New Zealand guidelines for selecting patients for drug management of cardiovascular disease (CVD) risk factors are based on US risk prediction research from the 1960s and 70s. Since then, the development of new drugs for treating CVD risk factors and changes in risk factor profiles has meant that current risk prediction algorithms are likely to be unreliable. New risk prediction equations have been developed internationally but none have incorporated drug treatment during follow-up. Using clinical data obtained in primary care by a locally developed web-based decision support programme (PREDICT), linked to comprehensive national healthcare datasets, we have an opportunity to develop new risk prediction algorithms based on very large datasets that include comprehensive follow-up drug treatment and laboratory data on individuals. This research will specifically investigate the impact of drug treatment during follow-up and novel exposures from laboratory testing on the accuracy of risk prediction.