Follow the Money, Then the Medicine · Morning Gist
Wednesday, October 215 stories · 15 min
Follow the Money, Then the Medicine
Who funds the doubts about ultra-processed food, CAR cells made inside mice, too many pills, the CDC budget fight, and cancer drugs from China.
Picked from 28 new articles across 12 sources · arrow keys for other days
Today in 30 seconds
30s left
Nine in ten of 69 papers criticizing the ultra-processed food definition had industry ties, as did 68.5% of their 232 authors. story 1 ↓
An mRNA injection turned about 80% of mice's natural killer cells into CAR cells; 3 of 5 melanoma mice fully remitted. story 2 ↓
Medicare now spells out how doctors and pharmacists can bill for drug reviews and tapering as more older Americans take five or more drugs. story 3 ↓
Congress looks set to reject Trump's CDC cuts again, but the administration is steering already-approved money through other channels. story 4 ↓
US cancer patients are flying to China for a lung cancer drug the FDA approved in 2025 that still isn't sold here. story 5 ↓
Nigeria confirmed an Ebola case in Abuja in a traveler from Uganda; Kenya has recorded nine infections since its first case this month.
1/5Food Politics (Marion Nestle)· Marion· Oct 20
Nine in ten papers criticizing the ultra-processed food definition have industry ties
A review in the American Journal of Public Health found that 91% of 69 papers attacking the Nova system for classifying ultra-processed foods had at least one food-industry-linked author. It doesn't test whether their criticisms are right.
Why it matters · Ultra-processed food is moving into dietary guidelines and policy. Knowing who funds the doubt helps you weigh the critiques, even though funding alone doesn't settle the science.
Know this first
Context from the editor, not the article · 25 sec
1 · What Nova classifies
Nova, developed by Brazilian researchers, sorts foods by how much they are processed rather than by nutrients. Group 4, ultra-processed foods, means industrial formulations such as soft drinks, packaged snacks and reconstituted meats.
2 · State of the evidence
Most evidence linking ultra-processed diets to obesity, diabetes and early death comes from observational cohort studies. One small 2019 NIH inpatient trial found people ate about 500 extra calories a day on an ultra-processed diet.
3 · The tobacco precedent
Internal documents showed tobacco companies funded scientists to cast doubt on smoking's harms. Public health researchers now look for the same "manufactured doubt" pattern in other industries.
2/5Fight Aging!· Reason· Oct 20
Mice get CAR immune cells made inside the body by mRNA injection
Researchers injected mice with fat-bubble particles carrying mRNA aimed at natural killer cells, turning roughly 80% of them into cancer-hunting CAR cells in the body — clearing tumours and autoimmune T cells without any cell manufacturing.
Why it matters · If in-body engineering holds up in humans, it could cut CAR cell therapy from a bespoke six-figure manufacturing job to an off-the-shelf injection.
Know this first
Context from the editor, not the article · 27 sec
1 · How CAR therapy works now
A chimeric antigen receptor is a synthetic protein added to immune cells so they recognise and kill a chosen target. Today it means harvesting a patient's cells, engineering them in a lab for weeks, then reinfusing them.
2 · What lipid nanoparticles do
The fat bubbles used in Covid mRNA vaccines. A cell swallows one, releases the mRNA inside, and makes the encoded protein for a few days until the mRNA degrades. Temporary, unlike gene therapy.
3 · Natural killer cells
A class of white blood cell that kills damaged or infected cells without prior training. Being not a T cell matters here: T cells engineered against T cell targets attack each other.
3/5KFF Health News· Blake Farmer, Nashville Public Radio and Cara Anthony· Oct 20
More older Americans take five or more drugs, and experts urge medication reviews
The share of Americans over 65 taking five or more drugs nearly doubled between 2000 and 2020. Pharmacists and doctors interviewed by KFF Health News say patients should ask their prescribers which drugs they could safely stop.
Why it matters · Taking many drugs is a common, preventable source of falls, memory problems and organ damage in older adults, and the patient is often the only one who sees the full list.
Know this first
Context from the editor, not the article · 28 sec
1 · What polypharmacy means
Polypharmacy usually means taking five or more medications at once. The more drugs a person takes, the more pairs can interact, and the harder it is to tell which drug is behind a new symptom.
2 · Why age changes drug effects
As people age, the liver and kidneys clear drugs more slowly and body composition shifts. A dose that was fine at 50 can build up at 80, which raises the risk of falls, confusion and stomach or bowel problems.
3 · Fragmented prescribing
Older patients often see several specialists who work in different health systems, each with its own records. Unless someone puts the whole list together, no single clinician may know everything the patient takes.
4/5Inside Medicine (Jeremy Faust)· Jeremy Faust, MD· Oct 20
Congress set to reject Trump's plan to gut the CDC for a second year
Trump's 2027 budget again asks Congress to cut the CDC and reorganize the health department under Robert F. Kennedy Jr. Congress is ignoring it, so the administration is trying other ways to control the money.
Why it matters · Which vaccine, HIV, birth-defect and research programs survive, and who picks the screenings insurers must cover for free, increasingly depends on who controls money already appropriated.
Know this first
Context from the editor, not the article · 30 sec
1 · Who decides federal spending
Each year the President sends Congress a budget request, but it is only a wish list. The House and Senate write appropriations bills that actually set spending, and they often ignore the President's numbers. The federal fiscal year starts October 1.
2 · The agencies involved
The Department of Health and Human Services (HHS) oversees the CDC (disease control and prevention), the NIH (medical research funding) and other agencies. Robert F. Kennedy Jr., known for his vaccine skepticism, has been HHS Secretary since 2025.
