Please Follow us on Gab, Minds, Telegram, Rumble, GETTR, Truth Social, X, Youtube, Instagram
Epidemiologist Eyal Shahar, writing for the Brownstone Institute, takes apart the CDC’s annual flu-vaccine effectiveness estimates. Each year the CDC publishes an estimate of how well the flu vaccine performed in the previous season. The NIH director recently criticized the test-negative design behind those estimates, and he was right to do so. The design is a two-edged sword: restricting the sample to people who sought medical care may reduce one kind of bias while introducing another — collider bias — that is far less appreciated. The net effect remains unknown.
That is not the only problem. Consider a large study of the flu vaccine in 2022–2023, a season when the vaccine was well matched to the dominant strain. It drew on the CDC-linked VISION Vaccine Effectiveness Network, and its results fall apart under scrutiny.
The first issue is confounding by the background risk of infection. Risk was high from October through December 2022 and low from January through March 2023. But vaccination tracks the calendar — most people got their shots by the end of December. As a result, the vaccinated accumulated more exposure time during the low-risk months and the unvaccinated during the high-risk months. That imbalance alone can make a shot look effective even when it is not. In an extreme illustration, injecting people with saline at the start of the low-risk period would appear to "work."
The same bias was documented for the Covid vaccines during the pandemic, including in a study from Ontario, Canada. It had not been recognized in the flu context, where the vaccine rollout typically follows the rising wave and finishes around the winter peak.
A second problem is immortal time bias. The study’s authors excluded events among vaccinated patients that occurred within 14 days of vaccination. Excluding those early events is a well-known source of bias that pushes estimates in the vaccine’s favor. Removing that bias in Covid studies has cut effectiveness numbers in half. Apply the same correction here, and estimates in the 30-to-40 percent range could drop to 20 percent or lower. How many early events were excluded? The number is hidden — nearly 2,000 outpatient encounters were dropped with no breakdown.
Other findings do not add up. Adjusted effectiveness against hospitalization was almost always lower than against ordinary outpatient visits — the opposite of what you would expect, since more severe outcomes should show similar or stronger protection. The authors call this "similar." It is not.
Then there is the healthy vaccinee bias. Across nearly every analysis, effectiveness shrank after adjustment, because the vaccinated were simply healthier than the unvaccinated to begin with. Regression models, no matter how elaborate, cannot fully scrub that out. The authors leaned on unusually complex methods — cubic splines, inverse-probability weighting, boosted regression trees — described so vaguely that no outside researcher could replicate the analysis even with the data in hand.
The age results are equally strange. We expect weaker protection in the elderly because immune response declines with age. Yet the study reported much stronger effectiveness against hospitalization in older adults (41 percent) than in younger ones (23 percent). The authors’ explanation — that seniors received "enhanced" vaccine products — fails to explain why that benefit showed up only for hospitalization and not for outpatient visits.
Most troubling is the outcome among hospitalized elderly flu patients. The case fatality rate was 50 percent higher among the vaccinated (4.0 percent versus 2.7 percent). The authors did not adjust for baseline differences and simply called the outcomes "similar." Following their own logic, one might just as easily wonder whether vaccination raised the risk of death in those patients.
The author chose to dissect a single paper rather than attack the methodology in the abstract, precisely because this one study displays so many of the flaws that run through CDC-based flu-vaccine research. Randomized trials will never be run, so better observational designs — cohort studies tracking both symptomatic and asymptomatic infection, or regression discontinuity analysis — are needed. So far, neither has shown the annual flu shot to be the reliable shield the official estimates claim.
Get the latest on Georgia delivered straight to your mailbox.
