Yesterday, in the Oval Office, President Donald Trump described what he believed happens to an American infant at a routine well visit. He said he had seen proof of a vaccination the size of a bottle of soda poured into a little child’s body. 

A few minutes later he described vats of vaccine being pumped in and explained that, under his new policy, children would receive 20% of that at each of five separate visits, with time allowed to elapse between them so the body could handle the massive amount of fluid. In the question period afterward, he returned to it again, describing the single visit as one where they pump in what seemingly looks like gallons.

Nearly every injectable childhood vaccine comes in a half-milliliter dose. Give a 2-month-old five separate injections at one visit and the total is two and a half milliliters, about half a teaspoon. In practice it is less, because combination products deliver several components in a single injection. A 20-ounce bottle of soda holds about 590 milliliters.

He said he had seen proof of a vaccination the size of a bottle of soda poured into a little child’s body.

That is a gap of more than 200-fold, and an executive order was signed on the strength of it.

The volume error would be a footnote if the document rested on something sturdier. It does not. The only evidentiary document the order cites is a January 2026 assessment by the Department of Health and Human Services (HHS) comparing the American immunization schedule to those of other countries. No trial, no safety signal, no dataset, as I noted in a CIDRAP op-ed then. And the central number in that assessment cannot be lowered by the policy it was written to justify.

The odd math behind ‘84 doses’

The White House fact sheet issued that afternoon states that the 2024 childhood schedule reached at least 84 vaccine doses in at least 57 shots. At the signing, the President said 72. Secretary Kennedy, standing a few feet away, said it could run as high as 94.

The 72 is at least traceable, as the assessment’s ceiling for injections rather than doses. The 94 matches nothing in the document, which caps doses at 88 and injections at 72.

The 84 is the one that appears in writing, so it is the one worth examining. It comes from the assessment, authored by Tracy Beth Høeg, MD, PhD, and Martin Kulldorff, PhD, and it reproduces exactly, provided you accept two conventions the administration rarely explains in public.

The first is that each disease that a combination shot protects against counts as its own dose. The legend to Table 2 on page 15 states it plainly: the MMR shot contains three vaccine doses, one each for measles, mumps, and rubella. So the two MMR injections a child receives become six doses on the ledger, and the five DTaP shots plus one Tdap booster become 18. (DTaP and Tdap vaccines protect against diphtheria, tetanus, and pertussis, or whooping cough.)

The second convention is projection. The assessment assigns influenza and COVID-19 18 to 19 doses apiece, up until a child’s 18th birthday. At the lower bound, which is where the 84 and the 57 come from, those two supply 36 of the doses and administrations. 

For influenza that is a fair model, because the 2024 schedule does specify annual vaccination. For COVID-19 it is not. That schedule calls for one or more doses of the updated formula and nowhere commits to a dose every year through age 18.

Someone defending the count will say that the Centers for Disease Control and Prevention (CDC) recommended each season’s updated COVID formula for everyone 6 months and older, so projecting an annual dose describes actual practice even if the printed schedule never says annual. The assessment’s own authors were not confident of that. Their footnote is conditional, beginning with the words if it is recommended annually, and the headline figure then treats the condition as settled. So 18 of the 84 doses are projected COVID shots that no schedule ever called for, and this administration went on to end the COVID recommendation for healthy children entirely. 

Officials inflated the number with a vaccine they removed, and then they cited the inflated number as grounds for removing more.

The 1980 comparison does not add up, either. Its stated components—one measles, mumps, and rubella (MMR), five DTP, one Td, and four polio drops—come to 24 doses under the same rule that makes MMR three. The document says 23, and the fact sheet repeats it.

The count reaches 18 diseases only by including nirsevimab, the respiratory syncytial virus (RSV) monoclonal antibody, which is not a vaccine. It delivers laboratory-made antibodies directly, protecting a baby through one winter and then wearing off, rather than teaching the immune system to build protection of its own. The assessment lists it apart from the vaccines in its own table, then counts it among the diseases anyway. If the concern is that we ask too much of an infant’s immune system, this is the one product on the list that asks nothing of it.

Splitting MMR into three single-disease products removes nothing from the total. The tally stays at six. The number of shots goes from two to six.

Now go back to the first convention, because it does something to the order that nobody at the signing appears to have noticed. If one MMR injection is already counted as three doses, then splitting MMR into three single-disease products removes nothing from the total. The tally stays at six. The number of shots goes from two to six. 

The order’s signature intervention, the one the president spent most of his remarks describing, does not move the statistic it was written to correct. It only triples the injections.

A document with nothing to enforce

The order sorts childhood immunizations into three categories and puts hepatitis A, hepatitis B, and meningococcal disease in two of them at once, both as recommendations for high-risk groups and as subjects for shared clinical decision-making. It never says who counts as high risk. The document that does, the January schedule memo, has been blocked since March, when a federal judge in Massachusetts stayed it, along with the reconstituted advisory committee’s votes in American Academy of Pediatrics v. Kennedy. The categories arrive with nothing behind them.

