In the linked randomised controlled trial (doi:10.1136/bmj.a3112), Ekeberg and colleagues compare the effectiveness of ultrasound guided corticosteroid injection in the subacromial bursa with systemic corticosteroid injection in people with rotator cuff disease.1 They found no significant difference in pain and disability between the two groups after six weeks. This suggests that the exact location of corticosteroid injections is not important.

The diagnosis of patients with shoulder pain can be difficult.2 This is illustrated by Ekeberg and colleagues’ study, in which general practitioners referred patients with suspected rotator cuff disease to an outpatient clinic, and those with confirmed disease were entered into the study. They referred 312 patients, but 140 were subsequently excluded because they had other diagnoses.1 Rotator cuff disease is diagnosed in up to 70% of people …