The Business of Sickness

For years I've listened to people, myself included, treat the cost of a hospital bill as the whole scandal. It isn't. The number on that bill is the visible symptom of a payment system that rewards volume over outcomes, and once you follow the money instead of the sticker price, a very different picture of American healthcare shows up.

This investigation traces that system through three specific places where it stopped being abstract and started costing lives: the routine flow of pharmaceutical money into physician prescribing decisions, the Sackler family's deliberate exploitation of that flow during the opioid crisis, and the reimbursement questions that surfaced around COVID diagnoses. Every factual claim below is sourced. Where this piece moves from documentation to my own judgment, that judgment is flagged as opinion.

Part One

The Premise: A System Built on Volume, Not Health

I want to start with the plumbing before I start naming names, because the plumbing is what makes everything downstream possible. American healthcare does not pay doctors and hospitals to make people healthy. It pays them to do things to people, tests ordered, procedures performed, prescriptions filled, and it pays them again the next time, regardless of whether the last thing worked. That model has a name, fee for service, and it has been the dominant payment structure in this country for most of my lifetime.[1]

Under fee for service, a doctor who spends twenty extra minutes talking a patient out of an unnecessary procedure gets paid less than a doctor who performs it. A hospital that keeps a patient healthy enough to never come back gets paid less than one that sees them again in six months. I am not describing a conspiracy in this section. I am describing an incentive structure, the kind economists study plainly and openly, and the kind that shows up in the data regardless of who is in the White House or which party controls Congress.

Researchers at JAMA have estimated that low value care, treatment that provides little to no benefit relative to its cost or risk, accounts for a meaningful share of total American healthcare spending every single year.[2] KFF and the Commonwealth Fund have tracked the same pattern from the payer side for years, spending keeps climbing, outcomes do not climb with it, and the gap between the two lines is where this piece lives.[3]

None of this requires a villain yet. A hospital administrator balancing a budget under fee for service is responding rationally to the rules in front of them, the same way a landlord responds rationally to rent control, the same way a tariff-protected industry responds rationally to a tariff. Opinion But rational responses to bad incentives still produce bad outcomes, and pretending otherwise is how this system has survived four decades of bipartisan promises to fix it.

What I want to do in this investigation is trace that incentive structure through three specific places where it stopped being abstract and started costing lives, the routine flow of pharmaceutical money into physician prescribing decisions, the Sackler family's deliberate exploitation of that flow during the opioid crisis, and the reimbursement questions that surfaced around COVID diagnoses. Then I want to ask what happens to the price of American healthcare once the incentive to overtreat is actually removed, instead of just regulated around the edges.

Part Two

The Money Behind the Prescription Pad

Before I get to the Sacklers specifically, I want to show that what happened with Purdue wasn't an aberration built by one unusually ruthless family. It's an extreme version of something I found happening across the entire system, every day, on a smaller and far less scrutinized scale.

The mechanism is called the Physician Payments Sunshine Act, and it created the CMS Open Payments database, the federal system that requires every drug and device company in the country to report, publicly, every payment they make to a physician. The database tracks payments that drug and medical device companies make to physicians, physician assistants, advanced practice nurses, and teaching hospitals. Over the most recent seven-year period for which CMS has published cumulative totals, the program has published a total of 84.05 million records, accounting for $72.80 billion in payments and ownership and investment interests.[4]

Not to a handful of bad actors, across the entire prescribing physician population of the country, in consulting fees, speaking fees, meals, travel, and research funding.

Here's what that money actually does. A team of researchers examined pharmaceutical marketing payments related specifically to opioids in 2014, then measured what happened to those same doctors' prescribing in 2015.[5] Doctors who received any opioid-related pharmaceutical marketing wrote nine percent more opioid prescriptions the following year than doctors who received none. The payments driving that shift weren't, for the most part, large. The median payment value was thirteen dollars, pointing to the potential influence of even small payments on prescribing behavior. A branded pen and a sandwich moved the needle on how many opioids a doctor prescribed.

A separate, larger study linked the same Open Payments data to actual Medicare prescribing records between 2013 and 2015 and found that physicians who began receiving opioid-related payments saw a significantly larger increase in opioid expenditures, in daily doses dispensed, and in expenditures per daily dose relative to physicians who received no such payments.[6] I want to be fair to the researchers' own caveat here rather than bury it. The study's cross-sectional design demonstrates only correlation between payments and prescribing, and it's possible manufacturers simply market to physicians who already prescribe higher opioid doses. Correlation isn't proof of corrupt intent in any single doctor's office.

But at scale, this pattern doesn't look like coincidence. Among the 369,139 physicians identified as opioid prescribers under Medicare Part D in 2015, seven percent had received a total of over 105,000 opioid-related pharmaceutical company payments the prior year, amounting to more than nine million dollars. The single company with the highest payment total by far was Insys Therapeutics, at over 4.5 million dollars.[5] That name isn't introduced casually. This is the same company whose founder was later convicted in federal court of racketeering for bribing doctors to prescribe its fentanyl spray.[7] The pattern that shows up quietly across the whole industry, small payments correlating with more prescribing, is the same pattern Insys ran openly and criminally, just at a scale that eventually caught federal prosecutors' attention.

Opinion This is the part of the story that rarely gets the outrage it deserves, because seventy two billion dollars spread across a million doctors doesn't produce a single villain a news camera can stand in front of. It produces something worse, a background hum of financial influence so normalized that most of the physicians accepting it would tell you, honestly, that it doesn't affect their judgment. The data says otherwise, and the data isn't interested in anyone's self-perception.

