Bias may affect outcome of trials

EDITOR,—Roger N Baird and Mark Lambert—and Peter C Rubin in his commentary—draw varying conclusions about the usefulness of carotid endarterectomy in the prevention of stroke.1 None, however, mention the likelihood of bias in assessments of the outcomes of the two quoted trials of the procedure.2 3 No placebo operations were done in the control groups, so the neurologists assessing the outcomes were almost certainly aware of whether each patient had been operated on. It is unfortunate that their preferred outcome measure, severe ipsilateral ischaemic stroke, is so clearly liable to unconscious bias …