Adverse events in anaesthetic practice: qualitative study of definition, discussion and reporting

Abstract

Background. This study aimed to explore how critical and acceptable practice are defined in anaesthesia and how this influences the discussion and reporting of adverse incidents.

Method. We conducted workplace observations of, and interviews with, anaesthetists and anaesthetic staff. Transcripts were analysed qualitatively for recurrent themes and quantitatively for adverse events in anaesthetic process witnessed. We also observed departmental audit meetings and analysed meeting minutes and report forms.

Results. The educational value of discussing events was well-recognized; 28 events were discussed at departmental meetings, of which 5 (18%) were presented as ‘critical incidents’. However, only one incident was reported formally. Our observations of anaesthetic practice revealed 103 minor events during the course of over 50 anaesthetic procedures, but none were acknowledged as offering the potential to improve safety, although some were direct violations of ‘acceptable’ practice. Formal reporting appears to be constrained by changing boundaries of what might be considered ‘critical’, by concerns of loss of control over formally reported incidents and by the perception that reporting schemes outside anaesthesia have purposes other than education.

Conclusions. Despite clear official definitions of criticality in anaesthesia, there is ambiguity in how these are applied in practice. Many educationally useful events fall outside critical incident reporting schemes. Professional expertise in anaesthesia brings its own implicit safety culture but the reluctance to adopt a more explicit ‘systems approach’ to adverse events may impede further gains in patient safety in anaesthesia.

The specialty of anaesthesia pioneered incident reporting in healthcare, transferring the concept from other high-risk industries such as aviation.1 In the UK, the Royal College of Anaesthetists (RCA) has both provided a widely used working definition of a critical incident (‘a critical incident is one which could have led to harm; it could have been prevented by a change of process’) and endorsed a reporting template.2 Incident reporting is now being more widely promoted. Many hospitals have had generic ‘clinical incident reporting schemes’ in operation for some years and now a national system for reporting of adverse incidents and near misses has been established by the National Patient Safety Agency (NPSA).3,4 Despite this widespread promotion of incident reporting, little attention has been paid to how potential threats to patient safety are recognized, defined, discussed and reported in actual clinical practice. If incident reporting is to achieve the same potential in healthcare as it has in other industries, the professional cultural factors affecting the use of such systems must be understood5 but these have hitherto not been widely explored. We aimed to describe the factors affecting these practices within the specialty of anaesthesia.

Methods

Approval for the project from which these data are drawn6 was granted by two local Research Ethics Committees. The study was conducted principally in one English district general hospital, with shorter periods of observation at a second English hospital. We used an ethnographic approach, grounded in detailed observation,7 followed by a series of in-depth interviews. Ethnography is a form of social research carried out in everyday settings, using a range of methods to focus on the meanings of individuals' actions and explanations rather than their quantification.8 The aim is to build up a picture of the phenomena under study which ‘makes sense’ to those who are being studied but which also allows, along with other qualitative approaches, for the inductive development of more general theories.9 The study focused mainly on the operating theatre environment, and included observation of, and interviews with, anaesthetists, operating department practitioners (ODPs), theatre and recovery nurses. Participants were recruited from the anaesthetic staff in the study hospital through a series of presentations to the anaesthetic department and theatre staff, outlining the aims of the research and inviting clinicians to take part. Regular meetings were held thereafter to inform all staff of progress of the study and to secure their continued co-operation. For comparative purposes a shorter period of observation was undertaken at a second site.

The staff participating in the study did so freely. Anaesthetists were asked well in advance of proposed observation sessions if we might observe them at work. At the time of the study, the UK research governance regulations requiring written consent from NHS staff taking part in research had not yet come into force. Patients on the operating lists were informed orally and in writing of the study and their written consent obtained.

Observation was performed without audio or video recording. Detailed contemporaneous notes were obtained during the observation period and transcribed immediately afterwards. Observations were performed principally by the research associate D.G. (a former anaesthetic nurse). C.P. (a medical sociologist) and M.M. (a sociologist of science and technology) performed two and five observations respectively ‘in tandem’ with D.G. to allow comparisons and internal validity checks on the data collection. Most of the observation took place in the operating theatre, starting when anaesthetists began their work in the anaesthetic room, although some observations were made of anaesthetists performing preoperative assessment. As the principal researcher had been an anaesthetic nurse in the department where she conducted the observations, staff were used to her presence and she was also able to position herself where she could note what was happening without being in the way. We were also able to observe two departmental audit meetings and gain access both to departmental critical incident reports and to the minutes of the remaining 10 meetings.

In our early interviews, to help participants describe how they used and acquired anaesthetic knowledge (the focus of the main study6), we asked them to recount a recent case, or on-call period. We also invited anaesthetists involved in critical incidents taking place during the study period to talk in more detail about them to our researcher. Interviews were tape-recorded and transcribed.

The qualitative analysis began with individual close readings and annotations of the observational and interview transcripts by each member of the research team. This was followed by collective discussions and comparison of the various readings of the data, from which our analytical themes and categories emerged.10 Four anaesthetists were involved in respondent validation11 of themes arising from project data. Documentary analysis of audit meeting minutes and critical incident forms was also undertaken.

Results

We have data from 19 interviews and over 130 h of observation of anaesthetic practice, gathered at the rate of approximately three sessions per month over a year. Although two of our interviews were ‘debriefing’ interviews, where anaesthetists were questioned about their work immediately after being observed, most interviews were unconnected with a specific theatre session. At the time of our study there were 13 consultant anaesthetists and 10 trainees at the principal study site. None declined to be observed, although two did not wish to be interviewed when invited. We interviewed 12 anaesthetists (7 consultants and 5 trainees), 4 ODPs and 3 nurses (2 working in the recovery room and 1 in anaesthesia). Observation included 31 operating theatre sessions of a variety of different types of surgery and a mixture of practitioners of varying experience. During these we observed 53 anaesthetics.

