Listen on: Apple Podcasts | Spotify | Amazon Music | Podcast Index | Overcast | YouTube

A new study says a supplement helps you live longer. A medical society tells parents to avoid peanuts—then years later urges early introduction. Hormone therapy is good, then harmful, and now… it depends. If you’ve ever wondered what to believe—and how to separate fact from fancy—you’re in the right place. I’m an open-minded skeptic: I will change my mind when the evidence is strong, but I’m cautious until it is.

In a recent conversation with Dr. Adam Cifu—professor of medicine at the University of Chicago, coauthor of Ending Medical Reversal, and cofounder of the Substack Sensible Medicine—we dug into why medicine sometimes flips, how to spot claims that are out over their skis, and the simple steps you can take to make confident decisions for your own health.

Why Medicine Reverses (More Than You’d Expect)

Most reversals don’t come from bad intentions; they come from moving too fast on ideas that seem biologically plausible. A treatment makes sense “on paper,” early observational data look encouraging, experts are enthusiastic, and the practice spreads before we have the kind of randomized, well-controlled human evidence that actually answers the question.

A few famous examples:

• Peanut introduction. For years, parents were told to avoid peanuts to prevent allergies. Then the randomized LEAP trial showed the opposite: early introduction reduced peanut allergy risk in high-risk infants (NEJM, 2015). Guidance flipped—appropriately—because better evidence arrived.
• Stents for stable angina. When coronary stents saved lives in heart attacks, it felt logical to extend them to people with stable chest pain. But trials like COURAGE (NEJM, 2007) and ISCHEMIA (NEJM, 2020) showed that for many stable patients, stents don’t reduce heart attacks or prolong life compared with excellent medical therapy—though they can help symptoms. The nuance matters.
• Hormone therapy after menopause. For years, estrogen-progestin was pitched almost as a public-health intervention. Then the Women’s Health Initiative found increased risks (e.g., breast cancer, stroke) with combined therapy in many women, forcing a hard rethink. Today’s stance is more measured: consider timing, symptoms, formulation, and individual risk.

The pattern is clear: when better studies arrive, good medicine adjusts. That’s a feature, not a bug. The challenge is living wisely in the gray zones before the dust settles—and staying calm when headlines shout certainty where there isn’t any.

Three Places Hype Outruns Proof

Longevity draws dreamers, and sometimes salespeople. Here are areas where claims often leap ahead of human evidence:

• Mitochondrial booster supplements. The narrative—“your powerhouses are tired; take X to recharge”—is tidy and tempting. What’s missing in many cases are large, long-duration randomized trials showing real-world outcomes that matter (function, disease, survival), not just shifts in lab markers.
• Microbiome miracles. Gut health is fascinating and important, but most “fix your flora, fix your life” promises are still in theory territory. Outside specific conditions (like recurrent C. difficile), robust trials for broad, durable benefits are scarce.
• Biological age tests. Epigenetic clocks and multi-omic panels may predict risk at the population level, but improving your “clock” doesn’t automatically mean you’ll live longer or avoid disease. Surrogate markers aren’t the same as outcomes.

Curiosity is good. Certainty without data is not.

A Practical Playbook You Can Use Today

Here’s how I approach new claims—and how you can, too.

1) Ask what kind of evidence it is.
Anecdotes and mechanisms are a starting point, not an endpoint. Randomized controlled trials in humans—when feasible—give the most reliable answers. If a claim rests on animal data, petri dishes, or uncontrolled “before/after” stories, file it under “intriguing, not proven.”

2) Prefer sources that synthesize evidence—honestly.
For supplements and nutrition, I like Examine.com for transparent, regularly updated summaries. When available, I read Cochrane reviews because they weigh all the trials and, crucially, often conclude “we don’t know yet.” That humility is a green flag, not a red one.

3) Separate “late-breaking” from “settled enough.”
Chat tools and social feeds are fine for brainstorming angles on established topics (say, vitamin D and bone health), but they can mislead on hot-off-the-press studies. For the newsy stuff—like emerging safety signals or early prevention claims—talk with your clinician, look for balanced analyses, and give it time.

4) Use safe N-of-1 experiments when the stakes are low.
When evidence is mixed and the outcome is measurable, test it on yourself—systematically. If you wonder whether melatonin helps you, track sleep for two weeks without it, two weeks with it, and compare objective metrics (sleep latency, awakenings, total sleep time) from a device or validated questionnaire. If you’re debating salt reduction for blood pressure, check home readings twice daily for two weeks at your usual intake, then two weeks lower-salt (while keeping everything else steady). You’re looking for a clear, reproducible signal—not wishful thinking. This approach doesn’t fit everything (e.g., bone density or cancer risk), but for sleep, blood pressure, daytime energy, reflux, or migraine triggers, it can be incredibly clarifying.

5) Watch for red and green flags.

• Red flags: one-size-fits-all cures, heavy reliance on anecdotes, affiliate links to the “solution,” and confident certainty where evidence is thin.
• Green flags: clear discussion of trade-offs, admission of uncertainty, alignment with randomized human data, and advice that doesn’t require buying something from the person giving it.

What to Do When Experts Disagree

First, breathe. Disagreement doesn’t mean “no answer”; it means the answer depends on your context—age, risk, goals, preferences, and values. That’s not a cop-out; it’s how good decisions are made. With hormone therapy, for example, the right call for a 52-year-old within a few years of menopause with severe vasomotor symptoms and low baseline cardiovascular risk may differ from a 62-year-old with multiple risk factors. The science informs the range of reasonable options; you and your clinician choose within that range.

One more mindset shift helps: think in probabilities and trade-offs, not absolutes. Even excellent trials describe averages. Your job is to ask, “How much benefit is likely for someone like me? What are the meaningful downsides? How will we monitor and adjust if it doesn’t help?”

The Bottom Line

Health information has never been more abundant—or more uneven. Some claims are accurate, some exaggerated, and some will be reversed. The antidote isn’t cynicism; it’s optimistic skepticism: be open to new ideas, insist on human outcomes when possible, and embrace “we don’t know yet” as a sign you’re listening to grown-ups.

If you want to dive deeper, listen to my full conversation with Dr. Adam Cifu. We unpack why reversals happen, where the biggest hype cycles live, and how to apply a simple, evidence-first approach to your day-to-day choices.

In the meantime, try one small experiment this week: pick a question you care about, decide what to measure, run a short, safe test, and learn from your own data. Progress, not perfection—and as always, you have the power to make meaningful changes, one small step at a time.