A wrong value slipped through in the table in this paper by Barbara Mintzes and colleagues (2 February, pp 278-9). Among patients who had not requested drugs before, the number who requested at least one drug that had been advertised direct to consumers was in fact 42 [not 52], and the percentage was 3.9% [not 4.8%].