3 · Last year's fight
5/5KFF Health News· Arthur Allen· Oct 20
US cancer patients fly to China for drugs not yet sold at home
A New York woman with lung cancer that had spread to her brain got sunvozertinib, a Chinese-made drug, by having it bought in Shanghai. Washington, meanwhile, argues over whether to court Chinese biotech or restrict it.
Why it matters · For patients with rare cancer mutations, China now develops some of the best options first, and US rules on trials and approvals decide when Americans can get them.
Know this first
Context from the editor, not the article · 26 sec
1 · Targeted lung cancer drugs
Some lung cancers, often in nonsmokers, are driven by one mutation in a growth gene such as EGFR. Drugs called tyrosine kinase inhibitors block that mutated protein, and each one fits a particular mutation subtype.
2 · Phases of clinical trials
Phase 1 is the first test in humans and checks safety and dose. Late-stage (phase 3) trials compare the new drug with standard treatment, often assigning patients at random, so some patients get the older therapy.
3 · Approval isn't availability
After the FDA, the US drug regulator, licenses a drug, a company still has to make, price and distribute it in the US. That gap can last months.
A 1 in the American Journal of Public Health examined 69 papers, published from 2009 to 2025, that criticized 2 or defended 3. Nova is the main system for classifying ultra-processed foods. Nestle, who writes Food Politics, highlighted the review. 91.3% of the papers had at least one food-industry-affiliated author, and 68.5% of the 232 authors had industry ties. About two-thirds of the papers appeared in journals run by nutrition societies with disclosed industry partnerships.
Why the definition matters
Nova sorts foods by degree of processing rather than nutrient content. Its top category covers industrial formulations such as soft drinks and packaged snacks. Nestle argues that having a definition is what lets researchers study these foods, and that research has linked high intake to obesity, diabetes and early death. Most of that evidence is 4. The key controlled study is a small 2019 5 trial in which people ate about 500 extra calories a day on an ultra-processed diet. So scientists have legitimate grounds to argue about the definition. Nestle notes that some independent nutrition researchers object to judging foods by processing at all.
Ties and tone
The review also checked whether papers went further and actively defended ultra-processed foods. They did so in 39.7% of papers with direct industry ties, 60% with organizational or governmental ties, and 16.7% with no identified ties. The authors compare this to tobacco companies' strategy of manufacturing scientific doubt.
Caveats
The untied group is tiny. If 91% of 69 papers had ties, only about six had none. The 16.7% figure is probably one paper, so treat the contrast as directional, not precise.
Only critics were studied. With no comparison of conflicts among Nova's supporters, the 91% is hard to put in context.
Funding shows a pattern, not intent. The tobacco comparison is a hypothesis. Concerns about Nova's vagueness still need answers on their merits.
The summary doesn't explain how ties were defined or how papers were found.
The takeaway
Nestle says the findings confirm long-held suspicions, which fits her years of documenting industry influence. Her practical conclusion is modest: eating less ultra-processed junk food is likely to improve health. For readers, the useful lesson is to check who wrote and funded a headline about ultra-processed foods "not being so bad."
The finding
A 1 in the American Journal of Public Health (November 2026, published online in September) looked at who writes the papers attacking 2, the most widely used system for defining 3. Most of those authors have links to the food industry. Marion Nestle, the nutrition-policy scholar who writes Food Politics, flagged the study. She came to it through a story in The Examination, a news outlet, and press coverage summed it up as "nine in ten" critical papers having industry ties.
Why the definition is contested
Nova was developed by researchers in Brazil. It ignores sugar, salt and fat content and sorts foods into four groups by processing: unprocessed or minimally processed foods, culinary ingredients, processed foods, and ultra-processed foods. Group 4 covers industrial formulations made mostly from extracted or modified ingredients and additives, such as soft drinks, packaged snacks and instant noodles.
Nestle's point is that a definition is what makes research possible. Once "ultra-processed" can be measured, researchers can link it to health outcomes, and a large body of studies has done that. Diets high in these foods are associated with obesity, type 2 diabetes, heart disease and higher mortality. Most of that evidence is 4, though. It comes from cohort studies that cannot fully separate the food from the lifestyles of the people who eat it. The main controlled evidence is a small 2019 5 inpatient trial in which volunteers ate about 500 more calories a day on an ultra-processed diet than on a minimally processed one. That leaves real scientific room to argue about the definition, and it also leaves room to exploit that argument.
Nestle notes that legitimate critics exist outside industry. Some nutrition researchers object to a focus on processing instead of nutrient content. Others say Nova is vague and that different raters classify the same food differently. Whole-grain bread and a candy bar can both land in group 4.
What the review did
The authors identified 69 articles published between 2009 and 2025 that either criticized Nova or defended ultra-processed foods. They then traced the affiliations of the authors and journals. The summary Nestle reproduces does not describe the search strategy or how a "tie" was defined: funding, employment, consulting or advisory roles.
Results
91.3% of articles (about 63 of 69) had at least one industry-affiliated author.
68.5% of the 232 authors (about 159 people) had food industry ties.
65.2% of articles appeared in journals linked to nutrition societies that disclose industry partnerships.
The review also asked whether articles went beyond criticizing the classification to actively defending or promoting ultra-processed foods. They did so in:
39.7% of articles with direct industry ties
60.0% of articles with organizational or governmental industry ties
16.7% of articles with no identified ties
The authors conclude that, given tobacco's "well-documented strategy of 6," the pattern raises concern that the food industry may be running a similar playbook as evidence of harm grows.
