5

Better Out of FIASCO

We often hear comparisons between surgeons and pilots (your life in their hands and all that). This has always struck me as somewhat superficial, because there are enormous differences between the two professions. For a start, a pilot flies an aeroplane, whereas a surgeon cuts people with a knife (‘very, very carefully’, according to my daughter Claudia, then aged five). An important similarity is that, of course, members of the public are usually prepared to place their trust — indeed lives — in the hands of pilots and surgeons alike. We rightly expect both pilots and surgeons to do their job ‘very, very carefully’, but there the similarities stop. On the one hand, the pilot is actually in the same boat (or aeroplane, to be literal) as the passengers. If the plane crashes and everybody dies, so does the pilot. A surgeon, on the other hand, will survive any number of his or her patients’ ‘crashes’. The surgeon’s reputation and career may not survive, but the surgeon almost certainly will. A surgeon takes risks with the lives of others. A pilot risks his or her own life. An even more important distinction between a surgeon and a pilot is that they are radically different in their modus operandi.

A few years ago, I thought that I would try to train for a private pilot’s licence. I had romantic delusions about flying to France for lunch, and the more prosaic notion that when I needed to travel somewhere quickly for a meeting, it would be easy and relatively inexpensive to charter a single-engine Cessna and go anywhere in the country without ever again having to endure the appalling degradation of flying with Ryanair. I also thought it would be cool. It was a short-lived venture. In the end, I gave it all up after half a dozen lessons, having learned to take off and to cruise but not yet to land. The reason for this was that aviation safety meant that I simply did not have enough time. The first flying lesson took an hour, and was fine. The second was preceded by a safety briefing that lasted 20 minutes. The third had a briefing and a debriefing. The fourth added safety checks to the briefing and debriefing. By the fifth lesson, I also had to become something of an amateur aircraft mechanic, checking obscure bits of the engine and fuel systems, and the supposed one-hour lesson was taking all day. From that ephemeral venture into aviation, however, I learned one of the biggest differences between pilots and surgeons. Pilots follow procedure, check and recheck, and then check again. They do their utmost to avoid leaving anything to chance. It’s a belt-and-braces hold-on-to-your-pants-with-both-hands-just-in-case approach. Surgeons, in comparison with pilots, are gung-ho wide-eyed cowboys.

Two examples from heart surgery illustrate the point. Surgeons join most blood vessels using a single-filament suture made out of polypropylene. The suture is usually Prolene, a brandname product made by Ethicon, an international suture-manufacturing company. Prolene has many interesting features, some of which are good and some of which are bad. It is blue in colour, which is ridiculous, as it often becomes invisible against the ubiquitous blue background of the sterile operating-theatre drapes; and it has the handling characteristics of pubic hair, being both crinkly and unwieldy.

Despite these disadvantages, Prolene is the preferred suture worldwide for joining up blood vessels. There are three main reasons for this. First, it is slippery, so it glides through tissue swiftly and smoothly without causing damage. Second, it is inert, and the body happily accepts it; there is no evidence that it causes inflammation or any other adverse reaction in the tissues. Third (and most important), it is very strong, and retains its strength more or less forever. Blood vessels carefully sutured or joined with Prolene can withstand high arterial blood pressure, even when the calibre of the Prolene suture is so small as to be finer than a human hair. In a quadruple coronary bypass operation, there will be at least seven suture lines where a Prolene suture is the only structure protecting the patient from fatal bleeding. It is, literally, the thin blue line between life and death.

