Too little has been done to discover the reasons for perioperative deaths, says a report published this week. Lynn Eaton looks at how that is set to change

It would be nice to think that any patient who ends up in hospital eventually gets better and goes home. Unfortunately, many die there.

Why they die - particularly in the 30 days immediately after surgery - is the subject of the annual report from CEPOD, the National Confidential Enquiry into Perioperative Deaths, published this week. It comes as the debate on improving clinical standards moves up a gear, with proposals for the National Institute for Clinical Excellence (NICE) and the opening of the public inquiry into the deaths of children who had heart surgery at Bristol Royal Infirmary.

In this climate of closer monitoring and regulation, the CEPOD report, though giving some insight into both NHS and independent hospital practice, is also frustrating. It does not identify individual consultants' practice, nor even the hospitals or health authorities involved. If there is another Bristol going on out there, you certainly won't find out about it here.

'It wasn't set up to be a clinical management tool,' explains Stuart Ingram, consultant anaesthetist at University College London Hospitals trust and principal clinical co-ordinator for anaesthesia with CEPOD. 'It is a learning process.

'Our enquiry works on a voluntary basis. The confidentiality protects hospitals and individual clinicians. When the enquiry was first set up in 1989, it was the only way to get information. We don't know the doctors involved. People know they can write honestly and it is not going to rebound. '

A total of 19,496 deaths of patients within 30 days of surgery were reported to CEPOD during 1996-7. Of these, 2,541 deaths, 13 per cent of the total, were looked at in detail. They covered patients who had undergone gynaecological, head and neck, minimally invasive, oesophageal, spinal or urological surgery. The report devotes a section to each, looking at the specific reasons for deaths.

In gynaecology, for example, 205 deaths were reported within 30 days of operation. Infection was a major factor in seven of them. 'It is a small number of cases, but there are lessons to be learned,' says Dr Ingram. Some of the women were high risk. But one, a 76-year-old woman who was very low risk undergoing routine surgery, died four days later. Dr Ingram highlights the fact that no prophylactic antibiotics were given.

Percutaneous endoscopic gastrostomy - an operation to introduce feeding tubes into the stomach - is deemed inappropriate in patients with very short life expectancy. 'Do we strive too hard to keep patients alive when the prognosis is hopeless and death is imminent?' the report asks.

It goes even further, to question whether many operations should be carried out at all: 'It is difficult to resist pressure to operate, whether this comes from the patient, relatives, or medical colleagues but it must be recognised that surgery cannot solve every problem,' says Ron Hoile, principal clinical co-ordinator for surgery.

One of the clincians' biggest worries is that surgeons and anaesthetists are not able to keep abreast of technological advances and learn new skills. 'We are good at training trainees, but it is more difficult with established consultants,' says Dr Ingram.

He argues that hospitals need to ensure they have enough people trained to use new equipment, such as key hole surgery instruments. 'Managers need to listen to clinicians in terms of what is becoming part of accepted practice. They should be looking to see what is going on in technology - not just the equipment, but the training implications.'

Death within 30 days of an operation will be one of the criteria NICE will use to measure clinical standards, which should raise the participation rate. At present, there is an 80 per cent response rate, but some key hospitals - including the Charing Cross in London, Royal United in Bath and Selly Oak in Birmingham - have no local reporter filing returns.

However, the 30-day statistic can be meaningless where the number of that particular type of operation is low.

'In Bristol, there were clearly large numbers of children, so you could identify poor outcomes,' says Dr Ingram. 'But what happens where a surgeon is doing only one or two of these cases? You would have to wait 20 years.

'Managers should be looking at the operations they are doing and if they are low numbers, asking whether referral to a larger centre would be more effective, rather than trying to be a jack of all trades.'

The report also highlights the low level of post mortems carried out where people have died within 30 days of an operation. Hospital post mortems were only eight per cent of the total cases.

'This is a recurring issue,' Dr Ingram admits. 'The general view among surgeons is that postmortems are an opportunity to learn. The assumption is made that we know exactly what caused the death, but when a post mortem is done, they often reveal things that would not be expected. There is a reluctance on the part of clinicians to request a relative's approval for a post mortem.

'We feel it is something hospitals should be looking at - what is their post mortem rate? That is the sort of indicator any chief executive can use.'

And it is not just about having a post mortem, but making sure the surgeon is there at the time, he adds, so they can learn what went wrong.

He also suggests managers review operations carried out less than 10 times a year by the hospital. At that frequency it may be difficult for the surgeon to maintain their skill base - and any errors will take years to come to light.

Discussions about how the CEPOD report will fit into any future clinical management tool are still under way. But for the time being, Dr Ingram wants to distance it from being a potential tool for NICE.

'What people should recognise is this is a process clinicians are undertaking and learning from. It is not there to be a national picture of what is actually happening. Managers might like it to be there in black and white, but what we are dealing with is shades of grey.'

+ 2 lines