This isn't confined to opioids either. Payments tied to brand-name diabetes and cardiovascular drugs have shown the same association with higher prescribing rates, despite lower-cost generic alternatives being available with no evidence of additional clinical benefit.[6] Opioids just happen to be the category where the consequence of that influence got measured in overdose deaths instead of an insurer's spreadsheet. Which is exactly where the Sackler family enters this piece, not as an isolated monster, but as the most aggressive and most consequential operators inside a system that was already built to reward exactly what they did.

Part Three

The Sackler Blueprint

This is the section where I stop describing a system and start naming the family that exploited it most aggressively.

In August 2018, Massachusetts Attorney General Maura Healey did something no state had done before, she filed a civil complaint that named individual members of the Sackler family personally, not just Purdue Pharma as a corporate entity.[8] The complaint charged the family and company with systematically deceiving physicians and patients about the safety and addictive nature of OxyContin. Purdue fought to keep the supporting memorandum sealed, and when a redacted version came out that December, more than half of it was blacked out. Healey refiled it weeks later, largely unredacted, and what it showed wasn't a company that failed to anticipate what its own product would do. It was a chain of command implicating eight named Sackler family members and nine Purdue board members or executives directly in the opioid epidemic.

What stands out most, reading through the Sacklers' own internal communications, isn't Healey's characterization of them. It's the family's own words. In 2001, as OxyContin abuse was becoming a national story, Richard Sackler, then Purdue's president, wrote internally that the company needed to "hammer on abusers in every way possible," calling them "reckless criminals."[8]

That wasn't a one-off outburst either. The complaint alleges Sackler-approved marketing plans labeled people harmed by opioids as "junkies" and "criminals," a deliberate strategy to distance the family from mounting research on OxyContin's addictive properties. And while that public denial was happening, company officials were privately acknowledging the addiction risk and working out how to profit from it further rather than curb it.[8]

The sales pressure documented in the memo was relentless. Purdue sales reps visited Massachusetts prescribers and pharmacists more than 150,000 times over eleven years, and Richard Sackler tracked the results personally. When one week of prescriptions doubled Purdue's own internal forecast, his response to the sales staff wasn't caution. It was "I had hoped for better results."[8]

Then there's the money, and it's worth flagging now because it echoes directly into Part Seven, when this piece gets to COVID reimbursement. In 2007, Purdue pleaded guilty to a federal felony for misrepresenting OxyContin's addictive potential.[9] The very next year, Richard Sackler suggested the family either sell the company or, failing that, extract as much free cash flow from it as possible for themselves, and they chose the latter. The complaint states the family voted themselves a 250 million dollar payment at a June 2008 board meeting, followed by another 199 million dollars that September, in the immediate aftermath of pleading guilty to a felony for the exact conduct still killing people.[8]

It's worth being precise here about what's allegation and what's admission. The dollar figures, the sales visit counts, and the internal emails come from documents Purdue itself produced or that were compelled through discovery. They aren't Healey's invention. Purdue and the Sacklers disputed her legal characterization of that conduct as unlawful, but the underlying documents themselves were never seriously contested as authentic. The emails are real. Whether they amounted to the specific legal violations she alleged was, at the time of filing, still a question for the courts.

One more thing matters here. In 2021, Purdue disclosed in bankruptcy filings that it had commissioned a 22,000 hour internal investigation into its own owners, run by outside lawyers and forensic accountants between April 2019 and early 2021, and then released none of the findings.[10] Healey called it what it looked like: "They're still trying to cover up the facts." The Sacklers have consistently and publicly denied any wrongdoing.

Opinion I don't think that denial and the documents can both be fully true. A family does not commission a 22,000 hour investigation into itself and then bury the results if those results were flattering. Part Five carries the weight of what accountability actually looked like once the legal process played out, because what happened there speaks for itself.

Part Four

What "Less Than One Percent" Actually Cost

I want to trace one specific number through this story, because I think it explains more about how the opioid crisis actually happened than any other single fact in this research.

When OxyContin came onto the market in 1995, Purdue's sales force promoted it for lower back pain, arthritis, trauma, fibromyalgia, dental procedures, broken bones, sports injuries, and post-surgical pain, essentially everything. And they built that promotion around a single claim, repeated to doctors over and over, that less than one percent of patients would become addicted to the drug.

That number didn't come from a study of OxyContin at all. It traced back to a misquoted letter, a five-sentence observation published in the New England Journal of Medicine about addiction rates among hospitalized patients monitored under close medical supervision, not a claim about outpatients taking a powerful opioid home with them for weeks or months at a time.[11] Purdue's sales reps didn't cite that context. When physicians raised concerns that a patient's pain was returning before the twelve hour mark, or that the medicine wasn't lasting as advertised, representatives were trained to reassure them that the risk of addiction was "less than one percent."

Here's the most damning part. Purdue produced roughly 14,000 copies of a promotional video making that exact claim, that fewer than one percent of patients became addicted, and the FDA later said the claim "has not been substantiated."[12] Government investigators found the company hadn't even submitted one of these promotional videos to the FDA for review in 1998, as required by law, before circulating it to thousands of doctors. Purdue called it an oversight. The FDA didn't review the 2001 version it did receive either, because the agency said it lacked the resources to do so. That doesn't absolve Purdue of what it put in front of doctors, but it does show a regulatory system that wasn't equipped to catch what was happening even when the company technically followed the rules.

