Why anaesthetising a patient is more prone to failure than flying a plane

Anaesthesia today is safer than it has ever been 1, yet public, media and institutional alarm about safety in modern medicine (including anaesthesia) is probably at an all time high 2-4. I will attempt to explain this apparent paradox, and more importantly, argue why it should act as a stimulus for an informed, systematic approach to more effective safety strategies. One of the central ideas to come from the study of high technology systems, including aviation, nuclear power and space exploration, is that the probability of an accident or failure can never be absolutely zero 5. No new safety device or extremes of human effort can ever remove entirely the chance of an error or failure occurring. Health care is one of the most highly developed technological systems in existence, and the number of health care procedures carried out each year is increasing. As the number of procedures increases, the number of patients harmed by them must also increase (given a constant, non-zero chance of harm). Thus, medicine can be safer than it has ever been in terms of relative risk, while simultaneously causing harm to a record number of patients. In our perception of risk, humans have been called non-Bayesian beings 1. This is because, for any phenomenon, we tend to pay attention to only the total number of bad outcomes, and ignore the associated number of trouble-free outcomes. We have a fixed idea of how many plane crashes or medical mishaps are tolerable each year for these technologies to continue to be considered safe – regardless of the total number of planes in the sky, or medical procedures conducted. The current alarm about the safety of medicine indicates that the number of patients harmed each year may be approaching the fixed level over which we will no longer view medicine as safe (Fig. 1). For example, between 44 000 and 98 000 deaths are estimated to occur in US hospitals each year due to preventable adverse events 3. What is required, as the Institute of Medicine (IOM) Report puts it, is a ‘threshold improvement’ in medical safety to reduce the total number of patients harmed by their treatment 3. The IOM Report does not mince its words when discussing safety in medicine, stating that: ‘The status quo is not acceptable and cannot be tolerated any longer’, and ‘Health care is a decade or more behind other high-risk industries in its attention to ensuring basic safety’. The Report promotes the systems approach to safety developed in other high-risk industries. Such an approach demonstrates that greater safety gains can be achieved by removing dangerous aspects of systems than by any other known method. Commendably, a recent letter by Dr Leng 6 calls for the support of the systems approach in improving safety in anaesthesia. However, the letter goes on to claim that recent safety discussion in this journal has looked no further than drug labelling, and suggests a number of more or less commonsense approaches to the reduction of error, unrelated to labelling, many of which are already standard practice. Obviously, system redesign should not stop at labelling, but it must be remembered that iatrogenic harm is not a homogenous problem. While it is important to keep the wider system in mind when considering safety initiatives, individual safety problems can only be identified and dealt with one at a time in a piecemeal fashion. Ignoring an obvious problem (like poor labelling) for the ‘wider picture’ carries the real danger of turning into an excuse to do nothing. Drug administration error is such a serious problem that it was singled out as a leading cause of harm in the IOM report 3. Our own multicentre study found that a drug error occurs once for every 133 anaesthetics conducted and estimates that in the course of a 30-year career, every anaesthetist can expect to harm an average of two patients through drug error 7. Incident data and problem reports in anaesthesia consistently point to drug labelling as a causative factor in errors. As such, there would appear to be an overwhelming body of evidence indicating that something needs to be done about this problem. Labelling would seem to be an excellent and the most obvious place to start in terms of redesigning systems in anaesthesia. Farcically, some hospitals in Britain currently have ampoules labelled in only Italian and German 8– could drug labelling get any worse, even if we tried? This is not a time for business as usual – existing safety strategies are unlikely to be sufficient to attain a threshold improvement in safety. Good quality incident data needs to be collected to identify problem areas throughout anaesthesia, and action needs to be taken to find significant, new ways to solve the problems thus identified 7, 9, 10. Such solutions must also take into account the psychological reality that we are all human 10, 11. Anaesthesia often looks to the aviation industry for ways to improve safety, and there is no doubt that this is an excellent approach. However, there are a number of important differences between flying a plane and anaesthetising a patient. The complexity of a modern aircraft is high and the number of potential failure modes even higher. However, the complexity of a patient's physiology is a number of magnitudes higher still, and this complexity is increased when the patient becomes part of the machinery when connected to a large amount of high-technology equipment in the operating theatre during an anaesthetic. In addition, the patient will likely be undergoing a destabilising and invasive procedure during the anaesthetic, all of which will be delivered and managed by teams of people in the operating theatre and beyond. The result is that the greater human-machine system, in which the patient is embedded, has many more failure modes than any man-made object, such as an aircraft. Therefore, anaesthetizing a patient would seem to be a more complex and unpredictable task than flying a plane. Ironically, it is the safety culture of anaesthesia that is the least developed. The traditional reliance on the resolve and vigilance of clinicians to avoid dangerous outcomes is no longer adequate to ensure patient safety. Systems must be redesigned to remove unneeded, dangerous alternatives and to better support human performance. Such redesign should start with existing, well-known problem areas.