It has been suggested that use of cardiovascular risk tables in hypertensive patients might improve clinical management. As part of a randomised controlled trial, we evaluated the accuracy of health professionals' use of the New Zealand risk tables in primary care consultations. Chance-corrected agreement between health professionals' assessments of absolute risk was only moderate (weighted kappa = 0.56 at the 12-month follow-up). Inaccurate use of cardiovascular risk tables may be a barrier to effective implementation of research evidence in the management of hypertension.