This site is intended for health professionals only

At the heart of general practice since 1960

Read the latest issue online

A faulty production line

Cancer risk tool rolled out to over 4,000 GP practices

Over 4,000 practices using EMIS Web can now access a cancer ‘symptom checker’ which helps GPs to identify patients at risk of having some form of cancer who need to undergo investigations.

The online QCancer tool – developed by the team of researchers behind QRISK– has been integrated into EMIS Web, so that GPs can see if a patient’s signs and symptoms indicate their risk of an existing, as yet undiagnosed, cancer is above a certain level.

The tool was developed using the QResearch database including over 900 UK GP practices, and has now been evaluated using the THIN database, which researchers said showed it have very good levels of accuracy at identifying patients with as yet undiagnosed cancer.   

The threshold for ‘high risk’ is currently set at a 5% chance of having any cancer, although this level could be changed, if for example NICE goes ahead with plans to lower the typical risk threshold for referral for suspected cancer to a positive predictive value of 3% for individual tumour sites.

The symptom checker is being developed into a full Cancer Decision Support (CDS) tool with help from Macmillan Cancer Support, which will be rolled out next month.

The CDS includes two extra features, a ‘batch processor’ that allows practices to score their entire patient population and review case notes of those identified at high risk, and an ‘alert’ that will flag up high-risk patients as the GP adds information to patient records.

Lead researcher Professor Julia Hippisley-Cox, from the University of Nottingham, told Pulse the symptom checker will help GPs carry out earlier, more targeted investigations of symptomatic patients.

Professor Hippisley-Cox said: ‘We are grateful for all the EMIS GPs who contribute to QResearch enabling [the] creating of tools such as these and for EMIS for integrating it into EMIS Web for the benefit of patients.

‘Early diagnosis of cancer remains a significant clinical challenge and we hope that this tool assist GPs in the consultation to identify those most at risk.’

Related images

  • IT - computer - computer screen - monitor - online records - online

Readers' comments (2)

  • The national rate of conversion of 2WW referrals to cancer is approx 10.5% i.e. 1 in 10, two week referrals is truly a cancer. Nationally approx 45% of all cancers are picked up by 2ww ( the actual number picked up quickly is probably slightly higher as if some is suspected to have e.g leukemia we don`t wait for 2 weeks but admit them ASAP -appropriate treatment but will not count towards 2ww referrals)
    To increase the pick up rate one can understand the 5% risk rule but the no of false positive could be high and false negatives if the symptoms are not read coded adequately.
    I am quite happy to run a automatic batch processor which we did for Q diabetes and Qrisk based on existing information to target screening but this is a different ball game.
    Calling up someone saying we would like to do a health check and then checking fasting sugar or lipids or HbA1c is different from calling someone for suspected cancer!!!
    Given that most symptoms listed are usually free texted and not read coded it may falsely reassure as well if symptoms are not read coded.
    Also I suspect most Acute Trusts would collapse completely if we use the 5% risk criteria -so although a noble and laudable effort there may not enough capacity in the system to absorb that.
    I would like to see if they have done any population based studies to see how much extra referrals this would generate.

    As usual caveat emptor!

    Some consultants believe the risk should be set at 30% ( the local trust 2ww letter states only 30% of those referred via 2ww have cancer-and they already can`t cope and have written for us to scrutinize our referrals well before sending!
    Wonder how they will cope at 5 % risk threshold or lower!

    Unsuitable or offensive? Report this comment

  • The challenge is in setting the threshold. The higher you set it the more cancers will be "missed", the lower you set it the more "normal" people will be seen for every cancer diagnosed. These tools give us the numbers. The voters (or the politicians) must decide where they want the balance. The trouble is they either lack the understanding or the honesty for an open dialogue.
    I do recommend viewing the video on how the checker template works though, it gives the opportunity to code the symptoms as you go which is very helpful in guiding routes of referral.

    Unsuitable or offensive? Report this comment

Have your say