Intubation training in the real world: a defence of the Northwick Park drill

Five years ago the Royal College of Anaesthetists (RCoA) [1] expressed concern over an apparent increase in airway problems; some have discounted this as anecdotal, but it is undeniable that the last three Confidential Enquiries into Maternal Death [2] reported a rising number of deaths related to general anaesthesia (0, 3 and 7 respectively). The statistical significance of this cannot be calculated for lack of denominators but, using reasonable estimates, the likelihood of this trend being due to chance is very low. Those who remain complacent should consider these facts. Failed intubation rates vary enormously in different units and an experienced anaesthetist writes that he is ‘disillusioned with intubation’[3]. A failure rate of 1 in 40 [4] understandably causes disillusion; by contrast, a failure rate of 1 in 1500 [5] shows that, with proper training, intubation is our most trustworthy technique, much more so than epidurals, for example. These facts surely require investigation. Failed intubation in obstetrics provides a valuable yardstick of trainee performance. The best figures (1 in 800) are from Northwick Park Hospital [6], the only comparable result (1 in 750) is from Durban in South Africa [5]; by contrast, a recent UK figure [7] is 1 in 112. Does this reflect a demographic difference? That seems unlikely, as two quite different populations, Northwick Park and Durban, gave similar results; a more plausible explanation is that those two units used similar training, based on the Macintosh method [8, 9]. In essence this uses a gum elastic bougie projecting 6 cm from the end of the tracheal tube. If used with a Magill tube then the tube tip may be obstructed at the cords [10]. To avoid this, Macintosh used a tube with the bevel rotated 90°; its safety and effectiveness was demonstrated 40 years ago in 18 000 patients [11] and verified in two later studies [9, 12]. To avoid confusion, trainee results will be labelled ‘first-up’ failure rate; if a senior colleague then succeeds, the ‘overall’ rate is lower. Thus in Durban [5] the first-up rate was 1 in 750, overall 1 in 1500. Traditionally, trainees are supervised for about 40 cases, then mainly teach themselves. This means that by the time they start working alone they may never have encountered a difficult intubation; we therefore argued the need for simulating difficult cases as a training strategy [9]. The Northwick Park drill uses the Macintosh method but instead of watching the tip of the bougie enter the trachea, the epiglottis is allowed to descend, hiding the cords, as in real grade 3s. Intubation is done blind and entry confirmed by capnography. Total time taken is no longer than a simple intubation, but only with practice [12]. If someone uses the Macintosh method for every intubation, regardless of whether the bougie is needed, then he will be using his most familiar technique when faced with an emergency grade 3. This is standard practice in a few areas. Our argument originally suffered from lack of data on the frequency of difficult intubation and the initial ‘guesstimate’ was too low, but it stimulated real surveys and these have supported our reasoning. The incidence of grade 3 cases is between 1 and 2%[5, 13, 14], which is close to the failure rate of 1 in 112 referred to above [7]; similarity between failure rate and grade 3 incidence is to be expected when there is no specific training for such cases. By contrast, the Durban study [5] of 1500 cases records 27 grade 3 views, all but one of which were successfully intubated, which confirms the benefit of regular practice with the bougie. If 1.3% of routine cases [13] have grade 3 views, then elementary laws of probability [15] enable us to predict what will happen (Fig. 1). Twenty-seven percent of trainees are expected to do 100 intubations without seeing a grade 3 case, but long before that they will be required to do unfasted cases unsupervised; if they encounter a grade 3 case then they are likely to fail, sometimes with a fatal outcome. Turning now to the right tail of the graph, 14% of trainees will see 3, 4 or 5 grade 3s in their first 100 cases and what happens then is still something of a lottery. Consider first the unlucky trainee. More than 20 different ways of handling difficult intubation are advocated and a leading textbook concludes [16] that there are no hard data as to their relative value. The trainee may be taught one method by one supervisor, another by the next, fails with both on the third, causing trauma and, not surprisingly, becomes disillusioned with intubation. Now consider the lucky trainee, who is shown the Macintosh method for his first grade 3; if he persists with this for a few cases then one might assume that he is well on the way to being expert. Unfortunately, that is quite wrong. Recent work [12, 17] indicates that learning takes much longer than was once thought; to get within 25% of the ‘expert time’ is a modest aim, but even that requires about 32 repetitions. A simple calculation shows that it would take 12 years, on average, to see 32 real grade 3s for someone doing 200 intubations per year. Someone doing most cases with the laryngeal mask airway would still be untrained when he retires. The number of grade 3 cases seen by 100 trainees in their 1st 100 intubations, assuming random distribution with a mean incidence of 1.3%. Twenty-seven percent of trainees would see no cases (black block), whereas 14% would see three, four or five cases (white blocks). These predictions are from Poisson's equation, which is a simple approximation. They were checked against the exact binomial distribution; the difference was negligible, as expected for sparse data. These facts highlight the need for training in difficult intubation but, although our grading system [9] and its associated drawings were quickly accepted worldwide, the text has been largely overlooked. Confidential Enquiries have discussed antacids etc., but not intubation training. The Obstetric Anaesthetists Association has advocated that trainees should do 20 supervised obstetric sessions [18], but by itself that is no guarantee that they will know anything about difficult intubation. A major Hammersmith review [19] strongly advocated our drill, but appears not to have been widely read. Grade 3 cases are rare enough to escape primary training, but common enough to occur many times in a whole career. Recent developments in our strategy have been summarised both in print [20, 21] and in abstract form (DAS Abstract(A): Cormack RS Training for Intubation. 