Risk of subsequent thromboembolism for patients with pre-eclampsia

Sometimes an error persists beyond the BMJ and is noticed not by a reader or author but by one of our sister journals preparing the article for publication—in this case BMJ USA. In this Research Pointer by Carl van Walraven and colleagues (12 April, pp 791-2) we allowed a wrong confidence interval to slip through. In the last sentence in the first paragraph of the Methods section, the 95% confidence interval for the specificity (67%) of the ICD-9 codes should be 59% to 74% [not 79% to 94%].