This site is intended for health professionals only

At the heart of general practice since 1960

Framingham must be consigned to history

There is no longer any place in guidance for a risk score that discriminates against deprived populations, argues Dr Peter Brindle

There is no longer any place in guidance for a risk score that discriminates against deprived populations, argues Dr Peter Brindle

Whether you like it or not, the primary prevention of cardiovascular disease occupies more of our practice time than any other single condition, and this is set to increase considerably. National guidance now recommends that statins and blood pressure treatment be considered for everybody with a 10-year risk of CVD of greater than 20%.

Depending on how it is calculated, this could mean an additional three to five million people being eligible for preventive treatment. The accuracy of CVD risk assessment is important to individuals because if a risk score misclassifies someone as being over the 20% threshold when their true risk is less, then it is likely they will end up taking lifelong preventive treatments when perhaps the benefits of treatment are outweighed by the risks.

Conversely, if someone's true risk is underestimated so they do not reach the threshold, then they will not benefit from treatment.

Framingham study

For the last 15 years or so, cardiovascular disease risk has been calculated using various formulae from the Framingham Heart Study. Framingham is a small town in the north-east of the US and most of the baseline data for these formulae were collected from 5,573 people between 1968 and 1975 when Framingham had a largely white and middle-class population.

Over the last few years, serious questions have been asked about the suitability of using such a risk score in modern European populations. A recent review assessed the accuracy of the Framingham score in 27 different populations and found it ranged from overestimating true risk by 2.9 times, to underestimating by 57%1.

Most worrying of all is its inability to detect the well-known increased risk associated with social deprivation. The Framingham score systematically under-estimates CVD risk in people from deprived areas compared with people with identical risk factors but from more affluent areas2,3. It is not hard to imagine how using such a score could exacerbate health inequalities.

Two new scores

So why has the Framingham score been published as charts at the back of the BNF and integrated into our GP software for all these years? The answer is simply that despite its flaws, there has been no better way of prioritising people for primary prevention.

Fortunately two new scores have been published that include a measure of social deprivation as a risk factor via a postcode-linked deprivation index. The first of these was ASSIGN, derived from about 13,300 Scots recruited between 1984 and 19924.

The second was the publication of QRISK in July 20075. QRISK, through collaboration with EMIS, is the first risk score to use real primary care data as the source of risk factors and CVD outcomes rather than the artificially controlled measures of traditional cohort studies like Framingham.

Like ASSIGN, it includes having a family history of CHD as a variable as well as social deprivation, but it also includes being on blood pressure treatment and body mass index as additional variables.

The result is a risk score derived from 1.3m people that performs better on every measure than Framingham in 600,000 people from other EMIS practices and a further 1.1m people from practices using In Practice Software6.

Although QRISK and Framingham identify many of the same people as being at high risk, nationally the numbers of differently classified people are equivalent to a city the size of Bristol.

We have reached the point that guidelines should no longer recommend the Framingham risk score as it could worsen health inequalities, and because QRISK, derived from British primary care data, is both fairer and more accurate.

It is the time to consign the Framingham score to history.

Dr Peter Brindle is a GP in Bristol and R&D lead for Bristol, North Somerset and South Gloucestershire PCTs – he is an adviser to NICE on cardiovascular risk assessment, but this opinion piece represents his personal view

Competing interests Dr Brindle is a co-author of ASSIGN and QRISK


Dr Peter Brindle: QRISK performs better on every measure than Framingham Dr Peter Brindle: QRISK performs better on every measure than Framingham

Rate this article 

Click to rate

  • 1 star out of 5
  • 2 stars out of 5
  • 3 stars out of 5
  • 4 stars out of 5
  • 5 stars out of 5

0 out of 5 stars

Have your say