The financial scale behind that one percent claim is what makes it more than a marketing footnote. Marketing files later showed Purdue spent more than 400 million dollars on OxyContin promotion, with roughly 75 percent of that spending occurring after 2000, the same year company officials told Congress they'd already learned of growing abuse and drug-related deaths.[12] Purdue spent more than 18 million dollars advertising in major medical journals, ads that federal officials later said "grossly overstated" the drug's safety, and poured more than 8 million dollars into a website and referral network called "Partners Against Pain," connecting patients directly to doctors willing to prescribe it. This wasn't a company slowly realizing it had a problem. By Purdue's own admission to Congress, the spending accelerated after the company already knew.

Purdue's defenders deserve their strongest version of the counter-argument here, because that's what fairness requires. Purdue has pointed out that its FDA-approved products account for only about 2 percent of all opioid prescriptions nationally, and the company has argued that attributing the entire multi-faceted opioid crisis to one company's marketing from over two decades ago is "deeply flawed." That's a real argument that shouldn't be waved away. Other manufacturers, other prescribers, and a broader medical culture that had already been primed to see pain as an undertreated crisis all played a role too, which is why Parts One and Two covered the broader system rather than resting this piece entirely on one company.

But the "less than one percent" number wasn't a rounding error or an industry-wide assumption Purdue merely repeated. It was a specific, misquoted, unsubstantiated statistic that Purdue built an entire national sales campaign around, distributed to physicians in FDA-unreviewed video form, after internally acknowledging in 2000 that the drug was already being abused. In 2007, that gap between what the company said and what it knew became a federal matter. Purdue Pharma and three of its current and former executives pleaded guilty to federal charges of misleading the public about OxyContin's addiction risk, paying 634.5 million dollars in fines, and prosecutors said Purdue had learned from its own 1995 physician focus groups that doctors were already worried about the drug's abuse potential, then told its sales force the opposite.[13]

Opinion That focus group detail is worth sitting with. Purdue didn't stumble into the "less than one percent" claim by accident and then fail to correct it fast enough. The company's own research told them doctors were worried about exactly the right thing, and the sales strategy was built specifically to talk them out of that worry. That is not a company that got the science wrong. That is a company that knew the real answer and sold a different one because the different one was more profitable.

Part Five

The Settlement and the Shield

Purdue filed for Chapter 11 bankruptcy in 2019, and by 2021 a bankruptcy judge had approved a plan that, on its face, looked like a reckoning. The deal would have directed 6 billion dollars to state and local governments to fight opioid addiction, plus more than 750 million dollars in direct payments to over 100,000 individual victims, ranging from 3,800 to 45,000 dollars each. In exchange, the Sacklers would give up ownership of Purdue and the company would be transformed into a nonprofit manufacturer of opioid addiction treatments.[14]

What that plan actually bought the Sacklers is the part that matters most. Buried inside the settlement was a legal maneuver that had nothing to do with money at all. The plan would have extended permanent immunity to the Sackler family from any future civil lawsuits related to the opioid crisis, even though the Sacklers themselves never filed for bankruptcy. They weren't offering up everything they owned to creditors the way an actual bankrupt debtor has to. They would have retained a significant portion of the wealth they'd extracted from Purdue while being released from billions of dollars in potential liability. There isn't a way to read that as anything other than buying legal peace at a discount.

A group of states refused to accept it. Nine states and the District of Columbia, along with the U.S. Trustee, a Department of Justice office, objected to and appealed the plan, and in December 2021, a U.S. District Court agreed with them, ruling that the bankruptcy court never had the authority to force states to release their claims against the Sackler family in the first place.[14]

That fight went all the way to the Supreme Court, and on June 27, 2024, in Harrington v. Purdue Pharma, the Court sided with the objecting states.[15] In a 5 to 4 decision, Justice Neil Gorsuch, writing for the majority, held that the bankruptcy code does not allow a court to discharge a third party, in this case the Sacklers, from legal liability when that third party isn't the one declaring bankruptcy. The Court found the kind of blanket shield the Sacklers were getting is only available to debtors who put all of their own assets on the table for creditors, and the Sacklers weren't doing that.

The dissent deserves real space here. That's a real cost. Blocking the shield also blocked, at least temporarily, the fastest path to money for the families who needed it. That tension doesn't resolve cleanly in either direction, and this piece isn't going to pretend it does.

What happened next is where this section earns its place in the piece. The states went back to the table without the immunity provision on offer, and it took most of a year to get there. On January 23, 2025, a bipartisan coalition of 15 states announced a new 7.4 billion dollar settlement in principle with the Sackler family and Purdue, this time without any grant of immunity from future lawsuits.[16] That announcement was only the first step. Over the following months, every eligible state and territory signed on, and by June 2025 all 55 attorneys general had agreed to the deal.[17] Under the plan as it moved toward court approval, the Sacklers must contribute up to 7 billion dollars over 15 years, and they permanently lose control of Purdue and are barred from ever selling opioids in the United States again. That's 1.4 billion dollars more than the settlement the Supreme Court had struck down the year before.

On November 18, 2025, a federal bankruptcy judge made it final. Judge Sean Lane approved the 7.4 billion dollar settlement, dissolving Purdue Pharma entirely and removing the Sackler family from any control of the company that made their fortune, while directing billions toward opioid abatement. Purdue itself is being replaced by a new company, Knoa Pharma. Most of the money goes to government entities fighting the opioid crisis, which has been linked to 900,000 deaths in the United States since 1999, with victim payments beginning next year.[18]

Opinion The Supreme Court didn't get the Sacklers a bigger check because it cared about optics, it got them a bigger check because it refused to let them buy their way out of ever facing a courtroom again. That's the part people miss when they just look at the final dollar figure. The family didn't pay more because they finally admitted fault. They paid more because the one thing they wanted most, permanent legal immunity without ever having to answer for what the emails and testimony showed, got taken off the table. Money was always something they were willing to spend. Accountability was the thing they fought hardest, and most expensively, to avoid.

