An article can be accurate the day it’s published and misleading by the time you read it. Nothing about the text changes. The evidence underneath it does.
That’s the part search engines and social feeds don’t tell you. A page looks current (clean design, working links, top of the search results) while its core recommendation is stale, superseded by a trial, a safety alert, a new diagnostic cutoff, or an updated guideline. The publication date only tells you when something was written. It says nothing about whether anyone has gone back to check it.
A timestamp isn’t enough. Medical content needs something closer to an expiration date.
This isn’t hypothetical. A 2014 analysis of clinical guidelines found that roughly 1 in 5 recommendations was outdated within 3 years. An earlier analysis of 100 systematic reviews found a median of 5.5 years before a signal appeared that an update was needed, but nearly a quarter, 23 percent, reached that point within 2 years. Seven percent already showed a signal that an update might be needed by the time the original review was published.
Real physician voices, twice a week
Free, and one click to unsubscribe.
None of this means every claim has a countdown timer. How insulin works isn’t going anywhere; an article comparing this year’s obesity drugs might not survive the quarter. Some claims barely decay. Others decay fast. You can’t tell which is which from a publication date alone. That is the problem with treating a publication date as if it tells you anything about freshness.
It gets worse once an article leaves its original home. Search engines keep old pages ranking. Social platforms resurface them stripped of context. AI tools lift a single sentence and hand it back as a current fact. Every step pushes the claim further from the evidence that once supported it.
Here is a practical fix: Every evidence-based medical article should visibly carry four things. The date through which the evidence was last reviewed. The date of the next planned review. The developments that would trigger an earlier review. And a record, visible rather than buried in a changelog, of what has actually changed.
That trigger clause matters most. Medicine doesn’t wait for your calendar. A regulator can issue a safety notice tomorrow. A large randomized trial can change the balance of benefit and harm. A professional society can revise a diagnostic threshold. An article shouldn’t sit untouched until its annual review after its central claim has already been challenged.
Not every article needs the same clock. Low-volatility educational content can run on a longer interval. Treatment comparisons, screening guidance, drug safety information, and fast-moving fields such as clinical AI need shorter intervals and clearer triggers. If a publisher cannot keep up with a high-risk article, the page should say so and direct readers to current guidance.
This isn’t an argument for deleting the archive. Old articles still have historical and educational value. The point is to tell readers which category a piece belongs to: archive or active guidance. Stating “Evidence reviewed through March 2026” plainly does more work than quietly changing the date at the top of the page. Readers deserve to know whether a piece was actually reassessed or merely republished.
Fixing this isn’t only the publisher’s responsibility. Writers should identify which claims are most likely to age quickly. Editors need a real review queue, not a habit of trusting that someone will notice when something has gone wrong. Health organizations should treat this as patient safety and public communication, not website maintenance to address eventually.
Admitting that content can expire doesn’t cost credibility. It builds it. Authority in medicine was never supposed to be permanent; it is conditional on the evidence holding up. A visible update trail earns more trust than an article that keeps sounding certain long after the science has moved on.
We already put expiration dates on medicines because stability matters. Medical information deserves the same discipline. Nobody needs the exact day a claim goes bad, only when it was last checked, when it is due to be checked again, and what would move that date forward.
An article shouldn’t stay persuasive just because nobody bothered to take it down.
Aman Mishra is a medical writer.
