In this paper by Julia Hippisley-Cox and colleagues (25 March, pp 842-5) the fourth potential confounder in table 2 should have read “rural practice v urban practice” (not urban practice v rural practice) as teenage pregnancy rates were lower in rural areas.