Reading the numbers carefully
The comparison groups are lopsided. If 91.3% of 69 papers had industry ties, only about six had none. The 16.7% figure for untied papers is then probably one paper out of six. (This is our back-calculation; the summary does not give counts.) The contrast with 39.7% or 60% points in a clear direction, but it rests on a tiny group and should not be read as a precise effect size. The summary also doesn't make clear how the "direct" and "organizational or governmental" categories overlap.
Three further limits matter:
The review studied critics only. It doesn't tell us how often pro-Nova papers carry conflicts of their own, such as ties to advocacy groups or to authors invested in the framework. A comparison base rate would make the 91% easier to interpret.
Association is not proof of influence. The tobacco comparison is a hypothesis about motive. The review shows a funding pattern, not documents showing that companies directed the arguments.
Funding doesn't make an argument wrong. Complaints about Nova's reproducibility and its lumping of very different foods have to be answered on their merits. What the review justifies is extra scrutiny of where the volume of doubt comes from, not dismissing every critique.
Nestle's take
Nestle says the results "confirm what I've long suspected." Her reading is that the industry dislikes the concept because the research it enables shows harm, and lower consumption threatens sales, which she says is already happening. She has spent decades documenting industry influence on nutrition science, so this fits a position she already held. Her bottom line is modest: not eating too much ultra-processed junk food is likely to improve health.
What's open
The questions that remain are whether better definitions emerge that keep Nova's insight while fixing its fuzziness, and whether trials beyond the single NIH study can show which features of ultra-processed foods cause harm. The candidates include energy density, how fast the food is eaten, additives and palatability. Both answers will shape how regulators write labels and guidelines, and the review shows that much of the pushback they face comes from interested parties.
Remember
Industry-tied critics were far more likely than untied ones to actively defend ultra-processed foods, but the untied group was about six papers.
Most evidence that ultra-processed diets cause harm is observational; controlled trial evidence is still thin.
Conflicts of interest call for closer scrutiny of an argument but don't disprove it; Nova's vagueness remains a real debate.
Researchers have shown in mice that the expensive manufacturing step at the heart of 1 cell therapy can be skipped: inject a targeted nanoparticle carrying mRNA, and the animal's own immune cells build the cancer-killing machinery themselves. The open-access paper was highlighted by the aging-research writer Reason at Fight Aging!
Why this matters
A chimeric antigen receptor (CAR) is a synthetic receptor planted on an immune cell's surface; when it binds its target, the cell kills whatever it is touching. Seven CAR T cell products are approved in the US. But each dose is bespoke: harvest a patient's T cells, engineer them, grow them for weeks, reinfuse. That is why the therapy costs what it does.
2 — white blood cells that kill damaged cells without prior training — have one distinct advantage. A CAR T cell aimed at a T cell protein will attack its own kind, so T cell diseases are hard to treat this way. NK cells have no such conflict. But making CAR NK cells needs the same factory.
The mechanism
3 are the fat bubbles from the Covid mRNA vaccines: a cell swallows one, translates the mRNA inside, and makes the encoded protein for a few days before the mRNA degrades. The hard part is aiming them — injected LNPs usually end up in the liver.
The team coated particles with engineered 4 ligands: antibody-derived fragments that grip two different spots on the same NK cell surface receptor at once, making uptake receptor-specific. They call them 5. The cargo is mRNA encoding a CAR, and which CAR is swappable.
The numbers (all in mice)
Caveats
The short-lived CAR is a safety feature — you cannot recall an infused cell therapy — but it means repeat dosing, and repeated antibody-coated particles may eventually draw an immune response of their own. Five mice is five mice, and the summary says nothing about durable remission. Reason adds a non-biological obstacle: LNP intellectual property is largely locked up, with only a few formulations in trials or approved. Mouse immunotherapy results flatter translation more often than not.
The news
Researchers have shown, in mice, that you can skip the factory step of 1 cell therapy entirely: inject a nanoparticle carrying mRNA, let the animal's own immune cells read it, and cancer-killing engineered cells appear inside the body. In the lead experiments, up to about 80% of the targeted natural killer (NK) cells took up the instructions. The open-access paper, "In vivo engineering of CAR natural killer cells for immunotherapy using ligand-directed lipid nanoparticles," was flagged by the aging-research writer Reason at Fight Aging! as an example of what mRNA delivery may be genuinely good for.
Why CAR therapy needs fixing
A chimeric antigen receptor (CAR) is a synthetic receptor protein planted on an immune cell's surface. It binds a chosen antigen — a cancer marker, say — and when it does, the cell activates and destroys what it is touching. The US Food and Drug Administration has approved seven CAR T cell products over the past decade, and in several blood cancers they work well.
The catch is manufacturing. Each dose is patient-specific: collect the patient's T cells, engineer them to carry the CAR, expand them in culture for weeks, then infuse them back. That is slow, technically demanding, and expensive enough to keep the therapy out of reach for most patients and most health systems.
Natural killer cells — white blood cells that kill damaged cells without needing prior exposure — have recently been engineered the same way, and they bring one specific advantage. If you want to attack diseased T cells, CAR T cells are a poor tool, because T cells carrying a receptor against a T cell protein will shred each other along with the target. NK cells do not have that problem. But CAR NK cells still require the same costly manufacturing.
The mechanism: aiming the nanoparticle
2 are the fat-and-cholesterol bubbles that carried mRNA in the Covid vaccines. A cell engulfs one, the mRNA is released into the cytoplasm, the cell's own ribosomes translate it, and the encoded protein exists for a few days until the mRNA is broken down. As Reason puts it, that places LNP-mRNA halfway between gene therapy and a small-molecule drug: it makes a protein do something specific, but only for a while.
