How to improve quality of diagnosis
Little research exists into issues around GPs and acute clinicians getting diagnoses wrong. Ingrid Torjesen asks what the NHS is doing about this crucial quality issue
- Misdiagnosis is the big fear for three in five patients seeing a GP.
- Computerised diagnostic aids are now available for doctors.
- Specialist advice may have to be called in earlier.
The rash that was in fact meningitis; the painful limb that turned out to be fractured; the cancer not picked up until it was too advanced to treat; or the aching shoulder that was actually a sign of Parkinson’s disease. All are serious examples of misdiagnosis.
Most people will have been misdiagnosed at some stage in their lives. The consequences may not have been fatal or caused permanent harm, but it would most certainly have delayed care and perhaps subjected that person to investigations or treatment they did not need. Not only will this have been unpleasant, it will have increased NHS costs and caused bottlenecks.
“Excessive testing might ensure serious conditions are unlikely to be missed, but it is not an indicator of good diagnostic skills”
The right diagnosis is vital because it determines the pathway a patient takes and ultimately the care they will receive. Diagnosis is a journey, rather than a single event, and begins when the patient is first assessed, usually in general practice. This initial decision will determine whether the patient needs further care and from whom.
A YouGov poll in 2005 found that one in five people had either been misdiagnosed in the previous five years or had a friend or relative who had been; three in five said being misdiagnosed was their biggest fear when seeing a GP.
These fears are not unfounded. While problems with diagnosis accounted for 9 per cent of patient complaints to the Healthcare Commission in 2007-08, they made up a quarter of complaints about general practice.
But, despite obvious concerns, little or no research appears to have been done on the quality of the GP consultation and diagnosis.
Not much more has taken place on diagnosis in hospitals. A review of 1,000 hospital admissions found that at least one adverse event was experienced by 8.7 per cent of patients and that 5 per cent of these were due to diagnostic error and preventable. Another review, of 840 emergency admissions, revealed five (0.6 per cent) had been diagnosed incorrectly - all by junior doctors.
Doctors make hundreds of diagnostic decisions every week, so it is not surprising that they get some wrong, particularly while training. But what is surprising is that there are no internationally agreed methods of assessing the quality of the consultation and diagnosis. It is usually only when something goes very badly wrong that the diagnostic journey is reviewed.
The National Patient Safety Agency has only looked at diagnosis in cancer. It reviewed the records of 141 patients who had died or been severely harmed, due to late diagnosis or misdiagnosis, to determine the contributing factors.
Rather than the timeliness of GP referrals, the focus was on the longest delays, which appeared to result from glitches in the reporting of test results, misinterpretation of tests, administrative errors and patients not being chased to be told the diagnosis if they did not attend.
Evidence suggests delays in GPs identifying and referring suspected cancer patients is one reason why the UK has poorer cancer outcomes than elsewhere in Europe and North America, so national cancer director Mike Richards has requested an audit of cancer diagnoses in primary care. This is being led by the Royal College of GPs.
Expensive diagnostic tests
Case note reviews occur more frequently in the US because it is more litigious and it has been estimated that 40,000-80,000 people die in US hospitals every year due to late or misdiagnosis. If those figures are extrapolated to the UK, purely on population, it would equate to somewhere between 8,000 and 16,000 - and that does not take account of the much readier access to expensive diagnostic tests in the US.
King’s Fund senior fellow Nick Goodwin says: “In the US a lot of the larger primary care units have MRI scanners and just because you have insurance they say we will just run you under. They do it because they have got the technology, because it is paid for and because it reduces their liability if something goes wrong, but it is tremendously more expensive by a factor of three or four.”
While excessive testing might ensure serious conditions are unlikely to be missed, it is certainly not an indicator of good diagnostic skills. Every investigation impacts on the patient’s holistic health; even with something as simple as a blood test there is a risk of infection.