There are, however, two features of a Prolene suture that should modify the behaviour of surgeons when they are handling it. The first is that Prolene is slippery. The downside of its smooth passage through tissue is that a knot made out of it can slip if not properly ‘locked’. To lock a knotted suture requires either a change of hands or a slight change in hand position while tying the knot, not exactly an arduous task. There are about three or four manoeuvres that will securely lock a knot, and all surgeons are aware of at least two of them. Careful surgeons begin by tying a number of slipknots, anything between four and seven. The slip allows the knot to be tightened to the desired firmness. They then do one of the available manoeuvres to lock the knot securely. In terms of time, the locking manoeuvre adds between half a second to three seconds to the procedure. In most operations on the heart and major blood vessels, a single running Prolene suture is used for joining and repairing blood-containing structures, including arteries that hold blood under high pressure, such as the aorta, which carries the entire blood output of the heart within its walls. So the slipping of a knot on the aorta can lead to the unravelling of an entire suture line and the patient bleeding to death in a matter of minutes. Knowing this, you would have thought that all surgeons would lock a Prolene knot, especially on an artery, and even more so if the artery happens to be the aorta. The shocking truth is that they do not. Of the surgeons I know, fewer than half properly lock every Prolene knot, even on an artery, even on the aorta itself.

The second drawback of Prolene is that it is a monofilament. Most strings in everyday use, such as the cotton you sew a garment with, the mooring line for a boat, or your shoelaces, are braided or multifilament: the string or rope is made up of several smaller strings entwined about each other. Some sutures other than Prolene are also braided or multifilament. The main advantage of a braided suture is that a fracture of one of its tiny strands does not jeopardise the integrity and strength of the whole suture. Prolene is different. It consists of merely one strand of material. In fact, if you look at Prolene under a powerful magnifying glass, all you will see is a single shiny, smooth, blue cylinder. A fracture of this one filament weakens or breaks the entire suture. So the manufacturers quite rightly advise against handling the suture with metal instruments such as surgical forceps, clips, or clamps for fear that a crack or fracture might weaken the suture, causing it to snap and (being slippery) unravel so that an entire suture line risks coming undone. Knowing this, you would have thought that no surgeons in their right minds would dream of touching a Prolene suture with metal forceps, or applying a metal clip or a clamp to it. The shocking truth is that most surgeons do, even if that suture is being used to repair an artery, and even if that artery is the aorta itself.

Fortunately, most multiple slipknots do not actually slip, and most Prolene somehow survives the assault of metal instruments, but not all. Over a 20-year career, I have personally seen a handful of patients bleed out and die from slipped knots and fractured Prolene. One was a man who was sipping his morning cup of tea after a successful CABG, when he bled out into the chest drains and died in a couple of minutes before anything could be done to retrieve the situation. Post-mortem examination showed that one of the joins of the coronary bypass had come undone because the Prolene was fractured. A metal clamp had been applied to it during the operation. Another was a man who had survived a difficult and high-risk heart transplant, only to collapse two hours later, having bled from an unravelled suture on the right atrium, a chamber of the heart. The team had got to him in time to re-open the chest, stop the bleeding, and massively transfuse him, but it was too late: he was brain dead. When his chest was re-opened, lying in a corner of the membrane enclosing the heart was an accusing, finger-pointing little blue twisted thread: the slipped Prolene suture, which had not been locked.*

[* That unfortunate patient became an organ donor, and his kidneys and liver were taken for transplantation to others.]

The numbers of such calamities are minuscule, but even one is too many when the safety step required to avoid it is so simple and quick. Take a few extra seconds to lock a Prolene knot securely, and handle the Prolene with gloved fingers only or, if you must, with metal instruments whose sharpness is tempered by a rubber sheath (so-called rubber-shod forceps), and these calamities will never happen. The two surgeons involved in these cases learned their lesson the hard way, and now always lock sutures and never handle Prolene with metal. A third, a colourful and engaging bon vivant who was one of the greatest heart surgeons in Britain, was a supremely confident operator, thinker, and innovator. He had excellent clinical outcomes, and won the respect and admiration of the medical establishment in Britain and internationally. He was not shy of publicity, and at one point in his career came as close to being a ‘household name’ as any could in this profession. He, too, had a near miss, of a similar nature, when a slipped Prolene knot on the aorta unravelled a few hours after an operation. He managed by the skin of his teeth to save the patient from death by blood loss. He nevertheless continued his merry practice of not locking slipknots until his eventual retirement a few years ago.

