Miscarriage, the COVID Vaccine, and Calling Fauci to Account

“Trust the science” became a recurring injunction in an age of crisis. But what, precisely, does it ask us to trust? An individual scientist's statement? A government agency's recommendation? The conclusion of a peer-reviewed paper?

Newly released material about the safety of COVID-19 vaccination during pregnancy forces that question back into view.

In January 2021, Anthony Fauci reportedly discussed in private messages with Rochelle Walensky and others whether fever or a strong immune response after a second dose could, “theoretically,” contribute to miscarriage during the first trimester. Several days later, speaking publicly, he acknowledged the need for caution but said there were no safety “red flags” at that point for vaccination during pregnancy.

A paper published in 2025 in Science, Public Health Policy and the Law later analyzed the US Vaccine Adverse Event Reporting System, or VAERS. It argued that all 37 pregnancy- and neonatal-related outcomes it examined produced disproportionately high reporting “safety signals” for COVID-19 vaccines compared with other vaccines. The authors accused governments, medical organizations, and the pharmaceutical industry of overstating safety.

Placed side by side, these facts can support an alarming story: scientists knew of a danger and concealed it from the public. Yet stopping the inquiry there would not itself be scientific.

“Theoretically possible” is not the same as “actually increasing”

The private message attributed to Fauci concerned a theoretical pathway, such as fever, rather than an observed increase in miscarriages. Considering causal pathways before harm has been detected is an ordinary part of scientific risk assessment.

Conversely, saying that no safety signal had been identified did not mean that danger was impossible. It was a narrower proposition: the observations collected up to that point had not revealed an unusual pattern. A theoretical concern and an absence of a detected signal can both be true at the same time.

Subsequent cohort studies and meta-analyses generally did not find higher rates of miscarriage, preterm birth, or stillbirth after COVID-19 vaccination. A reanalysis of the CDC's v-safe pregnancy registry, for example, estimated an age-standardized risk of spontaneous abortion of 12.8 percent—within the expected background range. A Dutch study of 4,640 participants likewise found no statistical association between vaccination before or during pregnancy and increased miscarriage.

Those findings did not prove perfect harmlessness. Observational research remains vulnerable to confounding, selection bias, limited follow-up, and incomplete measurement. A Canadian study published in 2025 observed a small increase in gestational hypertension in a vaccinated group, while finding no increase in preeclampsia, preterm birth, placental abruption, stillbirth, or most other placental outcomes. Scientific safety assessment is not a declaration that risk equals zero. It is the continuing work of stating what outcomes are supported, how confidently, and what remains unresolved.

What does a VAERS “signal” mean?

The critical VAERS study should not simply be ignored. VAERS exists to help detect rare adverse events that clinical trials may miss and to generate questions for more detailed investigation. If an unusual reporting pattern appears, public authorities owe an explanation of how they examined it.

But VAERS numbers alone cannot tell us the probability that a vaccine caused an event.

The system accepts reports of events that happened after vaccination regardless of whether the vaccine caused them. Reports may be submitted by anyone, may contain incomplete information, and may increase because of news coverage or public attention. VAERS also does not follow every vaccinated and unvaccinated pregnant person under the same conditions. It therefore lacks the stable denominators needed for a direct comparison of incidence rates.

A proportional reporting ratio above a specified threshold means that a reporting imbalance deserves investigation. It does not mean that the vaccine multiplied the real-world risk by the same amount. When reports from COVID-19 vaccination—administered at unprecedented scale and under extraordinary public scrutiny—are compared with 34 years of reports about influenza and other vaccines, differences in reporting behavior and surveillance intensity also have to be considered.

When the paper moves from VAERS signals to the language of “catastrophic” effects, there remains an evidentiary distance between early-warning data and a causal conclusion. That does not make the study worthless. Its strongest value lies not in proving that harm occurred, but in identifying questions about follow-up analysis that authorities should answer clearly.

Scientists must explain the route to a conclusion, not merely announce it

Most citizens cannot master immunology, obstetrics, epidemiology, and statistics and then reanalyze every raw dataset. Modern society depends on a division of epistemic labor: we rely on knowledge held by other people and institutions. Trust therefore cannot be eliminated. A person who distrusts a government agency still trusts someone else—the author of a paper, the narrator of a video, or the administrator of a database.

The real question is not whether to trust. It is under what conditions trust is justified.

Experts possess more specialized knowledge and greater institutional authority than most members of the public. Because of that asymmetry, scientific accountability requires more than announcing a final conclusion. At minimum, experts should distinguish four things:

1. Observed facts — What happened, how often, and in which population?
2. Scientific inference — What do those observations permit us to say about causation?
3. Uncertainty — Where are the missing data, possible confounders, and risks of oversight?
4. Policy judgment — By what values were known risks from infection weighed against uncertain risks from vaccination?

If all four are compressed into the word “safe,” citizens may reasonably feel deceived when internal concerns or conflicting data later emerge. A conclusion can be statistically defensible at the time and still be communicated inadequately if the uncertainty behind it is withheld.

