|
|
|
|
|
Reader Opinions How Did Evidence-Based Medicine Come to Oppose Vaccines and Undermine
Public
Health? |
|
Author:
Michael A. Stoto, PhD Note: This article appeared simultaneously on SSRN on September 16, 2026. Intended to restore trust, President Trump’s August 2026 executive order on “gold standard” child vaccine policy relies on a report that contains no evidence regarding vaccine efficacy or safety and is full of errors. Although it remains to be seen whether the order has any legal power, it will succeed in sowing confusion and contribute to waning measles vaccination rates while cases are soaring. This is the latest attempt by Trump and Health and Human Services Secretary Robert F. Kennedy Jr. to use “gold standard,” “scientific,” and “evidence-based medicine (EBM)” to justify their policies. EBM is an approach to grounding clinical and public health decisions in the best available research. Yet some of EBM’s most visible champions have invoked a narrow, interpretation that elevates one study design – randomized clinical trials (RCTs) – to cast doubt on vaccines, delegitimize long‑standing regulatory and advisory processes, and justify sweeping policy changes. Together, these actions reveal a striking inversion: the rhetoric of evidence-based medicine is being used selectively to question vaccine benefits while accepting weak or nonexistent evidence for vaccine harms. Beyond the dubious rationale and vast consequences of these changes for vaccine access and child health, these actions delegitimize the contributions of science and undermine trust in public health. Ironically, restoring confidence in vaccines – and public health generally – will require preserving the well-developed vaccine regulatory and advisory system Kennedy seems determined to dismantle, which is actually a crown jewel of evidence-informed public health. Evidence-Based Medicine during the pandemic: from nuanced framework to RCT absolutism As described by Guyatt and others in 1992, Evidence-Based Medicine advocated the “conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients.” RCTs are regarded as the top of the “evidence pyramid” because they help establish a causal relationship. Case studies cannot establish cause and effect, so are at the bottom of the pyramid. There are many observational and quasi-experimental designs between these two poles that can provide useful evidence. Whatever the design, theoretical support and careful planning and execution are critical. This is especially true when RCTs are infeasible, as in many public-health interventions. Some EBM adherents see RCTs as not just the best but the only acceptable type of evidence. During the pandemic, for instance, some argued that mask requirements and social distancing measures were not supported by RCTs. However, observational studies showed that wearing masks, supported by mask mandates, generally reduced the transmission of SARS-CoV-2 infection. Based on observational evidence alone, Ioannidis wrote in March 2020 that the virus might take only 10,000 lives in the US. In April, he and others reported on a sample of Santa Clara County residents tested for antibodies to the virus. Serious biases in this and related analyses seriously underestimated the infection fatality rate, making the virus seem less harmful, including to children. These analyses falsely shaped public understanding of pandemic risks and the benefits of interventions, including vaccination. In 2025, Ioannidis and others published a study that underestimated the benefits of COVID-19 vaccines and exaggerated their harms, contributing to the false narrative that vaccines are dangerous and not effective. This was the result of assumptions – all biased in the same direction – about the infection fatality rate and other model parameters, simplifications in model structure such as ignoring the impact of vaccines on transmission and ignoring the impact on hospitalizations and non-health benefits to vaccinated individuals and the community. The role of RCTs in vaccine approval The EBM community showed no concern about the efficacy of the original COVID-19 vaccines, but some questioned the need for, and safety of, vaccines for new strains of the virus. In 2025, Prasad and Makary published a commentary calling for placebo-controlled randomized trials for new COVID-19 vaccines for all but the most high-risk individuals. This policy demonstrates a profound misunderstanding of ethics, epidemiology, and biostatistics. Ethics: Efficacy was established in 2020 in RCTs with 43,500 and 30,000 participants. Once proven effective, it is unethical to withhold the vaccine from the control group. Biostatistics: RCTs must be exceptionally large to identify rare adverse events. Post-marketing surveillance systems covering millions of patients provide better information on safety. Epidemiology: Once shown to be effective, surveillance on circulating strains and immunobridging studies are needed to see whether the viral strains incorporated in the vaccine match those circulating in the population. Requiring RCTs would delay the availability of the vaccine, perhaps until it no longer matches the circulating strain. Assessing vaccine risks Assessing vaccine risks requires the same evidentiary discipline that