“Safe and Effective”: Can Vaccine Studies Even Prove It?

A new JIM paper reveals an overlooked limit in vaccine safety studies: even large ones can’t show vaccines are safer than the diseases they target.

We’ve all heard the line that vaccines are safe and effective, full stop. You’d assume the safety studies had settled the question beyond a reasonable doubt. But a new paper in the Journal of Independent Medicine asks whether those studies can actually provide that kind of certainty.

Dr. Donald Adema, Dr. Paul Thomas and Dr. David Brownstein compare the risk of each disease with the smallest risk its safety studies are able to detect, to show where the evidence falls short and where research is still missing. In the video above, Brownstein, a physician and medical director of the Center for Holistic Medicine, walks through their case with statistician Hector Magno, whose research on disease risk the paper cites.

Brownstein, who says that message has held since before he became a physician, explains what it leaves out:

“Most believe that there is a plethora of vaccine safety studies that have proved beyond a doubt that vaccines are safe and effective, and their use far outweighs the risk of disability or death from the diseases they are supposed to protect against. Yet few are aware of an important limitation of those studies: their statistical limitation.” — Dr. David Brownstein

📖 Read and Download the Full Paper

Why Statistical Power Matters When Measuring Vaccine Safety (JIM Vol. 2, No. 4, 2026)
Authors: Donald S. Adema, Paul Thomas and David Brownstein

The Known Gaps in the Safety Record

Statistical power is only one limit of vaccine safety studies, and the paper starts with the others:

• An incomplete evidence base: In 2012, the Institute of Medicine published the most comprehensive review of vaccine safety studies to date, covering dozens of adverse events, including neurological disorders and autoimmune diseases. For most of them, it found the evidence “inadequate to accept or reject a causal relationship,” often because there wasn’t enough research or because of limits in study design.
• Short studies: Some trials observe for only days or weeks. Very few last years.
• Underreporting: The Vaccine Adverse Event Reporting System, or VAERS, is passive, relying on doctors and the public to report on their own, and passive surveillance systems may receive reports for as few as 1% of serious side effects.

That is why epidemiological studies, which track health outcomes across large groups of people, are important. They can study many conditions over long periods, using more reliable data. But they have limits too.

Vaccines Face a Different Test Than Drugs

Epidemiological safety studies usually compare two groups: a vaccinated group and a control group. Small differences between the groups can be the result of chance, so they aren’t considered statistically significant. The confidence interval defines that range, and its details reveal a study’s statistical power: how small a risk it can actually detect.

This is where vaccines differ from drugs. Drugs are generally designed to treat people who are already sick. Vaccines are given to healthy children to prevent an illness. So, the authors argue, a vaccine’s risk has to be compared with a healthy child’s risk of both catching the disease and suffering serious harm from it.

That comparison sets a high bar. In the paper’s words, the confidence intervals of vaccine safety studies “must be narrow enough to be smaller than the disease risk; otherwise, it is impossible for such studies to prove that a vaccine causes less injury than the disease it targets.”

1 in a Million a Year: The Disease Risk for a Healthy Child

The disease side of the comparison is harder to pin down. Incidence and severity change over time, and some childhood diseases were deadlier a century ago, before modern sanitation, nutrition and medical care. Risk also depends on individual risk factors, and the data are scattered across the medical literature.

Magno’s published work organized those data and accounted for factors such as pre-vaccine declines in mortality, so disease risk could be measured more accurately. The paper presents the risk of permanent disability or death for children under 10 without a vaccine. For children at normal risk of exposure, the paper finds, that risk was low:

• Hepatitis B: About 97% of fatal cases occur in people born to infected mothers, people who engage in high-risk behavior, and people who live with an infected person or in a community where hepatitis B is unusually common. For a child without those risk factors, the annual pre-vaccine risk was 1 in 7 million.
• Measles: About 92% of cases of permanent disability or death occur in patients with low vitamin A. For a child with normal vitamin A levels, the annual pre-vaccine risk was about 1 in a million.

Across the diseases vaccines target, Magno explains, the pattern holds:

“In fact, when considering most of the diseases targeted by vaccines, the median annual risk of permanent disability or death was about one in a million, with a lifetime risk in the range of one in tens of thousands.” — Hector Magno

Are Vaccines Safer Than the Diseases?

That is the question the paper puts to large epidemiological studies, some involving more than half a million children: did they have enough statistical power to detect a vaccine risk as low as the risk of the disease the vaccine targets? Two examples from the paper:

• Hepatitis B vaccine: A study of 135,000 children published in Nature Medicine could not rule out a 1 in 3,900 vaccine risk of demyelinating diseases (nerve conditions such as multiple sclerosis).
• MMR vaccine: A study of 537,000 children published in the New England Journal of Medicine could not rule out a 1 in 1,900 vaccine risk of autistic disorder, based on the study’s unadjusted data.

Both figures are larger than the lifetime disease risk Magno describes. Brownstein spells out what that means:

“The paper shows that even large vaccine studies do not meet the burden of proof of safety. They do not have the statistical power to rule out the possibility that vaccines may cause more harm than good.” — Dr. David Brownstein

A Word of Caution and a Way Forward

Magno adds an important word of caution:

“Limited statistical power does not mean proof of harm. However, it does leave unanswered questions about the safety of vaccines.” — Hector Magno

More research is needed to find the answers, and the paper points to where it should start: a statistical power database showing the smallest risk each vaccine safety study is able to detect. Healthcare professionals and families could compare those limits with the risk of each disease, and use that comparison to guide public health policy and personal decisions about vaccination.

Read and download the full study at the Journal of Independent Medicine.

Related Reading

• Post: Journal of Independent Medicine Marks America’s 250th With New Edition Advancing Independent, Patient-Centered Science
• Guide: Parent’s Guide to Childhood Immunizations
• Post: Childhood Vaccines: An Up-to-Date Review on Efficacy and Risks with Dr. Elizabeth Mumper
• Post: DARE-SAFE: A New Standard for Comparing Vaccine and Drug Safety
• Guide: Navigating Measles: An Integrative Approach to Care and Recovery