Doctors in the US will often ask for “a full blood work-up”, says the British Medical Association’s GPs committee chair Laurence Buckman, whereas what are really needed are appropriate investigations.
“Any test that was normal because you did a battery of tests - that was an unnecessary test and a waste of resources and you should have known better,” says Dr Buckman.
While a specialist is automatically looking for something that is wrong, the generalist is focused on dealing with the 95 per cent of cases that can be treated with advice or a prescription or a non-urgent referral. Recognising conditions is becoming increasingly difficult for GPs because there is a greater variety to choose between, and many patients will have more than one.
Aneez Esmail, professor of general practice at Manchester University and an expert in diagnostic error in primary care, says circumstances also make it easier for hospital doctors to identify problems.
“You don’t that often get a problem misdiagnosed in hospital because you are seeing disease at a much further end of the spectrum in effect, where things are much clearer,” he says. Hospital doctors have many more tools available to them: blood tests and x-rays can be ordered, or a nurse asked to monitor the patient. A GP must make a judgement simply on the history and a short physical examination.
Only a tiny part of the NHS research budget goes into projects looking at diagnostic quality, but Professor Esmail points out that to be able to assess and improve the quality of diagnosis, it is important to understand how and why particular diagnoses are reached and the thinking processes involved.
“It is dependent on the individual clinician, their knowledge base and the context in which they are working,” he says.
It does not help that how to hone and improve diagnostic skills forms a very small part of the undergraduate medical curriculum. In Professor Esmail’s view even that is poorly taught.
Royal College of GPs chair Steve Field says GPs coming out of training are “competent”, but many are not “confident” because they have not had sufficient experience of the conditions they are coming across, so the college is pushing for longer training. For example, only trainees in Liverpool are currently guaranteed a placement in paediatrics, he says.
“We need to do more work looking at the quality of diagnosis and referrals, we need to know more about what we are doing in primary care because when you start to look at that, and look at the variability, you can start to improve not just the bottom end of the curve but the lot,” says Professor Field.
When a patients first presents, the most important thing is for the GP to pick up when a symptom such as an ache, pain or rash is not a minor problem but a manifestation of something more sinister.
Computerised diagnostic aids are now available to prompt doctors to consider conditions they may not have thought of. One of these is Isabel, with information on more than 10,000 conditions.
Isabel Healthcare chief executive Jason Maude says there is no requirement for doctors to note the conditions they have considered and discounted.
“There is a tendency for people to think about one or two things and then forget the third or fourth thing that could be extremely important,” he says.
He asks: “How can you talk about quality and not talk about diagnosis? It is like aviation safety looking at every aspect of aviation except take-off.”
Doctors frequently refer to reference books and the internet to check things out, says Professor Field, who admits he always carries the Oxford Handbook of General Practice. He adds: “Some GPs might find a computerised aid irritating and some GPs might find it helpful as part of their development.”
One GP who would find such an aid intensely irritating is Dr Buckman.
“I think it is very dangerous to have computerised fail-safe mechanisms,” he says. “You have to have doctors who are competent and know their limits. You should not be doing something if you need to be reminded what to do; if you need to be reminded you need to educate yourself so you don’t.
“I am not anti rule book, I just think the rule book has to be written in such a way that it allows for variation. Otherwise, you end up with a pathway you can’t get off.”
Mr Goodwin says the consultation is the key to a quality diagnosis. He believes that the doctor often focuses on the main symptom being presented rather than looking for other associated signs.
“If you were able to collect all of the range of the presenting conditions that someone comes to you with, which is quite difficult to do in a consultation of five or 10 minutes, you are more able to make a more accurate and reasoned diagnosis,” he says.
Longer GP consultations or referral to a long term conditions nurse to collect a fuller history could help. Patients should also be educated, through initiatives like health coaching schemes, in how to have a good consultation by presenting everything they feel is wrong with them. The threshold at which GPs get more specialist advice might also need to be reduced if diagnostic quality is to be improved, Mr Goodwin adds. This does not necessarily mean more hospital referrals, it could involve asking advice from or referring to a specialist doctor based in the practice or a telephone call.
