Prediction of coronary heart disease risk by Framingham and SCORE risk assessments varies by socioeconomic position: results from a study in British men

Abstract

Evidence is limited on performance of the Framingham risk score (FRS) in different socioeconomic groups; similar limitations apply to the Systematic Coronary Risk Evaluation (SCORE). We examined the performance of coronary risk prediction systems in different socioeconomic groups in British men. In a socially and geographically representative cohort of British men aged 40-59 between 1978 and 1980, predicted 10-year coronary heart disease (CHD) (fatal and non-fatal) risk was calculated using FRS, and CHD mortality using SCORE. Prevalent cardiovascular disease cases were excluded. Occupational social class ranged from I (professionals) to V (unskilled workers), and was summarized as non-manual (I, II, III non-manual) and manual (III manual, IV, V). Both FRS and SCORE over-estimated 10-year CHD risk; over-prediction by both was particularly marked in high social classes. With FRS, predicted/observed risk fell progressively from 2.30 in social class I to 1.19 in social class V. Sensitivity of FRS at a ≥20% threshold (27% of men) fell from 53% to 37% from social class I to V; specificity varied similarly. With SCORE, predicted/observed CHD mortality fell from 1.53 to 1.26 from social class I to V; sensitivity at a ≥5% threshold (29% of men) fell between non-manual (61%) and manual (57%) groups, as did specificity. However, including social class in FRS barely improved risk prediction (net reclassification improvement = 0.18%). Framingham and SCORE predictions varied between socioeconomic groups and are more likely to identify those at greater CHD risk in higher socioeconomic groups. To ensure equitable primary prevention, strategies to adequately estimate risk in lower socioeconomic groups (at increased CHD risk) should be developed.