The guideline that was updated

A clinical guideline changed its advice, and the newspapers reported that doctors had been wrong for twenty years. The guideline's own documents tell a different story.

The original recommendation was made when the available evidence consisted of two small trials and a plausible mechanism. It was issued with a stated strength — weak, in the guideline's own grading system — and a note that it should be reviewed when larger trials reported. Eighteen years later they reported, and the recommendation changed. The system worked exactly as designed, and the only thing that had gone wrong was the public communication, in which a weak recommendation had been reported, and repeated by clinicians, as an instruction.

Grading systems exist precisely to carry this information. A strong recommendation means that almost all well-informed patients would choose the same thing and that the evidence is solid. A weak one means the balance is close, the evidence is limited, and different patients would reasonably choose differently. The two look identical when they arrive as advice, which is why the grade has to travel with the sentence.

The uncomfortable part is that the weak recommendations are the ones most likely to be useful in conversation, because they are the ones where the patient's own priorities decide. Stripping the grade out makes the advice easier to deliver and removes the patient from the decision.

The practical move for a reader is to find the grade. Most published guidelines state it beside each recommendation, and the difference between 'we recommend' and 'we suggest' is not a stylistic choice: in several major systems those two verbs are defined terms carrying exactly this distinction.