Objective: To compare the accuracy of five risk assessment methods in identifying patients with uncomplicated mild hypertension at high coronary heart disease (CHD) and cardiovascular disease (CVD) risk.DESIGN Comparison of risk estimates using each risk assessment method with CHD risk 15% and CVD risk 20% over 10 years calculated using the Framingham risk functions.
Setting: British population.
Subjects: People aged 35-64 years with uncomplicated mild systolic hypertension (systolic blood pressure (SBP) 140-159 mmHg, = 202) from the 1995 Scottish Health Survey.
Main outcome measures: Sensitivity, specificity, positive and negative predictive values.
Results: Compared with CHD risk 15% over 10 years, the Sheffield table and Joint British Societies (JBS) Chart had good sensitivity and specificity ( 90%). The New Zealand (NZ) Chart had sensitivity 83% and specificity 89%. Compared with CVD risk 20% over 10 years the Sheffield table had sensitivity 81%, the JBS Chart had sensitivity 63%, and the NZ Chart had sensitivity 75%. All had good specificity ( 90%). For CHD risk and CVD risk the World Health Organization/International Society of Hypertension (WHO-ISH) and United States Joint National Committee VI (JNC-VI) methods had high sensitivity at the cost of very poor specificity ( 50%).
Conclusion: In patients with uncomplicated mild hypertension, the Sheffield table and JBS Chart both identified CHD risk 15% over 10 years with acceptable accuracy, while the NZ Chart was less accurate. Compared with CVD risk 20% over 10 years, these three risk assessment methods were all less accurate, but the Sheffield table retained the highest sensitivity ( 0.05 versus JBS Chart, = NS versus NZ Chart). The WHO-ISH and JNC-VI methods had unacceptably low specificities compared with both measures of risk and failed to differentiate between those at high and low risk.