Feature Medicine and the Media

How many new cancers are diagnosed after emergency admission?

BMJ 2013; 347 doi: https://doi.org/10.1136/bmj.f6857 (Published 20 November 2013) Cite this as: BMJ 2013;347:f6857
  1. Margaret McCartney, general practitioner, Glasgow
  1. margaret{at}margaretmccartney.com

And what conclusions about general practitioners’ performance can be drawn from these statistics? It’s more complicated than the newspapers might have us believe, finds Margaret McCartney

On 18 October, NHS England London tweeted, “Did you know? A third of cancer diagnoses are still only made in A+E [accident and emergency].” A few days earlier, the Teenage Cancer Trust had released a report that concluded that “over a third of young people with cancer” are diagnosed through admission to emergency departments.1

Last year, there was widespread media coverage when the National Cancer Intelligence Network, part of Public Health England, reported findings2 that led to headlines such as the BBC’s “‘Too many’ cancers only diagnosed in A&E, study suggests.”3 The article reported Jane Maher, chief medical officer at the charity Macmillan Cancer Support, as saying, “It is appalling that so many cancer patients are still diagnosed through emergency admissions. It can be more difficult to spot cancer symptoms in older people who have other health conditions but this does not excuse such a high number of people being diagnosed in this way.”3

The NHS England tweet was incorrect and was later clarified. It used the same fraction of “a third” as data publicised by the Teenage Cancer Trust, which it had taken from its own annual survey. This found that of 300 young patients with cancer, 111 (37%) had had their condition diagnosed when visiting an emergency department. However, of these, only 29 (9.6%) had not been referred by their general practitioner.4

Research from the NHS North West Cancer Intelligence Service, published earlier this year, used hospital episode statistics to find that 24% of diagnoses of cancer in children and young adults were made at emergency presentations.5 The proportion was the same for the population overall. Of these, a third (8%) are in patients who have been admitted by their GP; the remainder are in patients who have attended emergency departments—but some will also have been sent by their GP.6

Additionally, a paper using data from the general practice research database integrated with hospital episode statistics found the proportion to be lower still, at 13.9% of cancers being first diagnosed at emergency admission.7

This is confusing because the Teenage Cancer Trust’s press release seemed to implicate GPs for not acting sooner, and Cancer Research UK told the BBC that some people are diagnosed in emergency departments because they “could be slipping through the net as symptoms may be dismissed as ‘the usual aches and pains’ or ‘old age.’”3 However, GPs may have been trying to access investigations as quickly as possible for the patient by referral to the emergency department.

Maher told the BMJ, “The key issue here is that no one group of professionals is to blame for the undeniable fact that many cancer patients are currently only being diagnosed [as having cancer] when they reach an emergency admission stage. Witch hunting is absolutely not the answer.”

Instead she thinks that “every opportunity” should be used to “help raise awareness of the signs and symptoms of cancer so that people know what to be looking for.”

Greg Rubin, a professor at the University of Durham who researches the diagnosis of cancer in primary care, told the BMJ, “On the whole, emergency presentations are associated with worse outcomes—there is evidence for that in colorectal and lung cancers. But the category of emergency admissions comprises both admission to hospital by the GP and the patient going to the emergency department—and that patient may have gone there because the GP told them to do so.”

So how many new diagnoses of cancer in emergency departments are “too many,” and how many are appropriate? “Emergency presentation is seen as a bad thing, but we need to be more sophisticated in understanding why people present in an emergency and not assume that the GP got it all wrong,” said Rubin.

Although symptoms of some cancers, such as breast lumps and pigmented lesions, are relatively straightforward to identify, others are far more difficult. GPs sometimes get the diagnosis wrong or make it later than they might have.

“But we don’t understand what is happening. Some people come to see their doctor for the first time with a symptom and are sent [to hospital] as an emergency. Some people will be having tests done in primary care, and in the process, be sent into hospital as an emergency. And in some cases, there will have been a missed initial diagnosis and the patient will need an emergency admission,” said Rubin.

As it stands, the data are relatively crude and require further understanding rather than the blame mongering of the Daily Express (“GPs fail to spot cancer in 1 in 3 elderly”)8 or the Daily Mail (“GPs are still failing to spot cancer”).9

Many research reports make an additional inference that symptoms later found to be related to cancer and which have taken more than three GP consultations before referral are undesirable and indicative of less good care.10 Rubin said, “Overall, we have found that GPs are very good at detecting cancer; 75% of patients with cancers are referred after the first or second consultation.” However, cancers with vaguer symptoms are less easy to diagnose quickly. Many GPs have been taught the “three strikes and out” rule—that is, if unexplained symptoms persist after the third appointment, referral should be considered.

Rubin agrees with this approach, particularly for children and young people, but also thinks that diagnoses of cancer after the third consultation shouldn’t necessarily be considered poor care. He gave the example of the National Institute for Health and Care Excellence, which says that rectal bleeding of six weeks’ duration should be investigated, meaning that a patient with the symptom for a shorter duration might be asked to return to primary care before referral.

“A GP could be clocking up investigations, for example, for suspected ovarian cancer, and appointments, and doing everything right. Sometimes you need time to evaluate the development of symptoms.” GPs are skilled at dealing with uncertainty and reviewing symptoms. However, even “red flag” indicators for cancer are relatively unspecific,11 meaning that their use for referral management is flawed. Indeed, an audit of urgent cancer referrals in Scotland found cancer in a large proportion of patients referred outside of the guidelines.12 Enforcing strict guidelines for referrals, therefore, is likely to be counterproductive.

It is easy to run media campaigns for awareness and “earlier diagnosis.” It’s far harder to work out what is going on, where the delays are, and how to support GPs who are balancing one set of uncertainties and potential for harm against another. The NHS has recognised that openness is essential. But with complex statistics and a failure to properly explain them to the public, we risk turning an opportunity to understand and improve into a blame game.

Notes

Cite this as: BMJ 2013;347:f6857

Footnotes

  • Competing interests: I have read and understood the BMJ Group policy on declaration of interests and declare I am an NHS GP.

  • Provenance and peer review: Commissioned; not externally peer reviewed.

References

View Abstract

Sign in

Log in through your institution

Free trial

Register for a free trial to thebmj.com to receive unlimited access to all content on thebmj.com for 14 days.
Sign up for a free trial

Subscribe