How well can we predict coronary heart disease? Findings in the United Kingdom Heart Disease Prevention Project.

The probability of myocardial infarction developing over five years in a group of middle aged men was predicted with knowledge of their ages, blood pressures, cholesterol concentrations, and smoking habits as recorded in an initial screening examination. Although the top 15% of the risk distribution predicted 115 (32%) of the subsequent cases of myocardial infarction, there was a considerable overlap in predicted risk between those subjects who did and those who did not go on to develop a myocardial infarction. Of the subjects in the top 15% of risk, only 72 (7%) of those initially free of coronary heart disease and 43 (22%) of those initially with coronary heart disease actually developed a myocardial infarction over the subsequent five years. Thus, although a group of subjects at high risk can be identified, among whom will be a high proportion of potential victims of heart attack, many subjects will be wrongly classified. These findings may explain part of the difficulty in persuading patients of the potential benefits of reducing risks and highlight the need for research to improve the prediction of the development of coronary heart disease.