Part Six

Low-Value Care, By the Numbers

Time to zoom back out. Everything so far has centered on one family and one drug, but the opioid crisis wasn't a uniquely evil exception inside an otherwise healthy system. It was the most lethal version of a pattern that runs through the entire way American healthcare pays for itself.

In 2019, researchers published a comprehensive review in JAMA that tried to answer a blunt question: how much of what America spends on healthcare every year actually does nothing for patients.[2] They pulled 71 separate cost estimates from 54 different publications, covering six recognized categories of healthcare waste, and found that wasteful spending totaled somewhere between 760 billion and 935 billion dollars a year. That's roughly 25 percent of all US healthcare spending, one dollar out of every four, though the number is actually down from an estimated 30 percent in 2011.

Breaking that figure apart matters, because "waste" can mean a lot of different things. The largest single category was administrative complexity, billing and coding overhead, physician time spent on quality reporting paperwork, at 265.6 billion dollars a year. Pricing failure, meaning prices that diverge from what a functioning market would produce, ran between 230.7 and 240.5 billion. Failure of care delivery, poor execution or a failure to adopt best practices, cost between 102.4 and 165.7 billion. Fraud and abuse accounted for 58.5 to 83.9 billion, and failure of care coordination, unnecessary hospital admissions and avoidable readmissions, added another 27.2 to 78.2 billion.

The category that matters most for this piece is overtreatment, what researchers call low-value care, medical tests, screenings, and procedures that carry real cost and real risk without providing meaningful benefit to the patient. That category alone was estimated at 75.7 to 101.2 billion dollars a year, and a more recent JAMA Network Open piece puts current annual low-value care spending at more than 100 billion dollars.[19] That's real people undergoing real procedures that couldn't help them and could hurt them, because the system is built to pay for doing something rather than paying for doing the right thing.

A 2022 study out of Johns Hopkins gave a much more concrete picture of who's actually driving that number. Researchers analyzed Medicare claims from 3,745 hospitals between 2016 and 2018, screening for 17 specific examples of unnecessary care, things like mammograms for women over 85, colonoscopies for people over 80, hysterectomies for benign conditions, and spinal fusions for back pain.[20] What they found wasn't random. Hospitals with more beds and fewer primary care physicians were more likely to deliver unnecessary tests or treatments, and investor-owned hospitals specifically were associated with higher rates of low-value or no-value care. Teaching hospitals, and hospitals with the highest share of uncompensated care, were less likely to perform unneeded procedures.

That finding isn't a coincidence, and it doesn't need to be. It lines up exactly with the incentive structure laid out in Part One. A hospital chasing investor returns under a fee-for-service model has a direct financial reason to order the extra scan, schedule the extra procedure, admit the patient one more night. A teaching hospital, more insulated from that specific pressure and more oriented around training and research, doesn't have the same pull in the same direction. The researchers behind the study put it plainly: "this wasteful care is physically, psychologically and financially harmful to patients."[20]

Opinion This is the section that ties Parts Two through Five back to the argument this piece opened with. The Sacklers weren't an aberration the system failed to catch. They were what happens when the system's normal incentive to overtreat gets applied by people willing to go further than anyone else was willing to go. The same financial logic that got a doctor a free lunch and a nine percent bump in opioid prescriptions is, at a larger and more sterile scale, the same logic behind an unnecessary spinal fusion or a colonoscopy nobody needed. Neither one requires a mustache-twirling villain. It just requires a system that pays more for doing more, and a workforce that responds to incentives the same way anyone would.

Part Seven

The COVID Reimbursement Question

This section needs to be handled differently than everything before it, because it isn't one claim, it's two claims stacked on top of each other, and only one of them is settled fact.

Here's the part that isn't in dispute. Congress built a specific financial incentive into COVID-19 hospital payments through the CARES Act. Section 3710 of the CARES Act directed the Secretary of Health and Human Services to increase the Medicare inpatient payment weighting factor by 20 percent for any patient diagnosed with COVID-19 and discharged during the public health emergency.[21] CMS identified those discharges using specific ICD-10 diagnosis codes, U07.1 for COVID-19 and, briefly early in the pandemic, B97.29 for other coronavirus causes, and hospitals could later collect a New COVID-19 Treatments Add-On Payment on top of that, tied to remdesivir and other approved treatments. That's not a rumor or a leaked memo. It's a statute, implemented through public CMS regulatory guidance.

Here's the part that's contested, and it's important not to let the documented policy above quietly validate a much stronger claim that wasn't proven. In April 2020, Dr. Scott Jensen, a Minnesota physician and Republican state senator, said in a Fox News interview:

What Jensen actually said matters, because it got flattened into something bigger as it spread. He specifically said he was not accusing physicians in the emergency room of lying or "gaming the system." His concern was that hospital administrators, not doctors at the bedside, might feel pressure toward a COVID-19 diagnosis in ambiguous cases because of the payment difference attached to it.

