The Naked Surgeon, page 8
I never imagined that while working on this chapter about the medical profession’s inability to learn from a near miss, a catastrophic event in my own practice would poignantly and viciously throw the issue into unforgivingly sharp focus.
One afternoon, I was operating on a 73-year-old woman with a narrowed aortic valve. The plan was to replace the valve with one made from animal tissue, with a view to relieving her breathlessness and reducing her chances of heart failure. The patient had a few risk factors, but nothing prohibitive, and we expected the operation to be smooth, quick, and relatively easy.
We attached the patient to the heart–lung machine in the usual way, and I inserted a tube into the left ventricle to keep the field free of blood while I replaced the valve. This tube was supposed to suck any blood from the ventricle, but, unfortunately on this occasion, it did not suck: it blew. The heart was filled with air, which went to the brain, and the patient suffered catastrophic and irreversible brain damage from which she died. Someone in the vicinity of the heart–lung machine must have unwittingly pushed the button that reversed the direction of the pump, so that instead of sucking blood away from the heart, it pumped air into the heart. Needless to say, this was an avoidable catastrophe, and a direct hit. The entire hospital immediately swung into action to study the root cause of the problem and see what could be learned from it. What we found was that it was ridiculously easy to reverse the pump accidentally: all it took was for an object, a finger, or an elbow to touch one sensitive button on the machine, and that made it blow instead of suck. On a human level, this was an unmitigated disaster for the patient and her loving family. On a professional level, it was a direct hit from which lessons could be learned, but I could not stop thinking how easily the accident had happened, and being somewhat surprised that it hadn’t happened before.
In 20 years of working at the same hospital, I had not seen such a calamity until now, and this begged the question: if this event was the direct hit, was it preceded by any near misses? I asked the perfusionists and surgeons if they had ever witnessed such an accidental pump reversal before. To my horror, most said ‘Yes’. All of them had seen it or, at the very least, were aware of it happening to colleagues, but, by sheer luck, on these past occasions the patients escaped injury: the archetypal type 2 near miss.
Within a few days, all staff at Papworth had been warned of the danger, and shown how the mistake could happen. Plans have already been implemented to change the operating procedures for this equipment pending either a modification of the design or the wholesale replacement of the hospital’s stock. We have learned from the hit, but it would have been better to have learned from the near misses.
Interestingly, when the first batch of the new machines was delivered, the new model had a modification that looked as though it was specifically designed to prevent this very mishap. The controls were embedded in a steeply sloping surface, and any object placed there would slip off onto the floor. Did the manufacturers know something that we did not?
On the bright side, however, research into direct hits is beginning in medicine. In heart surgery as in many other surgical specialties, the boundaries of what can be successfully done, and to whom it can be done, are being constantly pushed back. Nowadays, patients are a lot older and a lot sicker than they were 20 years ago, and yet the overall results are better and the success rates are higher. Not surprisingly, the medical literature teems with reports of successful intervention in supposedly ‘hopeless’ cases.*
[* Surgeons like nothing more than to publish papers with the unwritten subtitle ‘Look how clever I am’.]
In the FIASCO study, (Freed 2009), my colleagues and I at Papworth took a diametrically opposite approach. Instead of looking at survivors of very high-risk surgery, we looked at those who died from very low-risk surgery. The reason for this was simple: we felt that the best way to identify any weaknesses in our care system was to take patients where nothing should have gone wrong and yet it did. We thought that, by analysing these cases, we could divide them into two groups. The first group would consist of those where death happened owing to ‘a bolt from the blue’ or ‘an act of God’: in other words, where death was a bit of bad luck that nobody could have foreseen or avoided. The second group would be those deaths that could have been foreseen and possibly prevented. It is in this second group that we believed we would find what went wrong that was avoidable, and, if a pattern emerged, we would know what needed to be fixed. The results were a real eye-opener.
We studied only those patients whose predicted mortality was less than 2 per cent, so they were low-risk patients who nobody expected to die. There were 4,294 such patients operated on between 1996 and 2005, and only 16 of them died, which gives a mortality of less than half a per cent, so that, on the face of it, we were, as a hospital, doing very well. We then meticulously reviewed the case notes of the 16 patients who died, and decided whether the death was a ‘freak’, and therefore unavoidable, or whether it happened because of a FIASCO (‘failure in achieving a satisfactory cardiac outcome’), and was therefore avoidable. Having thus classified the deaths, we sent the case notes to an independent outside expert to confirm that we had got the classification right. Both we and the independent outsider agreed that nine of the 16 deaths could not have been avoided, but that there were seven avoidable FIASCOs. A pattern emerged in which the two commonest reasons for death in this group were communication errors and an inadequate method of protecting the heart during the operation. Both have since been addressed at the hospital by a change of practice. Four years later, we carried out a follow-up study (FIASCO II) to see if these errors had been eliminated. The study showed that not a single patient in this period had died as result of communication errors or poor heart protection (Farid 2013). The lessons had been learned.