The hard part is aiming. Injected LNPs tend to end up in the liver. The team built what they call 3 — NK cell-targeted lipid nanoparticles — coated with engineered 4 ligands, meaning binding proteins (antibody-derived fragments) that grip two different epitopes on the same NK cell surface receptor at once. Two-point binding makes uptake receptor-specific rather than incidental, so the mRNA goes where intended.
Inside the cargo is mRNA encoding a CAR. Which CAR is a swappable choice. The researchers used one against 5, a protein on B cells; one against 6, found on the autoreactive T cells that drive certain autoimmune diseases; and, separately, mRNA aimed at 7, an inhibitory checkpoint receptor on NK cells.
An analogy, labelled as such: conventional CAR therapy is building a custom robot in a workshop and shipping it to the patient. This is mailing the patient a set of instructions that their existing workers read, follow for a few days, then throw away.
The results, with numbers
All of the following is in mice.
Two humanized blood-cancer models also regressed when the particles were optimised for human NK cells, which is the step that matters most for translation.
Trade-offs and what's unresolved
The transience cuts both ways. A CAR that lasts days is safer — CAR T cell therapy can cause severe cytokine release and lasting immune depletion, and a dose you cannot recall is a liability. But it also means repeat dosing, as the five-injection lymphoma experiment shows, and repeated injections of an antibody-coated particle raise the question of whether the immune system starts neutralising the delivery vehicle. The paper summary does not address durability of remission, and the melanoma result — 3 of 5 animals — is a very small number.
Reason adds a commercial caveat that has nothing to do with biology: the LNP field today resembles the viral-vector field a decade ago. Enormous invention, but much of the intellectual property on specific lipid formulations is locked up, and only a handful of formulations with a narrow range of delivery behaviours have reached trials or approval. A clever targeting ligand does not help if the lipid underneath it cannot be licensed affordably.
Also unstated: how much of this survives the jump to humans. Mouse NK cell biology, tumour models and humanized mice all flatter immunotherapies routinely. What the work establishes for now is a modular platform — same particle, different mRNA, different disease — rather than a treatment.
Remember
The CAR only lasts days, since the delivered mRNA degrades — safer than infused cells, but requiring repeat dosing.
The share of Americans over 65 taking five or more medications nearly doubled from 2000 to 2020, according to a study in 1. More than a third of older adults now take that many drugs a day. A KFF Health News and Nashville Public Radio explainer says patients worried about this should ask their doctors about "2": stopping or tapering drugs they no longer need, under a clinician's supervision.
The piece is practical advice built on interviews with clinicians. It reports no new trial and no data on outcomes of deprescribing.
Why drug lists grow
Taking five or more drugs at once is called 3. Amanda Mixon, a 4 (a hospital-based doctor) at Vanderbilt, says the older patients she sees average 16 medications from multiple prescribers and pharmacies. That is her impression of hospitalized patients, not a national figure. Each added drug raises the chance of a harmful interaction.
One patient in the story, Marine veteran Jonathan Bledsoe, took a mood stabilizer and a cholesterol drug among several others for nearly three years. Then his urine turned coffee-colored. The combination was damaging his liver. After his drugs were switched, his liver and kidney function returned to normal.
A common route to a long list is the 5: a drug causes a side effect, and a second drug is prescribed to treat that side effect. Mixon's example is heart failure. A water pill reduces fluid buildup but causes frequent urination, which can lead to incontinence, which is then treated with yet another drug.
Separate health systems make this worse, because prescribers often can't see each other's records. "There isn't this magical unicorn pharmacist just making sure everything is safe," says pharmacist DeLon Canterbury, who runs GeriatRx, a consultancy that helps patients stop unneeded drugs. Aging also changes how the body processes drugs, which raises the risk of falls, memory loss and stomach or bowel problems.
What patients can do
Keep a current paper list of all medications, vitamins and supplements, and carry it with you.
After a new symptom or side effect, ask your primary care doctor or specialist whether each drug is still needed.
Canterbury says the aim is safety, not the fewest pills. He also says the science of safe tapering is new and not widely taught in medical schools.
Policy
This year Medicare's administrator, the 6, clarified how doctors and pharmacists can be paid for reviewing medications and supervising tapers. Doctors had been unsure they could bill for this time-consuming work.
Separately, Health Secretary Robert F. Kennedy Jr. has pushed to wean Americans off psychiatric drugs he considers overused, a stance that has divided physicians. That is a different matter from a supervised review of one patient's drug list.
The share of Americans over 65 taking five or more medications nearly doubled between 2000 and 2020, according to a study in 1. More than a third of older adults now take that many drugs every day. A KFF Health News and Nashville Public Radio piece in their HealthQ consumer series looks at the risks of piling up prescriptions. Its advice is that patients should start a "2" conversation themselves, and it notes a recent Medicare payment change that may make doctors more willing to have it.
This is a practical explainer, not a report on a new trial. It cites one prevalence study and draws mostly on interviews with clinicians and one patient's story. It gives no outcome data on deprescribing itself.
The problem: drugs accumulate
Polypharmacy is the usual term for taking five or more medications. Amanda Mixon, a 3 (a doctor who cares for patients in hospital) at Vanderbilt University, says the older patients she sees take an average of 16 drugs, often from several prescribers and pharmacies. That figure is her clinical impression of hospitalized patients, not a population statistic. They are probably sicker than older adults overall.