Multiple independent fact-checkers examined this claim directly, and their findings matter more than either side's talking points. FactCheck.org found that Jensen's dollar figures for Medicare reimbursement were reasonable estimates.[25] PolitiFact rated the broader claim "Half True," noting the 20 percent add-on payment was real, but that the article amplifying Jensen's comments implied case counts were being deliberately padded, when the available evidence actually pointed toward COVID-19 cases being undercounted, not inflated, largely due to limited testing availability early in the pandemic.[23] Snopes rated the claim "Mixture," finding it plausible that Medicare payments fell in the range Jensen cited, while noting Medicare itself said it does not make a single standardized payment amount for COVID-19 admissions the way the flat figures suggested.[24]

The pushback from people with direct knowledge of how this system actually works matters too. A health economist told PolitiFact "this is no scandal," explaining Congress added the 20 percent specifically because hospitals were losing revenue from canceled elective procedures during the pandemic while also facing higher costs treating COVID patients, including PPE.[23] A member of the American Medical Association's Journal of Ethics editorial board told PolitiFact it was "very unlikely that physicians or hospitals will falsify data or be motivated by money to do so," citing strict reporting policies.[23] On the other side, a Johns Hopkins surgeon and health policy professor told USA Today that "what Scott Jensen said sounds right to me," though he didn't elaborate further on the record.

The financial incentive Jensen described is real and documented. Whether it actually altered individual physicians' diagnostic or death-certificate decisions, at any meaningful scale, was not established by his statement, and the fact-check record on it is mixed rather than confirming.

How death certificates were actually supposed to be coded is worth laying out plainly. CDC's National Center for Health Statistics guidance allowed COVID-19 to be listed on a death certificate as "probable" or "presumed" when a definitive lab-confirmed diagnosis wasn't available but the circumstances were compelling within a reasonable degree of medical certainty, and certifiers were instructed to use their best clinical judgment.[26] Deaths were coded U07.1 whenever COVID-19 was reported as a cause that contributed to death, with or without lab confirmation. That's a real, and reasonable, allowance for a novel disease during a period when testing access was inconsistent. It is also exactly the kind of soft judgment call that a financial incentive could theoretically nudge, even without anyone consciously deciding to defraud the system. This piece isn't going to assert that it did.

Opinion Jensen wasn't describing fraud. He was describing exactly the same incentive architecture laid out in Part One, applied to a diagnostic gray area instead of a prescription pad. A "probable" or "presumed" diagnosis, by definition, leaves room for judgment, and it isn't unreasonable to notice that a system offering a financial reward for landing on one side of that judgment call deserves scrutiny, even in the total absence of proof that anyone actually took the reward on purpose. Scrutiny is not the same accusation as fraud, and this piece is careful never to blur the two.

The Pattern Repeats at the Top

One more piece belongs in this section, because it's the same species of claim sitting at the very top of the federal pandemic response rather than at the hospital level. On March 11, 2020, Dr. Anthony Fauci, then director of the National Institute of Allergy and Infectious Diseases, testified under oath before the House Oversight and Government Reform Committee. "The stated mortality over all of this when you look at all the data, including China, is about 3 percent," Fauci told lawmakers. "It first started off as 2 and now 3."[27] That testimony helped set the tone for how dangerous the public, and Congress, understood this disease to be in its earliest and most frightening weeks.

Fauci's own private estimate, recorded weeks earlier, put the number nowhere close to that. In personal notes dated February 8, 2020, roughly a month before his sworn testimony, Fauci estimated the true case fatality rate at "more like 0.2 to 0.3% rather than 2.0%," once asymptomatic and unrecorded cases were factored into the total.[28] That's roughly a tenfold gap between what he privately calculated and what he told Congress under oath, and it's Fauci's own written estimate, released through his own diary, compared against his own recorded congressional testimony.

Fairness requires including the context that softens this somewhat. In that same testimony, when asked to compare COVID-19's mortality to seasonal flu, Fauci said that counting all minimally symptomatic and asymptomatic infections would probably bring the mortality rate down to around 1 percent, which is closer to, though still meaningfully higher than, his private figure. He wasn't hiding the concept that asymptomatic spread would lower the real fatality rate, he acknowledged that dynamic on the record in the same hearing.[27]

Opinion The flu comparison he offered in testimony was doing real work to walk the number back toward something closer to what he'd privately calculated, without ever stating the private number itself. That's a distinction worth being honest about rather than flattening into a simple "he lied" framing. But acknowledging that nuance doesn't excuse the gap. If a public health official's own internal math says a novel disease is roughly as lethal as a severe flu season, and the number he leads with under oath is ten times higher, the public and Congress deserved the actual number, not a walked-back approximation buried in a follow-up answer.

One more discrepancy surfaced this year, though it's important to be precise about what it does and doesn't show, because a version of this claim spread with a statistic attached to it that doesn't hold up under scrutiny, and this piece isn't going to repeat that number. In a January 25, 2021 text exchange, Fauci wrote to then-CDC Director Rochelle Walensky and then-Surgeon General Vivek Murthy that "since many people have significant cytokine storm and fever after the 2nd dose, this theoretically could be associated with miscarriage in the 1st trimester," raising a theoretical concern about the vaccine during early pregnancy. Days later, in a public JAMA livestream discussion, Fauci said the FDA had "found thus far, and we have to be careful, but thus far no red flags about that, about pregnant women."[29]

Those two statements aren't as contradictory as they've been made out to be. A theoretical concern raised internally isn't automatically a "red flag" in the clinical sense the second statement was addressing, and the studies that followed, including a CDC analysis of 2,456 women tracked through pregnancy outcome, found a 14 percent miscarriage rate, within the normal 11 to 22 percent background range for pregnancy generally, with no elevated risk tied to vaccination.[29] This exchange is included because it's real and documented, not because it proves what it's often used to claim.