Section 2(b) says MMR should be given as three separate shots at separate visits, once such products are domestically available and to the maximum extent feasible. Section 3(a) tells HHS to make single-disease MMR available while guaranteeing that combination vaccines remain available. A defender will call that a recommendation alongside preserved choice. The order never says which one a pediatrician is supposed to follow.

On everything medical, the order asks only for paperwork. Agencies are told to review the recommendations. The HHS secretary is told to deliver plans within 90 days. Both instructions come wrapped in qualifiers, to be carried out as appropriate and only to the extent the law allows. 

States, which are the ones that actually set school vaccine requirements, are not told to do anything at all. They are advised to consider changing their requirements.

Section 4 is written differently. There the attorney general is directed to bring lawsuits against states over religious and medical exemptions, and three departments are directed to press their contractors and grantees on the same question. That is the only place the order reaches past the executive branch, and it has nothing to do with medicine. It is also where the administration expects the work to happen. At the signing, Heidi Overton, MD, PhD, said the administration would work directly with states so that they would not be relying on the court case to resolve.

Recommending vaccines that don’t exist

Set aside the arithmetic and consider what the order asks American parents to do. It asks them to obtain three products that cannot be obtained.

There is no measles-only vaccine available in the United States. There is no mumps-only vaccine and no rubella-only vaccine, either. Three measles-containing products are licensed here: M-M-R II, PRIORIX, and ProQuad, all combinations. 

Merck told the CDC’s vaccine advisory committee in October 2009 that it would not resume making the single-disease versions, a decision it reached after requests from that committee and the American Academy of Pediatrics, on the grounds that separate shots would delay protection, leave children susceptible longer, and increase both injections and visits. 

The World Health Organization (WHO) has prequalified a standalone measles vaccine and a standalone rubella vaccine. It has never prequalified a standalone mumps vaccine anywhere in the world.

Bringing them back would mean three new products. Each would need its own biologics license under section 351 of the Public Health Service Act, and approval turns on inspection of the manufacturing establishment as well as evaluation of the product. Merck said in a statement to PolitiFact that splitting MMR would essentially create three new separately licensed vaccines requiring full clinical testing and regulatory review. It estimated more than 10 years to approval and commercialization. The order concedes the difficulty in its own text, calling for the split to happen once such products are domestically available.

Until then, the instruction to a parent who takes the order seriously is to wait. It arrives during the highest annual measles total since 1991, 2,465 confirmed cases as of August 6, as schools reopen in the hardest-hit states, and three months before the regional commission decides in November whether the United States has lost its measles elimination status, a review the United States asked to postpone from April.

The thing they call unstudied

Standing beside the president, Deputy Chief of Staff for Policy Stephen Miller said that parents are pressured into five, six, or seven shots when their children are weeks out of the hospital, and that “nobody has studied it, nobody has looked at it, nobody has thought about it.”

Nothing could be more detached from reality.

Vaccines on the childhood schedule are licensed on the basis of controlled clinical trials of the product or its components, and combination shots are tested as combinations. Guidance on combination vaccines from the Food and Drug Administration (FDA) directs that immunogenicity and safety data supporting simultaneous administration with already licensed vaccines should be obtained before licensure. The FDA also says that combinations should be compared against their components given separately at the same visit. 

Once a vaccine is in use, the surveillance continues. When reports of intussusception (a telescoping of the intestine) after the first rotavirus vaccine accumulated in the national reporting system in 1999, the recommendation was suspended in July, emergency case-control and cohort investigations followed, and the product was off the market by October. 

The schedule as a whole has been examined, too. The Institute of Medicine reviewed it in 2013 and uncovered no evidence of major safety concerns associated with adherence to the childhood immunization schedule, while calling for further study of specific questions the literature had not settled.

What that work shows is a schedule that has grown in injections while shrinking in what it asks of the immune system. The relevant measure is antigens, the pieces of a germ the immune system learns to recognize. 

The whooping cough vaccine used into the 1980s was made from the whole killed bacterium and carried on the order of 3,000 distinct proteins, more than everything else on that era’s schedule combined. Today’s schedule covers 16 to 18 diseases and contains between roughly 165 and 315 antigens, depending on the counting rules, a reduction of roughly 90% to 95%. We got there by working out which few pieces actually generate protection and making only those.

Offit and colleagues calculated that an infant could theoretically respond to about 10,000 vaccines at once, and that 11 given simultaneously would engage roughly a tenth of one percent of the immune system. The empirical answer came later. A 2018 study in JAMA compared children who developed infections unrelated to any vaccine against children who did not and found no meaningful difference in cumulative antigen exposure from vaccination, 240.6 against 242.9.

Offit and colleagues calculated that an infant could theoretically respond to about 10,000 vaccines at once.

The executive order tells HHS to develop alternatives to the aluminum salts added to some vaccines to strengthen the immune response. The assessment behind it does identify a signal: a 2023 analysis of Vaccine Safety Datalink records reporting a hazard ratio of 1.19 (a 19% increased risk) for persistent asthma per milligram of aluminum. That is the kind of finding that warrants a larger and better-designed test, and one was published a year ago. 

