Buckley was brutally tortured at the hands of Hezbollah for fifteen months. In early June of 1985, he was buried in a shallow grave on the side of the road in the Shiite section of the city. His body would not be returned to the United States until December 1991.