Opinion None of these discrepancies, on its own, proves Fauci set out to deceive the country. But the pattern matters, because it's the same pattern running through every other part of this piece. A hospital administrator responding to a fee-for-service incentive. A pharmaceutical rep repeating an unsubstantiated addiction statistic. A family extracting cash from a company mid-crisis. And now, a public health official leading with a mortality figure ten times higher than his own private calculation, at the exact moment that figure would shape how afraid an entire country became. Different actors, different stakes, but the same underlying question every time, what does the person with the real number gain by not saying it out loud.

Part Eight

What the Excess Mortality Data Actually Shows

Part Seven made a point of not asserting the death-coding incentive claim as proven. This is where that restraint gets tested against the actual numbers, because excess mortality data provides a way to check the coding question empirically instead of just weighing dueling quotes.

Here's how the method works, in plain terms. Researchers take the number of deaths that actually occurred in a given period and compare it to the number of deaths that would have been statistically expected based on prior years, using CDC baseline data. Whatever gap is left over is "excess mortality," deaths beyond the normal pattern that need some explanation. If hospitals and physicians were systematically over-labeling deaths as COVID-19 to chase the reimbursement bump covered in Part Seven, reported COVID-19 deaths should run higher than, or at least roughly equal to, total excess mortality. If the opposite is true, and excess deaths substantially outnumber the COVID-19 deaths actually reported, that's real evidence pointing the other way entirely.

The data points the other way entirely.

One analysis using CDC excess mortality data found that between January 3, 2020 and September 26, 2021, the United States experienced 895,693 excess deaths associated with COVID-19, a figure 26 percent higher than what was actually reported as COVID-19 deaths during that same window.[30] That's not a marginal gap. Nearly a quarter of the pandemic's actual mortality toll, by this measure, was never even captured in the official COVID-19 death count in the first place.

A separate study focused specifically on California reached a similar conclusion using a different method entirely, tracking death certificates directly rather than aggregate excess mortality modeling. Researchers found excess natural-cause deaths in California between March 2020 and February 2021 totaled 69,182, exceeding the number of deaths actually diagnosed as COVID-19 by 29 percent.[31] Nearly half of those excess deaths, 47.4 percent, occurred outside of a hospital setting entirely, where only 28.6 percent of the excess mortality was ever attributed to COVID-19 on the death certificate. The researchers' own conclusion, based on how closely the timing of those undiagnosed excess deaths tracked the timing of confirmed COVID-19 waves, was that this pattern suggests the excess deaths represented undiagnosed COVID-19 deaths, not indirect pandemic effects like delayed medical care, and that COVID-19 deaths were undercounted during the first year of the pandemic, not overcounted.

A multi-state analysis using National Center for Health Statistics data found the same directional gap held at the national level over a longer window. Comparing NCHS excess death estimates to actual observed COVID-19 deaths from January 2020 through December 2022, the analysis found roughly 1,233,366 estimated excess deaths nationally against 1,094,230 observed COVID-19 deaths for the same period.[32] That study did note one nuance worth including for fairness, New Jersey specifically saw its reported COVID-19 deaths slightly exceed its own excess mortality estimate, which the researchers attributed to the state's unusually early and detailed COVID-19 death classification system.

This fits a wider global context worth noting, because the American pattern isn't unique or uniquely suspicious. The World Health Organization estimated approximately 14.8 million global excess deaths through the end of 2022, a figure 2.7 times higher than the number of deaths officially reported as COVID-19.[33] And in countries with far less capacity for financially motivated hospital coding than the United States, India recorded roughly 2.4 million excess deaths in 2021 against only 335,004 officially reported COVID-19 deaths, a 7.2 times undercount, the same pattern at an even larger scale.[33] If a financial incentive built into American hospital billing were the driver of a COVID-19 death coding gap, an even larger version of the same undercounting pattern wouldn't be expected in a country where that specific incentive didn't exist.

Opinion The empirical evidence actually cuts against the version of the claim that spread the furthest. The theory that hospitals inflated COVID-19 deaths for reimbursement predicts a specific, testable outcome, official COVID-19 deaths exceeding the true mortality toll, and the excess mortality data says the opposite happened almost everywhere researchers looked. That doesn't mean Jensen's underlying concern about financial incentives shaping ambiguous judgment calls was worthless to raise, incentives are still incentives, and scrutiny remains reasonable. But this piece can't honestly claim the coding-for-cash theory bore out in the data, because by the measure that actually tests it, it didn't. If anything, the bigger and better documented failure here was months of undercounting, not the inflation Jensen worried out loud about.

Part Nine

Disincentivizing Harm

Everything up to this point has been documentation. This part is where the reporting turns into an argument about what it all means.

The entire premise of American healthcare is backwards. A system that pays more the more times you walk through its doors is a system that, whether anyone running it would ever say this out loud, does better financially when you're sick than when you're healthy. Part One laid out how fee for service rewards volume over outcome, and Parts Two through Six laid out how that reward shows up everywhere from a thirteen dollar lunch to a family's multi-billion dollar fortune. Opinion A hospital administrator under this system does better with a patient who needs to come back next month than with a patient who's cured and gone for good. That doesn't mean every doctor or administrator in this country is consciously rooting for you to stay sick. It means the system doesn't actually require anyone to be malicious for the outcome to bend that direction anyway, and that might be the more dangerous version of the problem, not a villain, just an incentive nobody has to admit they're responding to.