Her point is about how the risk adds up. "As you accumulate more and more medications, the chance of them interacting with each other and having bad interactions increases," she said. Drugs tend to be added one at a time, each for a good reason, and rarely taken away.
The article opens with Jonathan Bledsoe, a Marine veteran in Reading, Pennsylvania. For nearly three years he took a mood stabilizer and a cholesterol drug along with several other medications. He noticed his urine had turned "the color of fresh brewed coffee." His doctors found that an interaction between his drugs was damaging his liver. After his medications were changed, his liver and kidney function returned to normal. The article does not name the specific drugs.
How it happens: the 4
One mechanism the clinicians describe is the prescribing cascade. A drug causes a side effect, the side effect is mistaken for a new condition, and a second drug is prescribed to treat it. That second drug can cause problems of its own.
Mixon's example is a heart failure patient. A diuretic, or water pill, reduces fluid buildup. The frequent urination it causes can lead to incontinence, and incontinence is then treated with another medication. Each step makes sense on its own, but the patient ends up taking a drug only because of another drug.
Fragmented care keeps cascades going. Pharmacist DeLon Canterbury, CEO of GeriatRx, a North Carolina consultancy that helps patients stop unnecessary drugs, notes that prescribers in different health systems often can't see each other's records. "There isn't this magical unicorn pharmacist just making sure everything is safe. Yes, even at the local pharmacy," he said. Canterbury frames overtreatment as a result of training rather than carelessness: clinicians are taught to diagnose and treat, and much less to stop drugs.
Age makes things worse. Canterbury notes that older bodies process drugs differently, which raises the chance of constipation, diarrhea, falls and memory loss.
What deprescribing is
Deprescribing means deliberately reducing or stopping medications under a clinician's supervision, often by tapering them gradually. Canterbury says the practice has gained momentum over the past decade, but the science of how to taper safely is newer and still not widely taught in medical schools. He stresses that the goal is safety, not the smallest possible number of pills. Some drugs are essential, and stopping certain ones abruptly can be dangerous, which is why tapering is supervised.
The article does not report how much deprescribing reduces falls, hospital stays or deaths. Readers should know that published evidence on hard outcomes is mixed. Trials often show that drug counts can be cut safely, but clear gains in survival or hospitalization are harder to show.
What patients are told to do
The advice is modest and practical:
Keep a current written list of every medication, vitamin and supplement, and carry it. Jen Hammons of the Appalachian College of Pharmacy says, "Everyone needs a medication list in their pocket." Apps and QR-code bracelets exist, but paper works.
Raise deprescribing yourself, especially after a new symptom or unexpected side effect. Ask a primary care doctor or specialist whether each drug is still needed. Canterbury says clinicians are usually more open to the conversation than patients expect.
Don't assume anyone else is checking for interactions across prescribers.
The policy backdrop
This year the 5, the federal agency that runs Medicare and Medicaid, issued guidance on how physicians and pharmacists can be paid for reviewing a patient's medications and helping them taper off. Before this, doctors said they weren't sure deprescribing work could be billed, and it takes a lot of time. The article does not say which billing codes or payment amounts apply.
Deprescribing has also become politically charged. Health and Human Services Secretary Robert F. Kennedy Jr. has pushed to wean Americans off psychiatric medications he considers overused, a position the article says has polarized the medical community. Reviewing a drug list with your doctor is a different thing from a broad campaign against a whole class of drugs. Stopping antidepressants or other psychiatric medications without supervision carries its own risks.
What's left open
The piece shows that 6 is common and growing and that the mechanisms of harm are plausible and familiar to clinicians. It leaves several questions unanswered:
How much deprescribing improves outcomes.
Which drug classes are the best candidates for stopping.
Whether the CMS payment clarification will actually change how often clinicians review medications.
Remember
Some of the drugs older people take are there only to treat side effects of other drugs (a prescribing cascade).
No one reliably checks your full drug list across prescribers, so carry your own list, including supplements.
Medicare now explicitly pays clinicians for medication review and tapering, which removes one barrier to doing it.
In 2025 the White House asked for deep HHS cuts and also tried to withhold money through executive action. Congress largely refused, and courts reversed many of the sudden cuts. Some of those cases are still unresolved.
Table from last year's negotiations comparing the President's, House and Senate HHS funding proposals by agency.1 / 2
President Trump's 2027 budget request again asks Congress to cut deeply into the 1 and reorganize it. Part of the plan is a new agency, the 2, that would put a large pool of money under Secretary Robert F. Kennedy Jr. Jeremy Faust, an emergency physician who writes the newsletter Inside Medicine, reports that Congress rejected the same plan last year and is clearly going to reject it again.
What happened last year
For the CDC, the request was a $5 billion cut, roughly 40%. The House proposed cutting $1.8 billion and the Senate $0.1 billion. The final bill gave the CDC $9.2 billion, just $14.5 million (0.15%) below the year before. Congress also wrote new language into the law. HHS must keep the staff needed to run the programs Congress funds, and it can't move programs out of the CDC without independent review. Faust points out two weaknesses. Nobody has defined what "needed" staffing means, and the law gives no clear way to stop a reorganization.
Why "no one cares" this year
That is how a former HHS budget official put it. The CDC is behaving as it would in a normal year. It invited applications for an HIV prevention program and advertised for a director of its birth-defects center, even though the President's budget would eliminate both centers. In 2025, staff held back similar postings. In April, Kennedy told the Senate health committee: "If you appropriate the money, I will spend it."
