Discriminative ability of a risk-prediction tool derived from the Framingham Heart Study compared with single risk factors

Abstract

AIM: To compare the discriminative ability of a multivariate risk-prediction model with individual continuous cardiovascular risk factors in a free-living population. METHODS: Standard cardiovascular risk factors were measured in 6354 participants (4638 men and 1716 women) aged 35 to 74 years with no history of cardiovascular disease either enrolled on Auckland general electoral rolls or employed by a New Zealand-wide multi-industry corporation, in 1992-3. The sensitivity and specificity of individual risk factors versus a five-year cardiovascular risk-prediction equation and the corresponding New Zealand risk charts in predicting hospitalisation and mortality from cardiovascular disease in the subsequent five-year period were estimated over a range of risk thresholds. RESULTS: Discrimination between individuals who had or did not have subsequent cardiovascular events was poor for individual risk factors. Increasing age had significantly more discriminability than blood pressure or lipids and the Framingham Heart Study risk tool had better discriminability than any single risk factor. CONCLUSIONS: A Framingham risk equation and corresponding New Zealand risk charts discriminate between individuals who will or will not experience hospitalisation or death from cardiovascular events. Discriminability is only modest but it is better than that achieved using individual risk factors.