Evaluating online medical information

Start with the job the information is doing

Some pages teach anatomy. Some describe a study. Some sell a product or a procedure. The job matters, because a sales page can borrow the tone of a clinic. Ask who wrote it, whether a clinician’s name is being used with permission, and whether the page would still exist if you could not click through to a purchase. This library does not sell supplements or appointments. A page that exists only to hand you to a shop is not an independent explanation.

Fear is also a sales tool. A countdown, a secret doctors supposedly withhold, or a promise aimed at a single frightening disease is a reason to slow down, not a reason to act the same day.

Separate stories, observations, and experiments

One man’s recovery is a story. It can be honest and still be a poor guide, because you do not know how many men with the same treatment did worse, or what else changed in his care. A study that follows what people already eat or which supplements they already buy can find a link. People who take supplements also differ in other ways. A randomized trial assigns the product by chance and is harder to dismiss as that kind of mix-up. It can still use the wrong dose, the wrong people, or an outcome that does not matter to you.

SELECT is the example to keep. Earlier observations had made selenium and vitamin E look promising. The randomized trial did not show a preventive benefit, and vitamin E alone was later associated with more prostate cancers. A hopeful headline from before that result is not “balanced” against the trial. It is outdated.

Animal and laboratory studies are further away still. A change in a mouse or a dish is not evidence that a person will have the same change.

Read the outcome, not only the adjective

“Lower risk” can mean fewer cancers found, fewer deaths, or a small shift in a blood test. Those are different. The finasteride prevention trial, as summarized for patients by the National Cancer Institute, found fewer prostate cancers and the same number of deaths from prostate cancer. A page that quotes only the first half is incomplete. A page about a procedure should say whether it measured symptoms, side effects, cancer control, or survival, and for how long. Advertising that compares a treatment with other treatments without those details is not a result. It is a claim.

Watch for these patterns:

• A result promised on a fixed timetable, or language that offers a certain medical outcome.
• A statement that a supplement replaces a prescribed medicine.
• A single clinic’s success rate with no description of who was excluded.
• A testimonial presented as typical.
• A product described as approved for a disease when the seller is talking about a dietary supplement. Supplements are not approved the way drugs are before they are sold.

Check the date and the conflict

Guidelines change. The screening article on this site relies on a 2018 Task Force statement and notes that the Task Force has said the topic is being updated. A video reviewed in 2012 is not current just because a website still displays it. Look for the date of the medical source, not only the date the web page was refreshed.

Then look for money. A speaker, a clinic that owns the equipment, or a site that earns a commission can still tell the truth. The reader should know about the money before trusting the emphasis. This project’s advertising page says there are no affiliate product links today. If that changes, the change has to be labeled. It would not turn an article into a medical endorsement.

A short test before you believe a page

1. Does it name the condition clearly, or does it blur enlargement, infection, and cancer?
2. Does it cite a source you can open, such as a government health agency or a journal?
3. Does it describe harms and uncertainty?
4. Does it tell you to stop prescribed care?
5. Would the advice still make sense if nobody were selling anything?

If the answer to the last question is no, keep the question and leave the offer.