The reading room: two guides, five articles and where to start

Every article and guide on this site follows the same rules, and they are worth stating once in the place a reader is most likely to be checking.

Citations are pulled from the NCBI summary service rather than written from memory, and every record is checked for retraction at paper level and at author level before it is printed. Where a widely quoted paper carries a retraction notice it is named and refused rather than quietly dropped.

Amounts are the amounts a trial used, with the population and the measured outcome attached. Where the published record points away from a product this desk sells, it is printed anyway.

• Every external citation links to its PubMed record so it can be checked in about twenty seconds.
• Every price carries the date it was read off the seller's page.
• Every label fact is taken from the artwork and quoted as printed.
• Nothing on these pages is medical advice, and no named medical reviewer is invented.

It is a fair question and it deserves an answer that is not a slogan.

A retail desk that publishes only product pages has one kind of relationship with a reader: it wants a sale today. The pages in this room are written for a different one. They survive the product, they are about the category rather than about one bottle, and several of them will make a careful reader less likely to buy rather than more.

That is deliberate rather than noble. A reader who buys a capsule expecting a change in week one is a refund in six weeks and somebody who never comes back. A reader who buys knowing that the trials behind the ingredients measured arteries at four and twelve weeks is buying something they can actually evaluate. The second reader is better for this desk as well as better served.

The pages also do work the product pages cannot. The checklist applies to every product on this shelf, not only to the one sold here. The evidence guide prints the four interventions with randomised cognitive evidence, none of which is sold here. Both belong somewhere a reader can find them before deciding, rather than buried inside a page whose job is to sell.

• The checklist works on any product, including the ones this desk does not sell.
• The evidence guide names four things with randomised evidence, and none of them comes in a bottle.
• Two of the five articles argue against a purchase made on the wrong expectations.
• Every claim in the room links to the record behind it.
• Nothing here is written to be skimmed on the way to a checkout.

Five articles and two guides is more than anybody wants to read before buying a capsule, so here is the short route through them for four common situations.

If you are deciding between two products on a shelf, start with the buyer's checklist. It is twelve questions, it works on any product in this category, and it will usually shorten a shortlist faster than reading either product's own pages.

If you want to know whether the category is worth anything at all, start with the evidence guide and the article on what actually helps memory. Between them they set out the four interventions with randomised cognitive evidence and where supplements sit relative to them, which is the context every other page here depends on.

If you have already bought and are waiting, read the timeline article. If you have the bottle in your hand and the label is confusing you, read the two on labels: the one about a label with no numbers on it, and the one about a bottle and a sales page disagreeing.

Every route ends somewhere a claim can be checked rather than somewhere a purchase is asked for.