The House bill adds money for infectious-disease work and gives traditional CDC vaccine programs $15 million more than last year. The Senate is holding its bill until after the November 3 midterms. In the past the Senate has never proposed less than the House, so Faust treats the House bill as a reasonably reassuring worst case.
The fight moves to control
Faust lists five ways the administration is trying to steer money Congress has already approved:
NIH grants: a proposed rule would let political appointees decide grants. Congress ordered it halted.
$2 billion in infectious-disease funds would go to the Defense Department rather than the NIH. How much will move, and what it will pay for, is unknown.
Faust concludes that the total funding looks settled. The harder fight, over who gets to spend it, will take months or years.
For the second year running, President Trump and HHS Secretary Robert F. Kennedy Jr. have asked Congress to reorganize the 1 and cut its funding sharply. HHS is the federal department that runs the CDC (Centers for Disease Control and Prevention), the NIH (National Institutes of Health) and other agencies. Jeremy Faust, an emergency physician who writes Inside Medicine, reports that Congress turned down the same plan last year and will clearly turn it down again in this budget cycle, for fiscal year 2026–2027. His larger argument is that the real fight has moved. The question is no longer how much HHS gets, but who controls the money once Congress has approved it.
The plan, again
The centerpiece is the 2, a new agency built partly from programs taken out of the CDC and other agencies. Last year's version would have put about $20 billion in HHS programs under Kennedy's control, plus $500 million to pay for the reorganization. Faust describes the budget as an attempt to make legal what Russell Vought had been trying to do by executive order. Vought directs the 3 and was an architect of Project 2025, the conservative governing blueprint. Courts reversed many of those executive cuts, though some cases remain unresolved.
How last year ended
A President's budget is only a request, and appropriations bills passed by Congress set actual spending. In 2025 both chambers wrote their bills as if the President's HHS proposal did not exist. Here is how the CDC numbers narrowed:
President's request: cut $5 billion, about 40%
House proposal: cut $1.8 billion
Senate proposal: cut $0.1 billion
Final: $9.2 billion, $14.5 million (0.15%) below the previous year
"For Congress to spurn the President's Budget in 2025 was a dramatic statement," a former HHS budget official told Faust, because Trump "was at the height of his power and influence."
Congress also wrote guardrails into the final bill. HHS must keep the staff it needs to carry out programs Congress has funded, which rules out unjustified 4 (reductions in force, the government term for layoffs). It also cannot move programs out of the CDC without independent review. Faust sees two gaps. First, nobody knows what staffing level counts as "needed." He offers an example: if Congress requires $29 million for autism work, little stops the administration from having one or two political appointees spend all of it on anti-vaccine projects. Second, the law gives no practical way to stop a reorganization if the review objects.
Why this year is quieter
"No one cares," the former budget official said of this cycle. Faust treats that as a good sign, and the CDC's behavior supports him. In 2025, staff were nervous enough that they held back job postings and grant calls for units the President wanted to cut. This year:
Early in the year, the CDC invited community groups to apply for an ongoing HIV prevention program, even though its parent center would disappear under the President's request.
In July, it advertised for a director of the National Center for Birth Defects and Developmental Disabilities, another center the budget would eliminate.
Faust says this is how the CDC normally operates: it assumes current funding will continue, and that assumption is usually right. Congressional hearings this spring showed less alarm than last year's, when the administration was regularly trying not to spend appropriated money. In April, Kennedy told the 5 (Health, Education, Labor and Pensions): "If you appropriate the money, I will spend it." Faust does not know why the White House tried the reorganization again. He notes that Kennedy's influence fell last winter after Trump concluded that meddling with vaccine policy was unpopular.
What the House wrote
The Senate has not released its bill and plans to wait until after the November 3 midterms. In the past the Senate has not proposed less than the House, so the final deal will fall between the two chambers' bills, and the President's request can largely be ignored. The House bill goes against Kennedy's stated wish to "give infectious diseases a break." It increases several infectious-disease lines and gives traditional CDC vaccine programs $15 million more than last year. Faust adds a footnote: the President's budget seems to raise CDC immunization and respiratory funding by $50 million, but that money is for a BioThreat Radar Detection system, not vaccine programs. He also counted the word "woke" 34 times in the President's document and zero times in the House's.
Five end-arounds
Faust argues that the administration's new approach is to put as much congressionally mandated spending as possible in the hands of a few loyal appointees. He describes this as part of a "unitary executive" agenda, the theory that the President should have full control over the executive branch. His examples:
What's unresolved
The final spending numbers will come after the midterms. Faust expects that if Democrats win both chambers, Republicans will want to settle quickly. He considers the House bill a reasonably reassuring worst case for funding levels. Several questions remain open: how much of the $2 billion will move, what the reconstituted USPSTF will recommend, and whether Congress's staffing and review requirements can actually be enforced. In Faust's view, the fight over control of the money will last "months and years."
Remember
Last year Trump asked to cut the CDC by 40%. Congress cut 0.15%, leaving it with $9.2 billion.
Insurers must cover any screening the USPSTF rates A or B for free, which makes Kennedy's eight new members highly influential.
Congress requires HHS to keep the staff its programs need, but never defined 'needed', which makes the rule hard to enforce.
Debbie Doppelt, a 60-year-old Manhattan jeweler, has a rare lung cancer that mostly affects nonsmokers. By December 2024 it had spread to her brain as 10 small tumors. Chemotherapy, immunotherapy and surgery had stopped working. The drug that best matched her tumor's mutation was 1, made by China's Dizal Pharmaceutical and approved there in 2023. It was not available in the US.