Last month the King’s Fund announced plans to look at GP diagnosis as part of a wider inquiry into the quality of primary care. The aim is to develop indicators to assess and compare different practice and suggest ways in which it could be improved.
Of course, the ultimate audit of whether a diagnosis was correct is a post-mortem.
Royal College of Pathologists president Peter Furness says one study found major diagnostic discrepancies at post-mortem in one in five patients who died on an intensive care unit. These included undiagnosed bleeds, pulmonary embolisms, heart attacks and complications of infections which would have affected treatment.
“That is in patients on an intensive care unit,” he says. “What it is like out in the community, goodness only knows.”
Professor Furness adds that post-mortems are rarely carried out on patients who die at home because there is not a clear NHS mechanism for how the body should be transported to a mortuary.
Doctors used to request post-mortems more frequently when they were not sure that what they were putting on patients’ death certificates was correct. However, Professor Furness reveals that doctors mistakenly believe discrepancies occur less often these days because of improved and increased availability of diagnostic techniques.
“The direct evidence does not indicate that there has been a reduction in the number of discrepancies between pre- and post-mortem diagnoses, despite the improvement in imaging and diagnostic techniques,” he says.
A reluctance in society to interfere with bodies after death, bound up loosely with concepts of human rights and dignity, and heightened by the Alder Hey and Bristol organ retention scandals, has also contributed to the fall in post-mortems.
In 1999, before the scandals’ peak, post-mortems were carried out on 6 per cent of patients who died in hospital in Leicester, but in 2001 it was just 1 per cent. The same audit shows it was not relatives refusing consent - 51 per cent in 1999 agreed compared with 49 per cent in 2001 - but doctors not asking for it.
Not surprisingly, doctors who regularly have to consider asking for consent find it a difficult process. BMA consultants committee chair Jonathan Fielden says: “It is something you have to do very carefully. It is not a simple exercise to say we will just increase the number of post-mortems. Although it would give us more information, there is a knock-on effect for families.”
Vast amounts of data are already picked up through clinical governance systems, complaints and monitoring tools, Dr Fielden adds. “What we need to do,” he says, “is continually drive up quality by building systems to turn this vast amount of data into information that is then made available to more clinicians. The problem is that has not been, until very recently, a priority for the health service.”
Getting it right at Rotherham General
Rotherham General Hospital foundation trust is conducting an analysis of the quality of diagnosis in its accident and emergency department. The main aim is to reduce the number of delayed and wrong diagnoses, such as missed fractures and conditions where deterioration is not spotted promptly, flagged up by the serious untoward incident system.
Trisha Bain, deputy chief of quality and standards at Rotherham General Hospital, says that, as well as getting the right diagnosis quickly, it is also about reducing variations in diagnosis. The InterQual tool has identified that quite a high proportion of patients are not meeting the set admission criteria and some are staying in hospital much longer than expected.
“We were thinking is that is an issue around diagnosis as well,” she says.
The analysis involves an in-depth case note review of 50 patients readmitted within six months by a multidisciplinary team using the global trigger tool to ascertain whether any diagnostic errors contributed to the readmission. Diagnostic errors will also be looked for specifically in the 20 randomly selected case notes already reviewed for incidents every month and through existing mortality review processes.
The team will consider whether diagnosis was made promptly and correctly, whether the right investigations were carried out and on time, who made the initial diagnosis - a junior doctor or a consultant - and whether a consultant later changed the junior doctor’s diagnosis.
An aggregate root cause analysis will be conducted on cases where there appear to have been diagnostic issues to identify common themes.
Dr Bain says: “We hope that by looking at it in various ways and through different tools we should be able to pick up some general themes.
“We imagine it will be things like timeliness of tests, who is actually making the decisions, is it junior doctors, and does the diagnosis get changed when a consultant reviews the case.”