EBM claims to defend. Yet Kennedy and his colleagues have long argued that vaccines are dangerous, perhaps causing more deaths than COVID-19 cases they prevented, and even stated there are no vaccines that are safe and effective. Once in office, his associate Tracy Høeg was given access to the Vaccine Adverse Event Reporting System (VAERS), which contains unverified reports of side effects or bad experiences with vaccines submitted by doctors, patients, or even someone who sees a report on social media. In September 2025, officials announced that they planned to link vaccines to the deaths of 25 children. Two months later, Prasad asserted in an internal email that “at least 10 children have died after and because of receiving COVID-19 vaccination. These deaths are related to vaccination (likely/probable/possible attribution made by staff).” No details were included, but Prasad called for stricter regulation of all vaccines, not just COVID-19, based on case reports rather than RCTs. Shortly afterwards, MedPage Today reported that Prasad’s assertion was false. FDA scientists analyzed 7, not 10, cases, and all had an alternative explanation for the cause of death. Revisions to the childhood vaccine schedule In January 2026 the CDC sharply reduced the number of vaccines that are routinely recommended for children. Described as a “scientific, evidence-based, data-driven response,” the report contains no evidence at all regarding vaccine efficacy or safety. Echoing Prasad and Makary’s call for more RCTs, the report ignores many vaccine studies – randomized and observational – that have been done over the years. CDC’s latest decisions were never considered by CDC scientists or the Advisory Committee on Immunization Practices (ACIP). Rather than evidence about vaccine efficacy and safety, Caplan suggests that the changes are driven by a moral argument that prioritizes personal freedom to choose, regardless of the health consequences. Suppression of non-RCTs In April 2026, Jay Bhattacharya, who was temporarily overseeing the CDC, suppressed the publication of a report finding that the 2025-26 COVID-19 vaccine helps protect against serious illness by reducing the risk of hospitalization and emergency department/urgent care visits, adding protection for a population with significant existing immunity from previous infections and vaccinations. The study was eventually published in the peer-reviewed JAMA Network Open. As explained by statistician Natalie Dean, the study’s “test-negative” design estimates vaccine effectiveness by comparing vaccination status in individuals seeking care for COVID-19-like illness (and test positive) are compared to controls in the same care sites who test negative. Because the U.S. has no large cohorts of vaccinated and unvaccinated populations, this design has been used for decades for influenza vaccine monitoring. Bhattacharya’s justification does not acknowledge the long history and established merits of the study’s design. Its statistical arguments are inconsistent with those offered by statistician Martin Kulldorff (another EBM advocate), neither of which hold up to scrutiny. Given the established record of efficacy and safety of mRNA vaccines, it is hard to see the suppression as anything other than an undue preference for RCTs. Restoring confidence in vaccines and public health requires preserving the U.S. vaccine process EBM has three foundational principles: (1) not all evidence is created equal and medical practice should be based on the best available evidence; (2) it is better to be led by the totality of evidence rather than by cherry-picking individual studies that support a particular claim; and (3) decisions can only be reached by considering the values and preferences of the person being treated. Bhattacharya and his colleagues effectively treat the first principle as an absolute preference for RCTs, even when observational designs are better suited, e.g. for detecting rare adverse effects. In contrast to the second principle, they seem to frequently dismiss studies that demonstrate vaccine benefits and safety. The U.S. vaccine regulatory and advisory system, developed and refined over more than half a century, begins with FDA’s review of laboratory and clinical data as well as RCTs to assess a vaccine’s efficacy and safety. Post-marketing surveillance systems, including VAERS and the Vaccine Safety Datalink, identify potential risks and investigate them in large clinical datasets to assess safety. Consistent with EBM’s second principle, the FDA approval process incorporates many different types of evidence: virology, immunology, and molecular biology to characterize pathogens; epidemiology and biostatistics to identify who is at risk, routes of transmission, and the impact of prevention and treatment strategies; and social, behavioral, and managerial sciences to develop effective public strategies. Greenhalgh and colleagues’ call for the EBM community to embrace mechanistic evidence on a par with RCTs is a welcome step. Once approved, CDC in conjunction with the Advisory Committee on Immunization Practices (ACIP) makes recommendations about the vaccine’s use. CDC’s recommendations consider efficacy and safety in different groups plus epidemiologic information about the