Why New York couldn't help
The approved option:2, approved by the FDA (the US drug regulator) in 2024, targets the same mutation. Her son Max worried about its toxicity, and doctors told him it might not reach tumors in the brain.
A trial she couldn't join: slight lung scarring made her ineligible for a Memorial Sloan Kettering trial of 3, a similar drug.
A trial she wouldn't join: US trials of sunvozertinib gave half the participants chemotherapy alone, and chemo had already failed her.
A relative's contact in China got her a first supply. Max then flew to Shanghai and brought back 26 boxes, which cost about $10,000. Nineteen months later her cancer is still mostly under control, with some radiation and dose changes. That is one patient's outcome. Max's estimate that the drug helps "about half" of users is informal. One woman he helped found after three months that her cancer had spread anyway.
The FDA approved sunvozertinib in July 2025, but it still wasn't on US shelves in mid-October 2026. AstraZeneca, which signed a distribution deal with Dizal in September, plans US availability by early 2027. Max knows of at least 30 Americans who made the same trip.
China's rise in biotech
By the article's count, China accounted for up to 46% of new drugs tested worldwide in 2025. It leads in 4, a therapy that engineers a patient's own immune cells to attack cancer, and foreign patients travel there to join its trials. Biotech leaders say first-in-human studies start much faster in China and Australia. One CEO says nearly all 60 executives at a May meeting planned to run their 5 trials abroad.
Washington pulls both ways
The Trump administration's 6, modeled on Australian rules, aims to speed up FDA trial processes. FDA nominee Heidi Overton called current requirements "slow and prohibitive." In Congress, a bipartisan bill would require the Treasury Department to clear every US pharma deal with a Chinese company. Investor Peter Kolchinsky argues the countries are "collaborating" to cure cancer, not competing.
Only about 7% of US cancer patients enroll in trials. Max has built a free AI tool that matches patients to trials, and about 300 people have used it so far.
One family's route to Shanghai
In January 2025, Debbie Doppelt, a 60-year-old Manhattan jeweler, had run out of good options in New York. She has a rare form of lung cancer that mostly affects nonsmokers. She was diagnosed in late 2023 and had been through chemotherapy, immunotherapy and surgery. In December 2024, scans showed 10 small tumors in her brain.
The drug that looked best for her was 1, made by Shanghai-based Dizal Pharmaceutical. It was approved in China in 2023 and targets the mutation driving her tumors. A doctor told her son Max, "Too bad she's not Chinese. She could go to a pharmacy in China and buy it." Within hours the family had found a relative with a contact in China who got a small supply. A few months later Max flew to Shanghai and came back with 26 boxes in his luggage, which cost about $10,000.
The FDA (Food and Drug Administration, the US drug regulator) licensed sunvozertinib in July 2025. As of mid-October 2026 it still wasn't on the US market. AstraZeneca signed a worldwide distribution deal with Dizal on September 1 and plans US availability by early 2027. Max says he knows of at least 30 Americans who have made the same trip.
Why the US options didn't fit
Her treatment ran into three separate obstacles, and each shows a different weak point.
An approved drug with drawbacks. The FDA had approved 2, which targets the same mutation, in 2024. Max worried about its toxicity, and several doctors told him it might not reach tumors in the brain. Many drugs struggle to get past the 3, the lining that keeps most substances in the blood out of brain tissue.
Trial eligibility rules. She tried to join a Memorial Sloan Kettering trial of 4, another drug for her mutation. Slight scarring on her lungs ruled her out.
Trial design. Sunvozertinib was being tested in the US only in late-stage trials where half the participants got chemotherapy alone. Her cancer had already beaten chemo, so a coin-flip chance of getting it again was unacceptable to her.
From general knowledge, not the article: amivantamab, zipalertinib, sunvozertinib and firmonertinib are all aimed at 5 mutations. Most EGFR inhibitors work poorly against this subtype.
How it has gone
Nineteen months after she started sunvozertinib, her cancer was still mostly under control, with help from a few radiation treatments and dose changes. This is one patient's story, not trial data. Max estimates the drug helps about half the people who take it. That is his informal figure, not a published response rate. A 70-year-old woman he helped bought a six-month supply from China and learned after three months that her cancer had spread anyway.
Max expects the cancer to become resistant eventually. He is now looking at a US trial of firmonertinib, which China approved in 2021. He is also watching a 6 therapy being developed at the University of Colorado. CAR-T means a patient's own immune T cells are engineered to attack the cancer. Kyle Concannon, who helps lead that team, says regulatory uncertainty has slowed the work. Newer tests for toxicity and for where a therapy travels in the body catch more possible harms, but they add time, cost and risk. "The alternative to a new medicine, with a natural disease like lung cancer, is the individuals will die from their disease," he said.
China's rise in drug development
China used to be known mainly as a supplier of generic drugs and the chemicals used to make drugs. By the article's figure, it accounted for as much as 46% of new drugs tested worldwide in 2025, and it now leads in CAR-T trials, which draw patients from the US and Europe. Sam Rasty, who runs the biotech FynderBio, traces the shift to an overhaul China made after its weak vaccine and drug record during covid. Investors now favor early clinical work in China and Australia. Both countries let companies start first-in-human studies much faster than the US does.
Steve Potts, CEO of Breakthru Medicine, says nearly all of the 60 biotech CEOs at a May meeting planned to run 7 trials in Australia, Europe or China. Five or 10 years ago, most would have run them in the US. He says American biotechs now hide early results because they fear Chinese competitors will move faster with the same ideas.
