By Justin Yamashita, MSc. Benchtop, site, CRO: three levels of basic and clinical research, explained without spin.

A drug isn't finished when it's approved. That's the single most common misunderstanding about how any of this works, and this week’s claim depends on it.

Approval is a decision made on the evidence available at that moment, from populations of a certain size, over a certain length of time. Everything rarer or slower than that study could detect is, by definition, still unknown on the day of approval. So the law doesn't let the study stop. It changes shape.

Post-marketing surveillance, sometimes called Phase IV, is a regulated ongoing stage of a product’s life. It has real machinery: a safety database that every reported adverse event flows into, signal detection run against it, periodic safety reports submitted to FDA, and post-marketing studies that FDA can require as a condition of approval.

When a signal survives investigation, the label changes. That's the mechanism by which a boxed warning, the strongest alert FDA can put on a drug label, gets added to a product that was already on the market. This newsletter borrowed that name deliberately. A real black box warning isn't evidence that the system failed to notice. It's the receipt for the system noticing.

For vaccines specifically the monitoring is layered. VAERS accepts reports from anyone and is built for signal detection, which means it's deliberately noisy and can't on its own establish that a vaccine caused anything. The Vaccine Safety Datalink is the counterweight: linked health records across participating systems, which allow rates in vaccinated and unvaccinated populations to be compared directly. People quoting raw VAERS counts as though they were findings are using a smoke detector as a fire report.

Counting disease cases is a different system with different owners. A confirmed measles case is laboratory confirmed or epidemiologically linked, reported by a clinician or lab to a state or local health department, and passed up through the National Notifiable Diseases Surveillance System. Jurisdiction, onset date, and vaccination status travel with it.

That last field is why outbreak clustering isn't a correlation anyone reverse-engineered afterwards. It's collected at the point of report.

So: one system watches the product, the other counts the disease. The claim quietly requires you to believe neither exists.

Measles was declared eliminated in the US in 2000, meaning no continuous year-round transmission. In 2025 the US reported 2,289 confirmed cases across 48 outbreaks with three reported deaths, the worst year since 1991, and 90 percent of those cases were outbreak-associated. 2026 has already passed it: 2,566 confirmed by mid-August, with the year not over. And the Region of the Americas lost its measles elimination status in November 2025, after endemic transmission was re-established in Canada. The US determination is scheduled for November 2026. So American elimination status is at risk rather than lost, and that distinction now has a date on it [1].

Polio was eliminated in the US in 1979. In 2022 an unvaccinated adult in New York developed paralytic polio from a vaccine-derived strain, and the same strain was later found in wastewater across several counties, meaning it had circulated silently before anyone was paralyzed [2].

That case deserves a note, because the claim will reach for it. It was vaccine-derived, which sounds damning until you look at the mechanism. Vaccine-derived strains come from the oral polio vaccine, which the US stopped using in 2000. The strain arrived from elsewhere, and it paralyzed the one person in the chain who wasn't vaccinated.

Pertussis. This is the honest hard case of the three, and being straight about it means correcting the easy version. Whooping cough returned to pre-pandemic levels in 2024, stayed high through 2025 at 28,783 reported cases and 16 deaths, ten of them infants under one, and has been falling since [3]. That is not a generational high, and I am not going to call it one; the modern peak was 2012. What has not improved is who pays. Infants under six months carry the highest incidence of any group and are hospitalized at roughly 41 percent. Protection from the current acellular vaccine wanes faster than the older whole-cell version did. That's a real limitation, it's openly discussed in the literature, and it's why vaccination during pregnancy exists, to cover newborns in the window before their own doses begin. The claim raises none of it, because raising it would mean engaging with how the vaccine actually works.

The claim isn't persuasive because the evidence is close. It's persuasive because it's well matched to five documented features of human reasoning. Naming them is the point, because a mechanism you can name is a mechanism you can catch.

Availability heuristic. We estimate probability by how easily an example comes to mind [4]. No mental image of the disease, no felt risk. This is the engine of the whole Vanishing Act.

Omission bias. Harm from acting feels worse than equal or greater harm from not acting. This was demonstrated using vaccination decisions specifically: people were reluctant to vaccinate a hypothetical child when the vaccine carried a small risk of death, even when the disease posed a substantially larger one [5].

The prevention paradox. Geoffrey Rose’s formulation, coined in 1981 and restated in the paper cited here: a measure delivering large benefit to a population offers very little to any one participant [6]. Every individual vaccinated is almost certainly fine either way. The benefit is real and it's mostly collective, which makes it permanently harder to feel than the needle.

The continued influence effect. Misinformation keeps shaping reasoning after it has been retracted and after the person accepts the correction [7]. This is why simply posting the numbers doesn't settle it, and it's the reason this newsletter teaches a method rather than issuing rulings.