Andersson and colleagues followed about 1.2 million Danish children in a study published in the Annals of Internal Medicine, examining aluminum exposure from vaccines against 50 chronic conditions. Nothing came back. The hazard ratios sat at or below 1 across autoimmune, allergic, and neurodevelopmental outcomes, asthma included. The authors concluded that the data rule out moderate or large increases in risk, while noting that very small ones cannot be statistically excluded. That is the answer to the question the assessment raised, from the registry of the country this administration selected as its model.

MMR has nothing to do with any of it. Its package insert lists no adjuvant, and no live attenuated vaccine on the American schedule carries one, because a weakened virus that replicates briefly supplies its own immune stimulus.

Meanwhile, the intervention the executive order prescribes has never been tested. No trial has compared the recommended schedule against an alternative one, a gap the Institute of Medicine noted in 2013 and that nobody has since filled. The one element of this prescription that has been studied is delay, and it points the wrong way. 

Two Vaccine Safety Datalink studies covering hundreds of thousands of children found that a first measles-containing dose given at 16 to 23 months carried roughly twice the risk of a post-vaccination fever-related seizure as the same dose given at 12 to 15 months, for MMR alone as well as for MMRV (MMR plus varicella, or chickenpox). 

Those seizures are frightening to witness but carry no lasting harm. What the studies establish is narrower. The only measured effect of pushing measles vaccination later is increased cost, and the order proposes to push vaccination later.

What we can actually learn from peer nations

Peer nations are moving toward the American approach, not away from it. 

The United Kingdom added varicella vaccination on January 1 of this year, reversing a refusal that had stood since 2009. Britain’s advisers had declined on the theory that taking childhood chickenpox out of circulation would drive shingles up in adults. Decades of program data from the United States and elsewhere showed it had not, and the objection dissolved. 

Germany added routine infant meningococcal B in 2024. Countries across Europe are expanding RSV prevention. Among the 20 peer nations in the assessment’s own comparison table, exactly one does not use MMR.

That one is Japan, and it is the closest thing to a natural experiment. Japan suspended MMR in 1993 over cases of meningitis traced to the particular mumps strain in its vaccine, a real safety problem specific to that strain, and moved to separate measles and rubella shots. Measles kept circulating, and sentinel surveillance estimated about 286,000 cases in the 2001 epidemic. 

Japan returned to a combination product in 2006 with routine two-dose MR, and mumps was never restored to the routine schedule, which is why Japan still has mumps outbreaks and mumps-related deafness. Japan’s own answer arrived this year. In May, regulators approved Mimlit, the country’s first domestically licensed MMR vaccine in roughly three decades, built around a different mumps strain with a long record in combination products elsewhere. The remedy for a strain problem turned out to be a better strain rather than three separate shots.

None of this requires parents to decide against vaccination. Coverage falls when vaccination gets harder, not only when people refuse it. 

Coverage falls when vaccination gets harder, not only when people refuse it.

The WHO holds that eliminating measles takes at least 95% two-dose coverage in every district, and US national kindergarten coverage was 92.5% in the 2024-25 school year, with individual schools and counties far below that. About 286,000 kindergartners started school without a documented MMR series. Combination products were developed to reduce injections and visits, and the measurable consequence is that children who receive them complete the series more often, and more of it on time.

Bad things are already happening

Before he signed, President Trump said that nothing bad can happen from what the administration is doing, and then said it again.

Something bad had already happened. Earlier in the same event he said of the combined MMR that there could be a possibility they are quite lethal, and that separately they are not lethal at all. When a reporter asked him for the evidence, he attributed it to other people, allowed that there might be a 5% chance of it, and repeated the claim. That is the president of the United States telling parents, twice, on camera, that the shot their pediatrician recommends may kill their child.

If the schedule changes take hold, the rest follows on a delay:

• Fewer children will be vaccinated, because turning one appointment into three will lower vaccine uptake. 
• Children will be hospitalized who would not have been. Some will die. 
• Up to 90% of infants who acquire hepatitis B at birth become chronically infected, and roughly a quarter of those infected in childhood die prematurely of cirrhosis or liver cancer. 
• The infections will not stay among children: Measles, whooping cough, and hepatitis B move through households and workplaces and hospital waiting rooms, to infants too young for their first dose, to pregnant women, to people on chemotherapy, to grandparents.

None of that is a prediction about something unknowable. It is what happened in Japan, in the United Kingdom after Andrew Wakefield sowed misinformation, and in every community where coverage has slipped since. We know how this goes, because we have watched it go.

President Trump said nothing bad can happen. Bad things have already happened, and more are coming, and every one of them was avoidable.

Dr. Scott is a clinical associate professor of infectious diseases at Stanford University School of Medicine, and a coauthor of “Updated evidence for COVID-19, RSV, and Influenza Vaccines for 2025-2026” in the New England Journal of Medicine.

The opinions voiced in CIDRAP Op-Ed pieces are the authors’ own and do not necessarily represent the official position of CIDRAP.