Urgent referrals for suspected cancer vary threefold among general practices

The rate at which GPs refer patients for suspected cancer varies so widely that it indicates real differences in the standards of care being provided.

Some variation in referral rate would be expected because populations differ, but data published this week for individual practices by the National Cancer Intelligence Network shows that some GPs refer more than three times as many patients as others.1 Rates range from a low of under 830 in every 100 000 people a year to a high of over 2550. This is the first time such data have been made publicly available, enabling patients to compare the referral rate of every general practice in England.

Mick Peake, the network’s clinical lead, said, “The data are not easy to interpret, since we do not know what the optimum level is for these measures; and although the data are adjusted for age, there may be other differences in the characteristics of the patients of a particular GP practice that impact on local referral rates. However, the range of the variation is so wide that, at the extremes, it probably reflects differing standards of care.”

Past studies of GPs’ referral and prescribing behaviour have shown similarly wide variation. A King’s Fund study in 2010, for example, reported that among all referrals (not just those for cancer) there was a 10-fold difference between the general practices that referred the most and those that referred the least.2 So the newly published data do not indicate anything unique about cancer referrals. Each year the average GP will see only seven patients who are found to have cancer. These break down as one case each of breast, bowel, prostate, and lung cancer and three cancers of other types.

The data were gathered from the national cancer data repository, and NHS patient numbers were used to track patients back to the general practice they were attending at the time of diagnosis. The number of diagnoses counted was more than 90% of the total number for the country recorded by the Office for National Statistics; some were missed because patients had moved and changed their practice more than once and could not be traced. The data were available to NHS staff and GPs on a confidential basis but have now been made public as part of the government’s open data policy. Where small numbers of patients in a particular practice may make it possible to identify individuals, the data have been suppressed, but professionals can still access the complete dataset.

Di Riley, associate director of the National Cancer Intelligence Network’s clinical outcomes programme, said, “Although the number of people GPs refer isn’t on its own an indicator of how good they are at spotting the early signs of cancer, it’s clear from these data that there’s variation that needs to be addressed. It’s important to remember that GPs have a hard job, and many of the symptoms of cancer are very similar to many other illnesses. But we must do more to understand the reasons for the variation.”

Sarah Woolnough, executive director of policy and information at Cancer Research UK, said, “It’s very worrying to see a more than threefold variation in the rate of urgent cancer referrals among GPs. And although the number of people GPs suspect have cancer and send for further tests will naturally vary depending on the age and demographic of their patients, this level of variation suggests differing approaches by some GPs.

“We urgently need to learn more about what’s behind these differences and tackle any poor practice. We already know that some patients present several times with cancer symptoms before being referred for further investigation, which can lead to late diagnosis. And a delayed cancer diagnosis could prove critical for a patient’s chance of survival.”