Salience and information overload. The newest of the five, and the most uncomfortable. In a 2025 randomized trial across Spain and Bulgaria, two thousand people were shown the official regulatory product leaflet for a flu vaccine. Hesitancy rose. Making the positive information more prominent did not rescue it, and hesitant participants spent less time reading [8]. The authors attribute this to salience bias and information overload rather than to availability. Note what that means for us: accurate, official, complete information delivered as a wall of it can move people away from the thing it describes. It is the strongest argument I know for teaching a method instead of publishing a correction.

Put those together and you get something structural, not stupid.

Public health success is invisible by construction. When it works, the thing you prevented doesn't happen, so there's nothing to photograph and nothing to remember. The measles ward that was never built appears in nobody’s family album. Every year the intervention succeeds, the evidence for why it's needed grows quieter, while the cost of it stays exactly as visible as it ever was.

So the argument for continuing gets harder every year that it works. That's not a communications failure. It's a property of prevention itself, and it's why the Vanishing Act keeps earning long after the data answered it.

Tier 1, government surveillance data: the measles, polio and pertussis figures, from CDC notifiable disease reporting. Strongest available tier for incidence, because it's a census of reported cases rather than a sample.

Tier 2, regulatory and programme record: the 2000 elimination declaration, the 2000 US switch away from oral polio vaccine, and the FDA post-marketing framework described above.

Tier 2, peer-reviewed behavioral science: references 4 through 8. Note these are a different kind of evidence from the surveillance data. They describe how people reason in general, and reference 8 is a randomized trial of information presentation, not of this claim. None of them are measurements of the measles claim itself and should not be read as such.

Not cited here, deliberately: individual outbreak news coverage. It moves faster than confirmation does, and early counts are routinely revised.

Get the deck of questions: A Skeptic’s Toolkit — 10 Questions for Any Claim

Am I seeing the full picture? The claim uses today’s low counts and omits both the cause of the decline and the trend since coverage fell.

Does the timing support the story? Sanitation improved gradually. The measles and polio curves fell near vertically at vaccine introduction.

Who is missing? Infants under twelve months and the immunosuppressed, who appear in the case data and never in the claim.

What would change my mind? Sustained low incidence in communities where coverage has dropped. That's a real, checkable prediction, and it has been running as a natural experiment for two years.

That's the whole exercise in Does It Have a Receipt: take the thing, ask what paper trail sits behind it, and see how fast the answer arrives.

Next week we leave the safety room and go somewhere the money is louder.

[1] CDC, Measles Cases and Outbreaks (cdc.gov/measles/data-research). 2025: 2,289 confirmed US cases across 48 outbreaks, 90 percent outbreak-associated, 3 deaths. 2026 to date: 2,566 as of 13 August.

[2] CDC, Polio in the United States (cdc.gov/polio), and MMWR 71(33). The 2022 Rockland County, New York paralytic case, vaccine-derived poliovirus type 2 in an unvaccinated adult, with genetically linked virus later found in wastewater across several counties. The US has been polio-free since 1979 and stopped using oral polio vaccine in 2000.

[3] CDC, Pertussis Surveillance and Trends (cdc.gov/pertussis/php/surveillance), and the 2025 Provisional Pertussis Surveillance Report.

[4] Tversky A, Kahneman D. "Availability: A heuristic for judging frequency and probability." Cognitive Psychology. 1973;5(2):207-232.

[5] Ritov I, Baron J. "Reluctance to vaccinate: Omission bias and ambiguity." Journal of Behavioral Decision Making. 1990;3(4):263-277. doi:10.1002/bdm.3960030404.

[6] Rose G. "Sick individuals and sick populations." International Journal of Epidemiology. 1985;14(1):32-38. Rose coined the prevention paradox earlier, in "Strategy of prevention: lessons from cardiovascular disease," BMJ. 1981;282:1847-51.

[7] Lewandowsky S, Ecker UKH, Seifert CM, Schwarz N, Cook J. "Misinformation and Its Correction: Continued Influence and Successful Debiasing." Psychological Science in the Public Interest. 2012;13(3):106-131. doi:10.1177/1529100612451018.

[8] Ingebretsen Carlson J, Puppo F, Roca-Umbert A, Folkvord F, Lupianez-Villanueva F. "An experimental study on the effects of electronic medicinal product information on vaccine hesitancy." Scientific Reports. 2025;15(1):11197. doi:10.1038/s41598-025-96092-6.

For educational purposes only. Nothing in this newsletter is medical advice. Talk to your doctor before making any health decision.

Instagram: instagram.com/roottorx  |  LinkedIn: linkedin.com/company/roottorx  |  Facebook Lab: facebook.com/groups/theroottorxlab  |  Subscribe: RootToRx.com

You are receiving this because you subscribed at roottorx.com.

Disclosure: Root to Rx is published independently by Open Label Media LLC. Views expressed are personal views of Justin Yamashita and do not represent his employer or any affiliated organization. No employer resources or proprietary information are used. Every claim is sourced from publicly available materials.

Root to Rx | 228 Park Ave S, #29976, New York, New York 10003, United States