Opinion This is also where the limits of what can be said with confidence need to be honest. There's an argument that chemotherapy, one of the most expensive and physically brutal treatments in modern medicine, sometimes does as much damage to a patient's body as the cancer itself would have. This piece isn't qualified to tell you that's true across the board, outcomes vary enormously by cancer type, stage, and individual patient, and it isn't going to state something that specific as fact when it can't be backed up. But the question underneath that claim is a fair one to ask out loud. When a treatment is this expensive, this physically devastating, and delivered inside a payment system that financially rewards more treatment over less, it isn't unreasonable for a patient or a journalist to want a straight answer about whether the treatment protocol reflects the best available science or reflects what generates the most billable procedures.

The Olivieri Case

I went looking for a documented case of a doctor or researcher who was professionally punished for challenging a profitable treatment with something cheaper or safer, because that suspicion shouldn't sit on instinct alone. This is the one that makes the point better than instinct ever could, though it's worth flagging upfront that it happened in Canada, not the United States.

In the early 1990s, Dr. Nancy Olivieri, a hematologist at Toronto's Hospital for Sick Children, was running a company-funded trial of an experimental drug called deferiprone, meant to treat iron overload in patients with the blood disorder thalassemia. She became convinced the drug was both less effective than advertised and potentially toxic to the liver. When she moved to share those concerns with her patients and the scientific community, as her hospital's own consent policies required her to do, the drug's manufacturer, Apotex, terminated her research funding and threatened her with legal action if she went public. Neither the hospital nor the university stepped in to defend her. At the time, Apotex was in the middle of negotiating a 30 million dollar donation to the university and its affiliated teaching hospitals, and university officials maintained that donation had no bearing on their decision not to back her.[34]

An independent Committee of Inquiry, commissioned by the Canadian Association of University Teachers, later investigated the entire affair and concluded that Olivieri's academic freedom had been violated. Separately, the committee found that a colleague of hers, Dr. Gideon Koren, had authored a series of anonymous letters attacking Olivieri and the colleagues who supported her, and had then lied for months, denying authorship, until DNA evidence from the envelopes proved otherwise.[34]

He continued his academic career for another decade and a half before eventually resigning in 2015 amid an unrelated scandal.

Opinion That outcome answers a question this piece raised earlier almost perfectly. The doctor who tried to warn people about a drug's safety lost her funding, her institutional backing, and years of her professional life defending herself. The doctor found, by the hospital's own words, to have committed conduct serious enough to justify firing him, kept his job, his lab, and his career for fifteen more years. That isn't a coincidence specific to one Canadian hospital in the 1990s. It's what happens when an institution's financial relationships and reputational interests are allowed to weigh more heavily than the question of who was actually right. There isn't a stack of identical American cases to put next to this one, and this piece isn't going to pretend there is. But the incentive that produced this outcome isn't unique to Toronto, either, it's the same incentive this entire piece has been describing, applied to the question of who gets protected when they're inconvenient to a paying partner.

Burgers, Beer, and the Commercial Break

Something worth examining happened in real time during the pandemic, one of the clearest windows into how compromised this system's public messaging became. Governments and private companies spent 2021 offering vaccinated Americans free beer, doughnuts, marijuana, lottery tickets worth a million dollars, and even hunting rifles in exchange for getting a shot.[36] New Jersey ran a "Shot and a Beer" program through participating breweries. Anheuser-Busch partnered directly with the Biden White House, promising every drinking age American a beer voucher if the country hit 70 percent vaccination by July 4th. Krispy Kreme gave out free doughnuts. West Virginia gave away hunting rifles and shotguns. Ohio ran a weekly million dollar lottery funded by federal COVID relief money.

Opinion At the exact same time, major news outlets framed anyone asking questions about the vaccine as fundamentally anti-science, as though skepticism itself was the disqualifying position, while running pharmaceutical commercials in the very next commercial break. That's worth sitting with. A government literally bribing people with burgers and beer to take a product, while a media apparatus simultaneously treated any hesitation about that same product as a moral failing, is not what neutral, trustworthy public health communication looks like. It's what an industry with a financial stake in the outcome looks like when it's confident nobody's going to connect those two things happening in the same commercial break.

That isn't an accident of how American media happens to be funded. The United States and New Zealand are the only two developed countries in the world that allow direct-to-consumer advertising of prescription drugs at all. Researchers who've studied the practice have found direct evidence that it drives overdiagnosis and unnecessary treatment, raising both health costs and the risk of harm to patients, which is exactly why nearly every other developed country in the world has banned it outright.[37] Opinion Every other wealthy democracy on earth looked at direct-to-consumer drug advertising and decided the risk of turning medicine into a sales pitch wasn't worth it. This country looked at the same evidence and decided to let the industry buy the commercial break.

The Line Between a Question and an Answer

Two more examples belong here, because they get at something the cases above don't quite capture, the difference between asking a question and being treated as though you'd already answered it wrong. The earliest documented autism prevalence studies, conducted in the United Kingdom and the United States in the 1960s and 1970s, put autism at somewhere between roughly 1 in 2,000 and 1 in 10,000 children.[38] Today, the CDC's most recent surveillance data puts the number at 1 in 31.[39] The diagnostic criteria for autism changed substantially over that same period, most significantly in 1994, when the DSM-IV broadened the definition and folded Asperger's syndrome into the same diagnostic category for the first time. Public awareness, screening access, and willingness to seek an evaluation all expanded dramatically over the same decades, and most researchers attribute the majority of the increase to those factors. The largest study to test the aluminum-adjuvant exposure most often raised as the remaining vaccine-related explanation, tracking 1.2 million children over 24 years across 50 tracked health outcomes including autism, has not found a link.[40]

Opinion None of that makes it irrational to notice the two trend lines moving upward together and want an answer for why. It becomes a problem when noticing that coincidence gets treated as though it were already a conclusion, in either direction.