As philosopher Onora O'Neill has argued, accountability is not the same thing as releasing the greatest possible volume of information. Publishing unreadable documents or context-free numbers and saying “the data are public” does not make judgment possible. Even when VAERS is open to everyone, transparency can create more misunderstanding unless people are told why report counts are not incidence rates. What is needed is not an information dump, but intelligent transparency: reasoning presented in a form that can be understood, challenged, and revised.

A more candid public statement in early 2021 might have sounded like this:

Direct evidence from the first trimester remains limited, and fever after vaccination presents a theoretical concern. Surveillance so far has not identified a signal of increased miscarriage. COVID-19 infection during pregnancy, however, carries observed risks. We will update our recommendation as the evidence develops.

That wording would not eliminate anxiety. Science's task is not to abolish uncertainty, but to make its structure visible.

The precautionary principle does not point in only one direction

Medical decisions during pregnancy can affect both the patient and a future child. In the spirit of Hans Jonas's ethics of responsibility, scientists should not ignore the possibility of grave and irreversible harm merely because the evidence is not yet complete. Because pregnant people were excluded from initial trials, those recommending vaccination carried an unusually demanding duty of caution and follow-up.

But the precautionary principle cannot apply only to unknown harms from a vaccine. The possibility that forgoing vaccination could lead to infection and harm to the pregnant person or fetus belongs in the same ethical calculation. Inaction is not neutral; it selects a different set of risks.

Accountability therefore means neither “prohibit anything with uncertainty” nor “avoid discussing concern when a benefit is expected.” It means describing the risks of both action and inaction, identifying who bears them, and disclosing which uncertainties received the greatest weight.

Dissenting scientists owe the public the same accountability

The duty to explain does not fall only on governments or mainstream medicine. Researchers challenging the consensus must meet exactly the same standard.

If they use VAERS reports, they must explain how they addressed reporting bias, uncertain denominators, duplicates, coincidental events, and stimulated reporting. If they infer causality from a safety signal, they must also account for why the same increase has not been reproduced in large cohort studies.

The same symmetry should govern conflicts of interest. Pharmaceutical funding is a reason to scrutinize a study carefully, but it does not by itself make the result false. Conversely, receiving no industry funding does not prove that a method is sound. A declaration about conflicts must not substitute for examining data and design.

Nor does being “censored” or “persecuted by the medical-industrial complex” establish that a claim is true. Galileo was excluded by authority and was right; it does not follow that everyone excluded by authority is another Galileo. Heterodoxy is neither correctness nor error. It is a starting point for testing.

Do not demand trust; build institutions worthy of it

Sociologist Robert Merton named organized skepticism as one of the norms of science. Doubt is not science's enemy. But skepticism must be directed not only at authority; it must also interrogate the anti-authoritarian stories we would like to believe.

The legitimacy of public policy does not arise solely from experts possessing the correct answer. It also arises from a process in which officials offer reasons that citizens can examine, answer objections, and correct errors. Explanation is not publicity attached to research. It is part of public science itself.

In a Habermasian sense, the legitimacy of a public decision rests less on the status of an authority than on presenting reasons to citizens and exposing those reasons to criticism. Scientific findings are not decided by vote. But deciding what to recommend, to whom, and on the basis of those findings is a public act containing value judgments. The proper claim is not “obey us because we are experts,” but “this evidence, with these uncertainties, led us to this recommendation.”

Before asking people to trust science, scientists and institutions must show that they are trustworthy. That means publishing inconvenient results as well as reassuring ones, distinguishing theoretical concern from demonstrated harm, and saying in advance what evidence would change their view.

Citizens have obligations too. A dramatic headline or one line from a private message should not replace the work of distinguishing an early-warning signal from an observational study or a replicated result. Scientific judgment in a democracy is neither an expert monologue nor a vote on social media. It must be an exchange of reasons between those who explain and those who question, addressed to the same evidence.

To trust science is not to believe that scientists never make mistakes.

It is to trust that scientists accept a duty to disclose uncertainty, answer criticism, and correct error. When that duty is not fulfilled, the appropriate response is neither obedience nor wholesale rejection. It is the continuing demand for an intelligible account.

Sources

• New York Post report on Fauci's private messages
• JAMA interview from February 3, 2021
• Critical VAERS study in Science, Public Health Policy and the Law
• Official guidance on interpreting VAERS data
• Reanalysis of spontaneous-abortion risk in the CDC v-safe registry (NEJM)
• Dutch cohort study of vaccination before and during pregnancy and miscarriage
• Systematic review and meta-analysis of vaccination during pregnancy (Nature Communications)
• 2025 Canadian study of vaccination around conception and placental outcomes

Note: This essay does not provide individualized vaccination advice. Decisions about vaccination during pregnancy should be made with a qualified healthcare professional in light of current transmission, medical history, gestational age, and the latest guidance in the relevant jurisdiction.