This smashing little piece explores those situations where a new treatment is so good that a randomised trial is unnecessary. The mother’s kiss technique for removing foreign bodies from pre-school nostrils is a memorable example; there are many others from the field of surgery, and a formula for measuring real effects from background noise. (Conflict of interest: I often work with Paul Glasziou and he shows great patience with my signal-to-noise ratio).

A great many exercise ECGs are done for reassurance rather than confirmation of angina, but most patients continue to experience the same symptoms after a negative test, and remain unreassured, as the control group of this study illustrates. To spare their anxiety, and stop them coming back to you after each twinge of musculoskeletal chest pain, seize the opportunity to give them an information leaflet explaining the significance of a normal test, and then back this up with a personal chat.

We call it sinusitis though we’re rarely all that sure how to diagnose it. “It’s gone on for more than a week: worth a course of amoxicillin