RFK Jr. says he's following 'gold standard' science. Here's what to know

The gold standard can differ because science and medicine is complicated and everything cannot be tested the same way. That term simply refers to the best possible evidence that can be gathered.

“It completely depends on what question you’re trying to answer,” said Dr. Jake Scott, an infectious disease physician and Stanford University researcher.

There are many different types of studies. The most rigorous is the randomized clinical trial.

It randomly creates two groups of subjects that are identical in every way except for the drug, treatment or other question being tested. Many are “blinded studies,” meaning neither the subjects nor the researchers know who is in which group. This helps eliminate bias.

It is not always possible or ethical to conduct these tests. This is sometimes the case with vaccine trials, “because we have so much data showing how safe and effective they are, it would be unethical to withhold vaccines from a particular group,” said Jessica Steier, a public health scientist and founder of the Unbiased Science podcast.

Studying the long-term effect of a behavior can be impossible. For example, scientists could not possibly study the long-term benefit of exercise by having one group not exercise for years.

Instead, researchers must conduct observational studies, where they follow participants and track their health and behavior without manipulating any variables. Such studies helped scientists discover that fluoride reduces cavities, and later lab studies showed how fluoride strengthens tooth enamel.

But the studies have limitations because they can often only prove correlation, not causation. For example, some observational studies have raised the possibility of a link between autism risk and using acetaminophen during pregnancy, but more have not found a connection. The big problem is that those kinds of studies cannot determine if the painkiller really made any difference or if it was the fever or other health problem that prompted the need for the pill.

Scientists can learn even more when they see how something affects a large number of people in their daily lives.

That real-world evidence can be valuable to prove how well something works — and when there are rare side effects that could never be detected in trials.

Such evidence on vaccines has proved useful in both ways. Scientists now know there can be rare side effects with some vaccines and can alert doctors to be on the lookout. The data has proved that vaccines provide extraordinary protection from disease. For example, measles was eliminated in the U.S. but it still pops up among unvaccinated groups.

That same data proves vaccines are safe.

“If vaccines caused a wave of chronic disease, our safety systems — which can detect 1-in-a-million events — would have seen it. They haven’t,” Scott told a U.S. Senate subcommittee in September.

Simply publishing a paper online is not enough to call it open and transparent. Specific things to look for include:

— Researchers set their hypothesis before they start the study and do not change it.

— The authors disclose their conflicts of interest and their funding sources.

— The research has gone through peer review by subject-matter experts who have nothing to do with that particular study.

— The authors show their work, publishing and explaining the data underlying their analyses.

— They cite reliable sources.

This transparency allows science to check itself. Dr. Steven Woloshin, a Dartmouth College professor, has spent much of his career challenging scientific conclusions underlying health policy.

“I’m only able to do that because they’re transparent about what they did, what the underlying source resources were, so that you can come to your own conclusion,” he said. “That’s how science works.”

Anecdotes may be powerful. They are not data.

Case studies might even be published in top journals, to help doctors or other professionals learn from a particular situation. But they are not used to making decisions about how to treat large numbers of patients because every situation is unique.

Even single studies should be considered in the context of previous research. A new one-off blockbuster study that seems to answer every question definitively or reaches a conclusion that runs counter to other well-conducted studies needs a very careful look.

Uncertainty is baked into science.

“Science isn’t about reaching certainty,” Woloshin said. “It’s about trying to reduce uncertainty to the point where you can say, ‘I have good confidence that if we do X, we’ll see result Y.’ But there’s no guarantee.”

If you come across a research paper online, in a news story or cited by officials to change your mind about something, here are some questions to ask:

— Who did the research? What is their expertise? Do they disclose conflicts of interest?

— Who paid for this research? Who might benefit from it?

— Is it published in a reputable journal? Did it go through peer review?

— What question are the researchers asking? Who or what are they studying? Are they making even comparisons between groups?

— Is there a “limitations” section where the authors point out what their research cannot prove, other factors that could influence their results, or other potential blind spots? What does it say?

— Does it make bold, definitive claims? Does it fit into the scientific consensus or challenge it? Is it too good or bad to be true?

___

AP Medical Writers Lauran Neergaard in Washington and Mike Stobbe in New York contributed to this report.

___

The Associated Press Health and Science Department receives support from the Howard Hughes Medical Institute’s Department of Science Education and the Robert Wood Johnson Foundation. The AP is solely responsible for all content.