Last month, NICE decided a modified version of the Framingham cardiovascular risk tool was the best option for detecting lipid modification. It had delayed publication of its guidelines to review an alternative tool, called QRISK.

QRISK was developed from the UK QRESEARCH database of general practice. This has now been updated and tests comparing QRISK2 and the NICE-recommended tool show the latter will miss people from ethnic minorities.

QRISK2 adds patient-reported ethnicity as well as conditions including type-2 diabetes, treated hypertension, rheumatoid arthritis, renal disease and atrial fibrillation to the original QRISK criteria. To test it, both QRISK2 and the modified Framingham tool were applied to data on 2.3 million patients aged 35-74 to find people at 20 per cent or higher risk of having a cardiovascular event in 10 years.

QRISK2 placed 10.4 per cent of patients in this category, compared with 14.9 per cent using Framingham. Closer examination showed that the extra patients picked up by Framingham were actually at low risk.

Lead researcher Professor Julia Hippisley-Cox, from the University of Nottingham, said: 'QRISK2 was marginally better than QRISK at the population level, and both were better than Framingham.'

However, QRISK2 excels on the individual level too, she added. While QRISK2 would identify 14.2 per cent of Bangladeshi women at risk, Framingham would pick up just 7.2 per cent. For Indian women, 10.1 per cent would be considered at high risk on QRISK2, compared with 4.6 per cent on Framingham. Looking at white men, QRISK estimates 14 per cent are at high risk, compared with 22 per cent on Framingham.

Dr Peter Brindle, a Bristol GP who helped develop QRISK, said: 'If this country wants a socially equitable risk score then sooner or later someone has to take notice.'

A NICE spokesman said it would consider 'any new evidence that comes to light'.

- BMJ Online 2008

Comment below and tell us what you think