Two policy responses pulling apart
The Trump administration launched 8 in June. It aims to speed up some FDA processes and is modeled on Australian procedures that date to 1991. Heidi Overton, the nominee for FDA commissioner, called FDA requirements "slow and prohibitive" at her September 24 confirmation hearing and promised to keep the program. An HHS spokesperson said buying medicine abroad for personal use is sometimes legal, but the FDA warns about safety and quality risks.
Congress is moving the other way. Reps. John Moolenaar (R-Mich.) and Debbie Dingell (D-Mich.) have co-sponsored a bill that would make the Treasury Department clear every US pharmaceutical deal with a Chinese company. Investor Peter Kolchinsky of RA Capital, which has Chinese holdings, argues that "we're not competing with them to cure cancer. We are collaborating with them." Patient advocates, including Marcia Horn of the International Cancer Advocacy Network, say they will keep using "the China option" when they need to.
Finding trials
Only about 7% of US cancer patients enroll in treatment trials. Max quit his tech job and built a free AI tool. It combines trial listings from research and drug company sites with a patient's health records, then ranks trials by fit and distance. About 300 people have used it so far, all with lung cancer. Companies such as Paradigm Health sell similar tools to hospitals.
What remains open
The article doesn't say why sunvozertinib took more than a year to reach US patients after approval. It also doesn't show whether TrialBlazer has shortened any timelines yet, or what the Treasury-review bill would mean for deals like AstraZeneca's. And the quality risk the FDA warns about for drugs bought abroad is real, but the article doesn't measure it.
Remember
FDA approval doesn't mean a drug is available: sunvozertinib was still off US shelves 15 months after licensing.
Trial design matters to patients: half of the US sunvozertinib trial's participants got chemo alone, which ruled it out for her.
The outcomes here are anecdotes. Max's 'helps about half' figure is informal, not trial data.
Up to ~80% of targeted NK cells were transfected, with minimal off-target delivery.
A CAR against 6, a B cell marker, completely cleared circulating B cells in healthy mice.
In aggressive melanoma, 3 of 5 treated mice reached complete remission.
In a humanized T cell lymphoma model, tumours were eradicated after five injections — comparable, the authors say, to conventionally manufactured CAR NK cells.
In autoimmune hepatitis, where T cells attack liver cells, one injection of a CD3-targeted CAR destroyed nearly all the autoreactive T cells and liver damage healed.
Against solid tumours, mRNA aimed at 7, an inhibitory checkpoint on NK cells, improved the effect by lifting a "don't attack" signal.
Transfection: lead NtLNPs reprogrammed up to ~80% of targeted NK cells, with minimal off-target delivery.
B cell depletion: CD19-targeted CAR NK cells completely cleared circulating B cells in healthy mice — the standard proof that the machinery works in vivo.
Melanoma: in an aggressive mouse melanoma model, 3 of 5 mice treated with tumour-targeting NtLNPs reached complete remission.
T cell lymphoma: in a 8, tumours were eradicated after five injections — a result the authors describe as similar to treating mice with conventionally manufactured CAR NK cells.
Autoimmune hepatitis: in mice where autoreactive T cells attack liver hepatocytes, a single injection of CD3-targeted NtLNPs destroyed nearly all the autoreactive T cells and liver damage healed.
Solid tumours: NtLNPs delivering mRNA against the NKG2A checkpoint improved antitumour effect by relieving suppression signalled through Qa-1b, the mouse ligand that NKG2A reads as a "don't attack" cue.
The 3: Kennedy added eight voting members. Insurers must cover screenings this panel rates A or B with no cost to patients.
Foreign aid: $2.5 billion set aside for global aid went to the costs of shutting down USAID.
CDC money for pet projects: research into a vaccine–autism link that has been repeatedly debunked, and $1.6 million toward a hepatitis B vaccine trial in Guinea-Bissau that Faust calls unethical.
NIH grants. A proposed rule would move funding decisions from outside scientists and career staff to political appointees. It drew hundreds of thousands of public comments, and Congress ordered it halted in a summer stopgap bill that kept the government open.
$2 billion to the Defense Department. The administration has sought to send infectious-disease money to the Department of Defense, which it now calls the Department of War. Its argument is that the money would still be spent on infectious diseases, just not by 6 (the National Institute of Allergy and Infectious Diseases, the NIH institute for that field). Because Congress did not specify in detail how NIAID's money must be spent, the move may be legal. How much will actually move is unknown. A Republican lawmaker said NIH Director Jay Bhattacharya assured him no basic-research money would fund "unrelated efforts." Faust questions how much weight "unrelated" is carrying in that promise.
The 7. Under the Affordable Care Act, insurers must cover preventive services this panel rates A or B with no out-of-pocket cost to patients. The New York Times reported that Kennedy added eight voting members, after removing some and letting others' terms expire. He made similar moves at the CDC's vaccine advisory committee, 8 (Advisory Committee on Immunization Practices), and courts reversed them. The USPSTF's governing law gives him more leeway, so the new members could, for example, issue autism-screening recommendations that change what insurers must cover.
USAID funds. Politico reported that as the fiscal year ended, $2.5 billion set aside for global aid went to the costs of shutting down USAID, the foreign-aid agency. Faust says the report appears to be accurate.
CDC money for pet projects. Kennedy pushed the CDC to fund outside research on a vaccine–autism link that decades of high-quality studies have rejected. According to Rolling Stone, $1.6 million in "global health security" funds may go to a Danish trial of the hepatitis B birth-dose vaccine in Guinea-Bissau. Faust calls the trial unethical, and his newsletter first published its protocol. He says he has heard rumors of other such projects.
