Lest we forget : learning and remembering in clinical practice

This issue of Anaesthesia carries two articles inspired by a single incident in an intensive care unit (ICU) in which a patient suffered neuroglycopenia when insulin was infused to treat artefactually high blood glucose levels. This came about because of contamination of arterial samples by a glucose-containing arterial flush solution. The first article is a detailed analysis of the sequence of events leading to the final, fatal outcome 1. The second is a survey of ICUs in the UK, where the researchers enquired about the management of arterial cannulae, especially how flush solutions are used and how samples are taken from lines 2. This case is very similar to one reported by Sinha and colleagues in Anaesthesia in 2007 3. This journal would not usually publish duplicate case reports; indeed, novelty is our main criterion for consideration for publication. Before we open the editorial floodgates to a deluge of duplicate or subtly different case reports, we would like to explain why we have apparently contravened our own editorial policy. First, there are a number of clinical differences between the case reported here and the 2007 one. In Sinha et al.'s report, the events took place over a few hours in an (initially) conscious patient who then lost consciousness. In the present case, the patient was sedated at the time the incident started, making signs of hypoglycaemia harder to detect. The presumed period of hypoglycaemia was also probably much longer. Second, the authors of the present report have used a systematic approach to identify causal and contributory factors, based on a recently developed framework 4. As is usual in such events, one ‘active error’ was compounded by many other contributing factors. Third, the authors provide evidence that problems with arterial lines, similar to those described but not necessarily harming patients so severely, are much more common than might be imagined. So it becomes possible to view this piece not, like some recent case reports, as a rare clinical oddity 5, but as a visible, published example of a more diffuse and hence less easily discernible threat. Fourth, and most importantly, however, the present case occurred some years after national guidance was issued by the National Patient Safety Agency (NPSA) 6. This guidance was produced in response to incidents similar to, and including, Sinha et al.'s, and the present case occurred in an ICU that had followed the recommendations issued. As such, the case demonstrates that problems can still recur when preventive guidance is available and has apparently been followed. The authors' accompanying analysis explores what additional measures might be needed 1. To me, the case raises a number of general points over and above the specific clinical problem described. One relates to risk and our ‘blindness’ to it. Arterial cannulation is commonplace and is conceived as a low-risk procedure. Insertion is generally safe 7 and problems during use appear to be uncommon. However, this shows our blindness to risk in two ways. Problems with any single low-risk activity or substance may seem to be rare, but healthcare (especially in ICU) is complex and there are many possible ways in which problems can occur. Hence, although one specific problem may be uncommon, the cumulative incidence of all and any rare events is actually quite high. We do not see this because it is not intuitively obvious. In addition, because the problems do not usually occur in clusters, we do not see them because they are, indeed, mostly invisible to us as individuals. Only when we look on a larger scale, as the reports from the NPSA's National Reporting and Learning System cited in the articles 1, 2 reveal, can we ‘see’. This blindness encapsulates our difficutly tackling avoidable harm in healthcare. Deaths occur one at a time, in different areas of the hospital, or different parts of the NHS, rather than in easily visible masses such as in the battles of the First World War or large-scale transportation disasters. Another aspect of the reported case relates to guidance, recommendations and advice. The national survey 2 reveals that staff in many ICUs were unaware of the NPSA's 2008 guidance. This is not in itself surprising. Writing in the British Medical Journal in 2011, Carthey and colleagues noted that an individual hospital might have hundreds of guidelines on its intranet system, and practitioners must also understand and comply with many other local, professional and governmental policies 8. They observed that the sheer volume of guidance, the difficulty locating relevant material when it is needed, and the length and complexity of documents, all conspire to deter professionals from using them. Guidelines can also contradict each other. Sometimes two different guidelines on the same topic are produced for different settings 9; sometimes different groups of experts many disagree in their interpretation of the same research evidence 10. Sometimes, though, two apparently unrelated guidelines, each designed to address a specific safety problem, can act against each other. In 2007, guidance from the NPSA recommended avoiding hypotonic intravenous fluids in children to avoid iatrogenic hyponatraemia 11. Perversely, this was one of the reasons why a bag of 0.9% saline with 5% glucose was available in the ICU where the incident reported by Gupta and Cook occurred, enabling the half-covered bag to be misread as 0.9% saline only 1. Furthermore, medical practice is constantly evolving 12 and so guidance has to change and evolve also. Given this, coupled with the endless creativity of human beings in