Here's an example that actually earns the suspicion, rather than just gesturing at it. In 2021, a coalition of doctors and scientists sued the FDA under the Freedom of Information Act for the data underlying Pfizer's COVID-19 vaccine licensure. The FDA had taken roughly four months to review that same data before granting approval. When ordered to release it publicly, the FDA proposed a production rate of 500 pages a month, for a dataset of at least 451,000 pages, a schedule that would not have finished until the year 2096, roughly 75 years away.[41]

A federal judge in the Northern District of Texas rejected that timeline outright, calling it "unduly burdensome" to the public interest, and ordered the FDA to produce 55,000 pages a month instead, a timeline the agency then met.

Opinion There isn't an innocent explanation for proposing a 75 year release schedule for data behind a product being given to hundreds of millions of people, and it doesn't require a conspiracy theory to say so, a federal judge looked at the same proposal and rejected it in almost identical terms. This is exactly the kind of institutional behavior that makes distrust rational rather than fringe, and conflating the people who pointed at this specific, documented, court-adjudicated stonewalling with people spreading unfounded claims did real damage to the credibility of legitimate questions.

The ivermectin conversation belongs back in this section too, because it's a smaller-scale version of the same dynamic. Much of the loudest mainstream ivermectin coverage treated anyone discussing it as though they were recommending it as a standalone cure. In practice, a lot of the actual public conversation, including accounts describing personal COVID treatment on popular podcasts, described it as one part of a broader combination of treatments, alongside monoclonal antibodies and other therapies, not a substitute for all of them. Whether any of those combinations were actually effective is a separate clinical question this piece isn't qualified to settle. But it's fair to say the popular framing collapsed a nuanced, multi-part claim into its simplest and easiest-to-mock version, and then treated the strawman as the thing that had been debunked.

Opinion All of this, the autism trend line, the FDA's own proposed release timeline, and the ivermectin flattening, points at the same underlying failure. Somewhere in the last several years, this country lost the ability to distinguish "I have a question about this" from "I have already decided this is a lie." Those are not the same posture, and treating them as identical doesn't protect public trust in medicine, it corrodes it, because it teaches people that the system's response to a fair question is to call them names instead of showing their work. The FDA's own proposed 75 year timeline is proof that sometimes the people asking why were asking the right question.

Opinion Here's the thesis of everything in this piece. Everyone treats the price of American healthcare as the problem, and it is a problem, but price is downstream of something else entirely. If the people running this system were paid based on how healthy their patients stayed instead of how many times their patients came back through the door, the prices wouldn't need nearly as much regulation as everyone assumes. The five hundred dollar aspirin, the unnecessary spinal fusion, the drug commercial between two anti-vax segments, and a hospital's decision to protect a colleague who lied over a doctor who told the truth, all of it comes from the exact same place, a system engineered to reward the wrong outcome. Fix that, and the price problem mostly fixes itself, because the industry would no longer have a reason to manufacture more of the thing it's currently paid to sell.

Part Ten

You Decide

Nine parts have built a case, and it is a case, not a neutral recitation. This piece went in believing this country's healthcare system runs on the wrong incentive, and came out of the reporting more convinced of that, not less. But conviction isn't what should convince you. The documents should do that, or they shouldn't.

Here's what's known for certain, because it isn't opinion, it's the public record. A family took hundreds of millions of dollars out of their own company within a year of pleading guilty to a federal felony for misrepresenting the very drug that was killing people. A federal database shows seventy two billion dollars flowing from drug companies to physicians, and shows prescribing patterns moving in lockstep with even the smallest of those payments. A quarter of all American healthcare spending, by the government's own estimate, buys nothing of value to the patient. A Supreme Court had to intervene to stop a family from buying its way out of ever facing a courtroom again, and even then it took a second, more expensive settlement before that family lost control of the company for good. A hospital's own written decision said a colleague's conduct constituted grounds for dismissal, and then declined to dismiss him. An FDA proposed a 75 year timeline for releasing vaccine licensing data, and a federal judge had to force it down to eight months.

None of that requires anyone to say what they think about chemotherapy, or about a text message Dr. Fauci sent in January 2021, or about whether the autism numbers mean what any particular person wants them to mean. Those are the places in this piece flagged plainly as opinion rather than fact, where the evidence itself is genuinely mixed or unsettled. That line was kept visible throughout on purpose, not because a strong personal view is absent here, it clearly isn't, but because the entire premise of this publication falls apart the moment that honesty about where the documents end and the author begins stops holding.

Opinion Here's the actual opinion, laid out one final time without hiding behind anyone else's data. The price of American healthcare is a symptom, not the disease. The disease is a payment system that rewards more treatment instead of better outcomes, and every specific failure in this piece, the opioid crisis, the seventy two billion dollars in quiet influence, the low-value care nobody needed, the institutional protection of the wrong people, all traces back to that same root. The people working inside that system aren't uniquely evil. Most of them are responding, like anyone would, to what the system actually pays them to do. That's what makes this harder to fix than finding a villain would be. There's no one person to arrest that makes any of this stop. There's only a set of incentives that has to be dismantled and rebuilt around a simpler question than the one it currently answers, not "how much can this generate," but "did this person actually get better."

You decide.