No heart operation is without risk. Mortality, though fortunately rare, can still occur, even in low-risk patients. Nevertheless, we were astounded to find that, even despite an extremely low mortality in the low-risk group, FIASCO still accounts for nearly a half of deaths. This suggests, even proves, that mortality may be reduced even further as part of a quality-improvement programme. All hospitals are different, and systemic weaknesses and strengths will not be the same in all of them. We therefore recommended that all hospitals do a FIASCO-type analysis on their own patients to see what can be fixed, if anything, in their systems. The first FIASCO paper was published some five years ago. At the time of writing, the grand total number of hospitals worldwide who took our recommendations and did their own analyses is, to the best of my knowledge, five: the Karolinska Hospital in Stockholm, Sweden, and a group of four hospitals in Turkey.
There is therefore still some ostrich-like behaviour among doctors and surgeons regarding the concepts of quality management, monitoring, and learning from mistakes. The overwhelming majority of doctors, surgeons included, and other healthcare workers are passionate about doing the best for their patients. The only problem is that there is an attitude of risk-taking, and a resistance to change. I alluded to this earlier as eminence-based medicine, which I defined as continuing to make the same mistake over and over again but with ever-increasing conviction. There is no doubt that at least some of this attitude persists in medicine. We are nowhere near the levels of self-examination and scrutiny that the aviation industry, and so many other fields of human endeavour, have already achieved, but there are signs that this is finally changing. Many hospitals now encourage incident reporting to see what can be learned from adverse events. Some even encourage near-miss reporting. In 2002, the Virginia Mason Hospital in the USA introduced ‘Patient Safety Alerts’, essentially a system for reporting anything that could potentially harm a patient. Take-up of the system by staff was muted at first, but, in 2004, a 69-year-old patient, Mrs Mary McClinton, died at the hospital as a result of an injection error, and this galvanised the staff into reporting near misses. As a result, the numbers of errors and medical accidents were drastically reduced, and the hospital benefited from a substantial reduction in its medical-liability insurance premiums. Elsewhere, near-miss and direct-hit reporting is also slowly increasing, though the systems designed to deal with both of these are still relatively rudimentary.
In Better, the American surgeon and writer Atul Gawande reveals how doctors can improve their services to patients in many walks of life, from the latest Western high-tech medicine to the provision of the most basic healthcare and vaccination programmes in remote villages in India. When the book was first published in 2007, I had the good fortune to be asked to review it for a medical journal, and I thoroughly enjoyed reading it. Here was someone who clearly had a longstanding and deep interest in the fields of safety, efficiency improvement, and error reduction in surgery. What’s more, one of his ideas was to me an absolute revelation. He wrote that he realised that he could improve the care of his patients much more effectively not by finding new and cleverer things to do, but by continuing to do exactly what he already does, except doing it a little better.
In 2010, 817 people died in aeroplane crashes worldwide. In 2011, 17 American civilians died in terrorist attacks worldwide, which, incidentally, is about the same number as died by furniture falling on them. Between May 2010 and May 2012, nobody died of a terrorist attack in Britain. Now try to work out in your head the millions, if not billions, spent on aviation safety and terrorism prevention, and let us put cardiac surgery in context. About 39,000 major heart operations are carried out in the UK every year. The mortality is around 2 per cent, so that means around 780 patients die every year as a result of or after heart surgery.* This is the equivalent of two fully laden jumbo jets. Imagine the reaction of the media and the public to the loss of hundreds of lives if two jumbo jets crashed every year in the UK. Try to envisage the investigations, reports, mechanisms, safety measures, recommendations, and endless legislation that would follow such air-travel catastrophes.
[* In Australia, with a similar mortality rate but a smaller population, the comparable figure is under 400 patients per year; in the US, the comparable figure is more than 6,000 patients per year.]
Of course, we can expect there to be deaths in major surgery such as heart operations, and the death rate will never be zero. Nevertheless, I hope that I have shown by now that simple steps can be taken to reduce this death rate, that some of these steps have been taken, that there is yet more to do, and that this a field where a little effort to do things a little bit better may well be richly rewarding for all of us.