finding new ways of working round and within complex systems, and the pure capriciousness of chance, it is no surprise that problems can still occur even when the guidance is followed and practice is ‘optimal’ 13, 14. But guidance itself is not enough. Simply exhorting people to ‘do better’, ‘check more carefully’, ‘be more vigilant’ and so on, is easy but hollow. What really works is helping them understand quite how they might do better, what exactly they might be checking for, and what precisely they could hope to see if they remain watchful. What this needs is memory. Memory brings context, comprehension and continuity. A few examples will illustrate the point. It is recommended practice now not to rehang bags of intravenous fluid once they have been detached from the giving set to the patient (for instance, to substitute intravenous antibiotics). Some of us are old enough to remember the 1997 Hazard Notice from the then Medical Devices Agency, issued after two patients died when re-attached bags containing air as well as the rest of the fluid were pressurised to speed infusion, leading to fatal air embolism 15. Some who are struggling with implementing the change to non-Luer lock neuraxial needles will not understand how and why the move was conceived, but to those who remember the accidental intrathecal injection of vincristine to an 18-year-old called Wayne Jarrett in Nottingham in 1999, stimulating the search for ‘design solutions’ to prevent recurrence of the problem, it makes more sense 16. The exhortation to check the patency of the whole anaesthetic breathing system, including the facemask and angle piece, before use, stems from the death of Tony Clowes, a 9-year-old boy, from Essex, in 2001. Here, the cap from the end of an intravenous giving set had become impacted, unseen, in the opaque angle piece. The cause of the blockage was not recognised and the boy suffered hypoxic brain damage 17. There are many other such examples, but my point is that an awareness of the particular incidents concerned gives them meaning and makes them easier to remember and harder to repeat. Our collective memory for problems helps protect us and our patients from the encroachment of known harms. This was recognised nationally some years ago 18. The issue is one of remembering, forgetting and being reminded of things we have once known well but that have become less familiar to us with the passage of time. What will strengthen this collective memory? Explicit written safety guidance plays its part, but we need to acknowledge the informal knowledge gleaned from individual clinical experience too 19, 20. Depending on when we started in practice (and how well we remember), each of us will have a different series of snapshots of the ever-changing pattern of clinical activity. Talking about problems helps too; airing and discussing cases has a long history in medicine. It can generate lively interest during departmental audit meetings but may be squeezed out by other less engaging agenda items. The Safe Anaesthesia Liaison Group (SALG), a joint venture between the Association of Anaesthetists of Great Britain and Ireland (AAGBI), the Royal College of Anaesthetists and NHS England (which has assumed the responsibilities of the NPSA), will soon be releasing a ‘morbidity and mortality toolkit’ with the aim of breathing new life into this long-practised activity. Publishing case reports is also valuable 21, as the article in this issue demonstrates; journals have a unique function as a repository of clinical wisdom as well as scientific investigations 22. On a larger scale, analysing and learning from incident reporting is vital 23-26 and this is another of SALG's functions 27. Indeed, in Australia, a set of crisis management checklists to help anaesthetists deal with intra-operative emergencies was developed from the Australian Incident Monitoring System; these encapsulate the learning from the problems reported 28. It is commonplace to have safety materials from the AAGBI, for instance for machine checking 29, 30 and the management of anaphlaxis, available in the theatre suite. However, we could go further. I recommend, in line with the Helsinki Declaration on Patient Safety in Anaesthesiology 31, that the full set of crisis management checklists is to hand in each anaesthetic room; confirming this should be one element of the World Health Organization's pre-operative briefing 32. Similarly, perhaps a more formal procedure, akin to the surgical checklist, could be encouraged for shift changeovers on ICU. Guidelines, too, could incorporate practical experience as well as research evidence 33. Lastly, it would be possible to continue reminding ourselves of the lessons from history by publishing regular updates with current interpretations of old lessons and alerts. Formal safety guidance is always needed, of course, and the AAGBI's working party on Preventing Iatrogenic Hypoglycaemia due to Inappropriate Insulin Therapy in Critical Care Practice will report on this specific issue in the next few months. To complement such guidance, though, we need the collective memory described above, to remind ourselves continually of what we know; the individual ‘mindfulness’ that our patients are at risk both from old and new threats; the sharpness of mind and eye to recognise these when they appear; and the willingness to report and share mishaps when they occur. We owe this to ourselves but also to the people under our care. We need to make sure that past patients did not die or suffer in vain and future ones will not do so needlessly. No external funding declared. Dr Smith is a member of the Safe Anaesthesia Liaison Group referred to above.