2004 Annual meeting). One caveat is needed. Plastic-coated wire introducers are sometimes used instead of a bougie and some are rigid enough to cause trauma [22]; they may be valuable in genuinely difficult cases, but for simulations the bougie should be used. Various criticisms have been made and we welcome these provided that critics have read what we wrote. For example, it is alleged [23] that emphasising the importance of intubation may lead to prolonged attempts – but for anyone who has read our guidelines [9] there is no risk of that. Again, it is implied that we advocate returning to a mythical ‘golden age’– but even the most careless reader will get our core message, that traditional training has a crucial defect. It is alleged that emphasising the importance of intubation will produce a sense of failure in trainees – but that happens already, and our drill cures it. Our drill is described as ‘outmoded’[24], whilst advocating the intubating laryngeal mask airway (ILMA) – but no supporting evidence is given, whereas our drill is backed up by the best published figures [6]. The ILMA is not suitable for emergencies or obstetrics; our method has been extensively tested in obstetrics [5, 6]. The ILMA has caused fatal oesophageal perforation [25] when used by an experienced anaesthetist – our drill has a blameless track-record, even by trainees. We are accused of ignoring the use of manikins – wrong again, see our abstract as cited above. A manikin designed by the Belfast team was demonstrated at the Glasgow Difficult Airway Society (DAS), 2003; laryngoscopy gives a grade 3 view, convertible to grade 1 by manipulating the cricoid. This should help in teaching laryngoscopy and for grade 3 simulations; it merits testing, and details can be obtained from Dr James Murray (james.murray@qub.ac.uk). Manikins may accelerate training, but in the end there is no substitute for real patients. Trainees are doing fewer intubations than in the past, which inevitably damages performance. Can more intubations be provided? One solution is to restrict use of the laryngeal mask airway during the training period. Against this it is argued that intubation is more likely to cause trauma. Experienced anaesthetists can testify to intubating over decades without trouble, yet litigation reports show an appreciable incidence of trauma, mainly dental damage. The explanation is that trauma typically occurs from inadequately trained personnel attempting grade 3 intubation who become increasingly frustrated by repeated failures; efforts lasting more than 1 h have been reported. This was a key factor prompting us to develop the drill. Trauma prevention is not an argument for the laryngeal mask airway; it is an argument for proper training. Lack of confidence with intubation may be a good reason for nurses, or paramedics, to use the laryngeal mask airway, but for someone aiming to be a full-time, professional anaesthetist it is an extremely bad reason – he or she needs to seize every opportunity of learning our most fundamental technique. Performance deteriorates without regular practice [26]. Some authorities advocate informed consent for all training procedures, but recently the Association of Anaesthetists of Great Britain and Ireland (AAGBI) [27] concluded: ‘Specific consent may, or may not, be required depending on the circumstances’. For the Northwick Park drill it is debatable whether informed consent is in the patient's best interests; a full evaluation is required. Proper training with the bougie should give a first-up failure rate of 1 in 1000 or better. An overall rate of zero may be possible with fibreoptically guided intubation, so is it worth it? Consider the benefits. Successful intubation would eliminate all airway disasters. Antacid regimens, failed intubation drills and difficult airway algorithms have become increasingly complex and costly and none is 100% successful, partly because of the difficulty of memorising all the ramifications. They would all become redundant if failed intubation were eliminated, so the stakes are not trivial. Many years ago, Mason [28] argued the importance of fibreoscopy, but emphasised that, in emergency cases, only the expert should attempt it, adding later [29] that attending a fibreoscopy workshop is not enough. It is now clear [17, 30] that even 20 supervised fibreoscopies do not produce an expert; this explains why many fibreoscopy studies report appreciable failure rates, yet those who do fibreoscopy clinics every week obtain 100% success. Some anaesthetists have ethical clearance to do all intubations by fibreoscopy without informed consent; that may be the only way to achieve complete success. Clearly this is only feasible for a proportion of anaesthetists, therefore the strategy depends on achieving a very low first-up failure rate. Recent developments in fibreoscopic methods [31, 32] may accelerate learning. The ratio intubation/laryngeal mask airway usage varies greatly in different areas and it is not yet possible to identify the optimum because of a lack of large-scale morbidity data [33]. Information technology will answer that in the future, but meanwhile a vague feeling that the laryngeal mask airway is somehow ‘less invasive’ must not outweigh training needs. Our argument is said to be not statistically conclusive, but recall that John Snow's surveys were retrospective and not randomised, yet they correctly identified the cause of cholera [34] and founded the specialty of epidemiology. Perfect trials should be done when possible, but the advance of medicine often takes place without them. Secondly, the best way to get cast-iron proof is nationwide testing of our strategy. Failed intubation is rare, therefore one or two hospitals cannot collect enough data. If, after multicentre testing, the trainee failure rate stays obstinately at 1 in 250 then that provides robust data against which to measure other schemes. For example, the French strategy goes direct from basic intubation to fibreoscopy (DAS Abstract(B): Cros Anne-Marie, Bourgain JL. Airway Training in France. 2003 Annual meeting). It may, or may not, give better results. Is this classical French logic vs Anglo-American hopeless muddle? Time will tell. On the other hand, if our strategy produces the improvement we predict, then that is substantial progress at negligible cost – at present the bougie is not widely used outside the UK. Only large studies like this can put the practice of anaesthesia on a rational base. So long as any anaesthetist is ‘disillusioned with intubation’ something is badly wrong; a trainer has done well when his protégé sees that intubation provides as near a perfect airway as we are ever likely to get. The two DAS Abstracts referred to are unpublished, but obtainable on request from the Archivist of the RCoA. Dr Cormack took part in designing the tube described by West et al. [12], but has no financial involvement.