6
Decisions, Decisions
Decades ago, patients pretty much did what doctors told them to do. The doctor, as expert, was trusted to know the right answer to a problem, and the patient, as a general rule, went along with the treatment prescribed by the expert. This is changing. The role of the passive patient has gradually shifted into that of an active consumer of health care. The catchphrase is ‘No decision about me without me’. This is an understandable and legitimate shift. In making healthcare decisions, the patient makes a choice between no treatment and what can be an array of different available treatments, each with its own set of risks and benefits, and such decisions require information and intelligent process. In fact, whether a certain patient receives a certain treatment is perhaps the most important question that doctors and patients ever tackle. From the doctor’s point of view, this is probably the one thing for which a medical education is essential. Pretty much everything else that is done by doctors and surgeons can be done by someone else with the right training.
Here are some of the things doctors do routinely: they take and record a medical history from patients, take blood samples, insert a catheter into the bladder, do an electrocardiogram, order a chest X-ray, prescribe medications, put up intravenous infusions, harvest veins from the leg for CABG, resuscitate a patient in cardiac arrest, and so on. Only 30 years ago, the mere idea of a non-doctor performing these duties would have been anathema to doctors and non-doctors alike. Nowadays, all of these tasks and more are performed by nurses and other healthcare professionals routinely and safely in many hospitals and in many parts of the world. The one thing that really, absolutely needs a full medical education is the thorny issue of selecting the right treatment for the disease, and offering it to the patient. In short, it is the decision-making. For that, doctors need a thorough education to make the right decisions, and patients need a reasonable level of knowledge to agree to or reject the treatment on offer. This is all the more important in surgery, as the consequences of the wrong decision (or the right decision that goes awry) can be disastrous. One day, you or a loved one will be faced with the offer of an operation. Deciding whether to accept that offer is not straightforward, but it can be done intelligently with a little understanding of the mechanisms necessary for decision-making. Before I explain the evidence-based way in which doctors go about this decision-making, we need to ask ourselves a simple yet crucially important question: why do doctors treat patients?
Why indeed? It is a question that I often put to my fourth-year medical students, and some of them get the answer right, but many get it wrong, so let’s begin by clarifying why doctors should not treat patients.
Poor reasons for treating patients:
the patient has a disease
the doctor has a treatment
the doctor needs money
the doctor wants the patient to go away quickly
the patient wants to go home with some form of treatment.
All of these are not good reasons to administer medical treatment. There are only two good reasons for a doctor to treat a patient, and they are:
to improve symptoms (help the patient to feel better)
to improve prognosis (help the patient to live longer).
Any medical treatment that does not achieve at least one or the other of the above objectives should absolutely not be offered to the patient, because this would be unethical. A treatment that neither improves symptoms nor improves prognosis is at best useless and at worst both dangerous and a costly waste of resources.
The only exception to the above rule is in the field of immunisation, where on occasion it can be justified to offer vaccinations against a particular disease to all people (even those who are not susceptible to the disease) in order to increase the ‘herd immunity’ and thus eradicate or drastically reduce the impact of a nasty disease. Even then, receiving the vaccination can often have prognostic benefit to the individual, so that the rule remains intact in most vaccination programmes.
Surgery is no exception. The reason, or ‘indication’, for an operation is always either symptomatic (to reduce or abolish troublesome symptoms, such as pain, discomfort, breathlessness, itching, palpitations, and so forth) or prognostic (to improve the likelihood of survival). In other words, the symptomatic indication deals with quality of life, and the prognostic indication deals with quantity of life.
So our original question naturally branches out into two parts:
Is the operation the right choice to relieve the symptoms?
Will the operation improve survival?
Let us deal first with the symptomatic indication, which is the simpler one of the two. In surgery (as opposed to general medicine), the symptomatic indication must fulfil an essential criterion that is always the same, regardless of the patient, the nature of the symptom, the surgical specialty, the surgeon in question, and the envisaged operation. It is a simple criterion that applies to all: there is no indication for surgery for symptoms until medical treatment has failed. The reason for this is obvious: patients in their right mind would never choose to have a hairy-armed surgeon cut them open with a knife if there is a tablet that achieves the same symptomatic relief. So the first question becomes: is the symptom successfully controlled by tablets? If the answer is ‘Yes’, keep taking the tablets. If the answer is ‘No’, consider surgery.
When medical treatment with tablets has failed, and surgery is being contemplated for symptoms, the next decision is whether to take the plunge and have the operation, and that is where risk comes into it. The patient must weigh up how troublesome the symptom is against the risk of the operation. For example, if a patient has angina, and the risk of a CABG to cure the angina, as calculated by a risk model (and perhaps adjusted to the surgeon’s own performance) is a mortality of 1 per cent, the patient needs to decide if he or she is prepared to accept a risk to life of 1 per cent to get rid of the angina. This is a relatively easy decision to make: the patient knows how troublesome the angina is, and its impact on quality of life, and most patients can easily understand what 1 per cent, or one in 100, means. The patient can therefore make an informed decision as to whether he or she thinks the benefit of the operation is worth the risk.
