Publishing bowel surgery death rates will save lives
Posted Apr 11 2011 12:00am
Today a team of researchers, led by our stats expert Dr Eva Morris from Leeds University, have completed the first detailed analysis of bowel surgery outcomes across the whole of England.
The team looked at how the proportion of patients who die within 30 days of surgery – called ‘post-operative 30-day mortality’ – varies from hospital to hospital, and published their findings in the peer-reviewed journal Gut .
This is without doubt a sensitive issue. Comparing deaths after surgery between hospitals is bound to ruffle a few feathers – but it’s important to stress that these findings don’t suggest why these variations exist. And there’s plenty of reason to believe that this isn’t, as some might assume, simply a reflection of the competencies of individual surgeons.
But if the figures don’t tell us the reason for the variation, what’s the point of measuring and publishing them? How does ‘naming and shaming’ the poorest performing hospitals improve things, when it also runs the risk of negative media coverage , or of upsetting patients who are due to have similar surgery?
Thankfully, as we’ll see, there are plenty of reasons why publishing data and making comparisons can, and do, improve the way health systems work, and ultimately improve things for patients.
The first and most important thing to say about these figures is that, overall, things have got substantially better since 1998, when 6.8 per cent of bowel cancer patients died within 30 days of major surgery – that’s a just under seven people for every hundred having an operation.
This had fallen to just under six per hundred by 2006. And there’s even a suggestion, looking at the most recent stats on the NCIN website, that things have got better still (although these data have not yet been formally peer-reviewed).
But as we pointed out in our press release , the data show worrying variations between hospitals across England – variations that don’t seem to be down to differences in the number or type of patients the hospital treats.
In fact, at a small number of hospitals, the proportion of patients dying within 30 days of surgery was much greater than average. Yet other hospitals seem to be performing significantly better than the rest.
Our chief clinician, Professor Peter Johnson , thinks it’s premature to assign a cause to the figures, and that 30-day post-operative mortality, as a measurement, is “a question, not an answer”:
“What data like these tell you is that you need to go and look very carefully at the areas where there are outliers [hospitals that are unexpectedly above or below the average] and see if there’s a genuine explanation – a rational explanation – for why they are different to the rest, or if there are genuine difficulties. It may be factors other than individual surgeons – it could be the way the teams work, or factors such as late diagnosis, ” he told us.
For him, the most likely explanation is that the variation is influenced by a combination of many factors, notably the way hospitals’ bowel cancer services are structured – and since 2006 there have been big improvements, especially through the introduction of ‘multidisciplinary teams’ of different specialists, and increased use of newer surgical techniques such as keyhole surgery .
But why publish these data if we can’t draw firm conclusions? How does knowing how a hospital’s vital statistics affect the way care is delivered? Given that we know that things have likely improved since 2006, how does these ‘old data’ help things?
Firstly, as Professor Johnson pointed out, “if you don’t put data out there for comment and criticism, you’ll never get better data” – a point echoed by Dr Mick Peake , a cancer clinician based at Leicester and clinical lead of the National Cancer Intelligence Network (NCIN).
Dr Peake pointed out that as a result of similar data published over the past few years, as part of the National Lung Cancer Audit , the simple fact of ‘knowing you’re being observed’ has led doctors to record more and better data, and in some cases, change how they treat their patients. According to Dr Peake, this has had a dramatic effect on lung cancer surgery rates in this country.
He highlighted figures showing that the proportion of lung cancer patients who receive surgery has increased year on year since the relevant data began to be published in 2006, and this is not just because more data are being reported. Intially, just 11 per cent of patients were offered surgery – less than half the international average.
But figures soon to be published as part of the 2011 Lung Cancer Audit will show, according to Dr Peake, that we’re now on the verge of catching up with the rest of the world and that this will ultimately, he thinks, be reflected in improved survival rates for patients.
And there is “no doubt” in his mind that the publication of the figures caused these improvements.
Another area where better data collection has led to lives being saved is in heart surgery. In 2001, data began to be published on how patients fared, amid concerns that this would make surgeons more defensive and less likely to perform high-risk operations. But an analysis in 2009 showed that the opposite had occurred – mortality rates fell as the surgeons shared ‘best practice’ and actually took on more difficult operations.
A final, and more indirect example of data saving lives is to be found in the EUROCARE international cancer survival comparisons. These studies began to be published in 1995, and looked at survival rates between different countries across Europe. Successive reports led to a huge amount of media interest, not least because the UK appeared to be lagging behind other countries and was branded the ‘sick man of Europe’ by some newspapers.
But despite these criticisms EUROCARE, if anything, lit a touchpaper under cancer policy in the UK, and focused policymakers’ minds on how to improve things. As a direct result, the UK government published a series of ‘ cancer strategies ’, which have led to improvements in cancer care across the UK.
Completing the virtuous circle, the data published today were made possible as a direct result of policies brought in by one of these strategies – namely the creation of the National Cancer Intelligence Network, who were involved in Dr Morris’s analysis. Data, it seems, begets improvements – which beget more data, and further improvements.
Professor Mike Richards , the UK’s National Cancer Director, pointed out at our press briefing that bowel surgery in this country should be judged, not by these figures, but by what happens next – this is very much the “start of the process”, and he’s “extremely optimistic” about the future.
Identifying the best performing hospitals has already led to plans for visits from other bowel cancer surgeons, to learn from them. There are also plans to add in more data in future, such as the causes of post-operative deaths, and how teams are structured in different hospitals, which will help pin down the source of any problems.
And it’s likely that the roll-out of new techniques and procedures for treating bowel cancer – such as multi-disciplinary teams, keyhole surgery, and Enhanced Recovery Plans – will be further accelerated, benefitting patients around the country.
And it doesn’t stop with bowel cancer. There are now plans to use the techniques developed and refined by Dr Morris and her team to look at other cancer types – which should identify more areas where improvements can be made.
Slowly, but surely, the collection and publication of data like these are making things better for cancer patients. Beating cancer isn’t just about developing expensive cancer drugs , studying genes and DNA , or campaigning for tobacco control measures . It’s also about taking time to understand the complexities of a modern health service and using this knowledge – ultimately – to improve things for people with cancer.
Morris EJA, Taylor EF, Thomas JD, Quirke P, Finan PJ, Coleman MP, Rachet B, & Forman D (2011). Thirty-day postoperative mortality after colorectal cancer surgery in England Gut : 10.1136/Gut.2010.232181