In an era where we are supposed to be doing everything possible to make surgery safer, to reduce to the best of our ability the potential and actual harm to patients, and to strive endlessly to improve results, basic safety steps are not being taken. A large part of the reason is that surgeons are natural risk-takers. They occasionally cut corners, often make decisions on the hoof, and tend to act in a cavalier way. The obsessive and self-examining culture that prevails in aviation has not yet taken hold in the field of surgery. In aviation, all possible care is taken all the time. In medicine, as the following study demonstrates, it is not.

Recently, one of my colleagues, Catherine Sudarshan, decided to look into the care of Jehovah’s Witnesses having major heart operations. Jehovah’s Witnesses are a broadly Christian denomination, but with several beliefs that distinguish them from the mainstream, including the rejection of festivals such as Easter and Christmas as essentially pagan, and the belief that the Kingdom of God was established in October 1914 and has room for only 144,000 believers. From the point of view of a heart surgeon such as Catherine, the most important attribute of Jehovah’s Witnesses is that they absolutely reject blood transfusion. For obvious reasons, blood transfusion can be essential for heart operations. In the early days of heart surgery, hardly an operation was carried out without the need for blood transfusion, and often in huge quantities. As we have got better at it, and our machines have become more sophisticated and compact, we have reached the stage where only around a third of patients need to be given any blood at all. A third, however, is a large minority, and the risk of needing at least some blood or blood products is very real.

A heart surgeon’s own heart usually sinks when a Jehovah’s Witness enters the consulting room. A patient who categorically refuses to have a blood transfusion takes one potential safety net away from the operation, and the surgeon knows that if major bleeding occurs during or after the operation, death is likely to follow. Some surgeons flatly refuse to operate on these patients, and most would refuse if they considered it likely that the proposed operation would result in excess bleeding. Catherine wanted to see what happens when Jehovah’s Witnesses are operated on. Do they do worse than those patients who are prepared to accept blood? In other words, does the removal of the safety net actually damage them?

She studied all the Jehovah’s Witnesses who received major heart operations at Papworth Hospital in the past few years, and found that they actually do not fare worse than other patients. There was no difference in their risk-adjusted mortality, and, overall, their cardiac outcomes were similar. In one important area, however, they actually did better than the others: they lost a lot less blood. The average blood loss in Jehovah’s Witnesses was a mere 272 millilitres 12 hours after surgery. Other patients had an average blood loss of nearly double that (498 millilitres). What is happening here is obvious: because the patients refused blood transfusion, the surgeons must have taken more care to control bleeding. This begs the question: if it is possible to take more care with Jehovah’s Witnesses, should we not take more care with everybody?

Common to both aviation and medicine is the concept of the near miss. It also demonstrates beautifully why surgeons are nothing like pilots. All it takes to constitute a near miss in aviation is for two planes to come reasonably close to each other so that an accident may have been possible. In aviation, therefore, a near miss is just that: an accident that could have happened but absolutely did not. In medicine, however, there are three types of near miss, which I classified in an article in The Lancet (Nashef 2003). In type 1, a mistake is made, the systems designed to detect it and correct it work as planned, and nothing happens. In type 2, the error is made and the safety systems fail, but no harm is done thanks to sheer luck alone. In type 3, harm is done, but it falls short of a direct hit — death or disability, or whatever is the outcome that is being studied. Near misses in aviation are overwhelmingly type 1, far more benign in nature than many near misses in medicine, and yet the aviation industry approaches them with absolute earnestness. All near misses are reported, collected, and scrupulously analysed. Lessons are learned from them, and changes in practice and protocol are introduced as a result. After all, it is so much safer (and more intelligent) to learn from a near miss than from a direct hit. Owing to near-miss observations and other technological improvements, the current rate of fatal accidents in air travel has dropped by about 65 per cent, to one fatal accident in about 4.5 million aeroplane departures, from one in nearly 2 million in 1997. We have no such systems in medicine. With occasional exceptions in the fields of drug prescription and blood transfusion, near-miss reporting is still in its infancy. There is some near-miss reporting in some surgical specialties, but in others it is almost non-existent, and in medical specialties it is totally non-existent. In clinical practice, on the whole, we are sadly still in the rudimentary stages of learning from a direct hit.

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.