risk of infection, as well as practical issues such as the standard schedule for well-baby care, thus aligning with the second and third EBM principles. ACIP’s “Evidence to Recommendation” framework systematically evaluates factors like the evidence quality, balance of benefits vs. harms (including safety data), and cost-effectiveness, to create comprehensive public health guidelines. And consistent with the third EBM principle, it considers affected populations' values. Impact on policy, health outcomes, and trust in public health The August 2026 Executive Order attempts to reinstate Kennedy’s vaccine recommendations that were blocked by the court on procedural grounds, resurrect long disproved claims that vaccines cause autism, split the MMR vaccine into three components (despite its infeasibility), and urge states to loosen school vaccine requirements (which are not a federal issue). Rather than “gold standard” science, the order is shot through with errors, rests on a selective view of evidence, and risks deepening public confusion at precisely the moment measles vaccination rates are waning and cases are soaring. Three weeks later, medical professional organizations issued their own guidelines for influenza, COVID-19, and respiratory syncytial virus vaccines. These were based on an evidence review conducted by the Vaccine Integrity Project at the University of Minnesota and a process like the one CDC abandoned, and many states are using these to set their own policies. While reasonable people can differ about the benefits, harms, and costs of some public health policies, the benefits of childhood vaccines to individuals and the community clearly outweigh the harms. But rather than argue the value questions, Kennedy and some prominent EBM advocates have argued the science. Beyond the consequences for vaccine access and child health, each side having its own “facts” can only delegitimize science and thus undermine trust in public health in general. Although the motivation for the EBM experts is unclear, their actions align with Kennedy’s efforts to undermine U.S. vaccine policy. The implications are already being felt in waning measles vaccination rates while cases are soaring. More broadly, they sow doubt on the scientific basis for vaccine policy, undermining trust in public health in general. Conclusion Evidence‑based medicine was never meant to privilege a single study design above all others, nor to provide rhetorical cover for policy changes made without credible evidence. Yet in recent years, a distorted version of EBM—one that treats randomized trials as the only admissible evidence and disregards decades of mechanistic, observational, and implementation research—has been marshaled to oppose vaccines and to weaken core public‑health institutions. The CDC’s 2026 revisions to the childhood vaccine schedule, the suppression of non‑RCT evidence on COVID‑19 vaccines, and the close alignment of certain EBM advocates with Robert F. Kennedy Jr.’s agenda all reflect this troubling trend. Restoring public confidence in both vaccines and EBM requires moving in the opposite direction. During the pandemic, the emphasis on “following the science” without explicit consideration of harms arguably contributed to conservatives’ declining trust in public health. Although Kennedy’s moves are intended to increase trust, Naomi Oreskes argues that we are warranted in placing “informed trust” in the “critically achieved consensus of the scientific community.” This consensus is informed by careful consideration of empirical evidence, which come in many forms, so a focus on one method (such as RCTs) above all others is a kind of “fetish.” Oreskes also reminds us of the critical role of the social character of science. Objectivity depends on scientists’ participation in the give-and-take of critical discussion, because the practices and procedures of scientific communities increase the odds that scientific consensus is reliable. Rebuilding trust requires that public health officials lead with honesty about uncertainty, explain their reasoning, and give people a genuine voice in decisions. The vaccine regulatory and advisory system that Kennedy wants to change is actually a model for evidence-informed policy that does precisely this by integrating diverse forms of evidence and incorporating stakeholder values. Public‑health leaders should reaffirm EBM’s three core principles: that not all evidence is created equal but many forms of evidence matter; that policy should be guided by the totality of high‑quality evidence, not isolated studies that support a preferred position; and that decisions must incorporate the values and preferences of affected individuals and communities. Protecting the independence and integrity of the FDA and ACIP, transparently communicating how evidence is weighed, and resisting efforts to sideline robust observational and mechanistic data are necessary steps. The U.S. vaccine regulatory and advisory system, far from being an obstacle to reform, remains a model of evidence‑informed policymaking. Preserving and strengthening that system is our best hope for aligning EBM with its original purpose: improving health by making decisions that are both scientifically sound and socially responsive. ■ |
|
|