What is a critical incident?

Our interview data show that respondents were aware of the official definition. However, they had personal, working interpretations of this official statement. One working definition related to speed of onset and potential severity. Hence, a recovery sister described an incident where a patient, who had had an epidural injection performed by an orthopaedic surgeon, rapidly became bradycardic and hypotensive. The patient had no i.v. cannula to enable resuscitation and the recovery staff asked urgently for help from an anaesthetist in another theatre. Another factor bearing on definition, related to the feeling of control over a given situation. In one interview extract (see Supplementary data for Appendix 1), a senior house officer (SHO) talks of situations becoming critical when he is unable to cope on his own, where the patient's condition was deteriorating beyond his control. He also recounts an incident that had occurred earlier on the day of interview, where a leak had developed in the breathing system. He and his supervising consultant had rapidly detected and corrected the problem. The SHO comments: ‘We were just doing our job and being observant’, suggesting that such events are an integral part of anaesthetic practice. He goes on to say that this incident was therefore not deemed critical.

The relevance of perception of control was echoed in another of our interviews, where an SHO with 4 months' experience recounted a case from a recent weekend on call. A previously fit 36-yr-old man had been listed for an appendicectomy. During the preoperative visit, the SHO had noted that the patient was pyrexial and clammy. The patient tended to persistent tachycardia and hypotension despite i.v. fluid replacement during the operation. He regurgitated gastric contents shortly after extubation of the trachea, became cyanosed and had to be re-intubated. Arterial blood gas analysis confirmed a metabolic acidosis suggestive of systemic sepsis of some hours duration. The SHO had considered this possibility intra-operatively but had not quite put all the clues together. It was only when he felt he had lost control of the situation—specifically, being unable to manually ventilate the patient's lungs in the recovery room after extubation—that the case became ‘critical’.

We also noted that, during an audit meeting early in our study period, there was some discussion as to what incidents satisfied the requirements for reporting on the ‘clinical incident’ reporting form. This was raised after discussion of a difficult intubation where the patient's trachea had eventually been successfully intubated on the third attempt. It was felt that as this had been anticipated, it did not satisfy the criteria for reporting.

Practitioners used the term ‘critical incident’ in another sense, that is of an opportunity for learning or reflection arising from practice. A consultant obstetric anaesthetist recounted trying to first extend an existing epidural block, then perform a spinal anaesthetic, then, when neither of these proved effective, converting to a general anaesthetic for Caesarian delivery of a woman in labour. An interview with a senior ODP (see Supplementary data online for Appendix 2) alluded to the fact that anaesthetists develop expertise by discussing such cases, and also commented that this very expertise makes them more skilled at recognizing when things are going wrong. A recovery nurse brought with her to her interview four written cases she termed ‘critical incidents’. One involved a disagreement with a trainee anaesthetist who refused to consider a patient controlled analgesia (PCA) machine for an opioid addict, while another related to the temporary holding of a critically ill patient in the recovery room until a transfer to an intensive care unit in another hospital could be arranged. It appeared that what made these and the other incidents ‘critical’ for her was the perception that she was unable to influence doctors' actions, leading to a breakdown in working relationships which thus, in her view, jeopardized patient safety.

The potential for adverse events in anaesthetic practice

During our observations, we did not observe any incidents where patients were harmed. However, we witnessed a number of events which, depending on the definition of criticality, could be construed as reportable (Table 1). For instance, we observed the disconnection of the breathing circuit as the patient's head was turned during a minor nasal operation. This was immediately recognized by the anaesthetist and corrected, but seemed to be regarded almost as an expected occurrence in this context. Some of the events we have classified as ‘lack of smoothness’ may be regarded as an integral part of anaesthetic practice, but our classification ‘deviation from protocol’ is based on our interpretation of instances where national or local policies governing safe practice have been breached.

Table 1

Classification of ‘minor events’ observed

Category	Examples	Frequency
‘Lack of smoothness’ (n=25)	Cough/hiccough/chomping on inductionLow blood pressure	65
	Inappropriate patient movement during surgery	5
	Low $S{\text{p}}_{{\text{O}}_2}$ on pulse oximeter	2
	Residual neuromuscular block in recovery	2
	ET tube inserted into right main bronchus	1
	Attempted use of nerve locator in patient who had been given neuromuscular blocking agents	1
	Vaporizer turned on but no fresh gas flow	1
	Slow to regain consciousness in recovery	1
	Anaesthetic turned off too early at end of operation	1
Procedural difficulty/failure (n=18)	Venous cannulationLaryngeal mask airwaySpinal anaesthesia	652
	Arterial cannulation	2
	Bag and mask ventilation	2
	Tracheal intubation (medical student)	1
Deviation from protocol (n=23)	Re-attachment of partially used i.v. fluid bag No monitoring during some part of anaesthetic	43
	Silencing monitor alarms	3
	Confusion over consent/laterality/X-ray	3
	Inappropriate handling of sharps	2
	Anaesthetist left room briefly	2
	No i.v. antibiotic given to patient having total hip replacement	1
	Anaesthetist eating crisps in scrub area	1
	Anaesthetist declined use of fluid warmer for expected long orthopaedic operation	1
	Patient not visited before operation	1
	Infection control policy breached	1
	Local anaesthetic infusion nearly connected to i.v. cannula	1
Equipment and monitoring (n=20)	Monitor alarmed when no problem with patient	5
	ECG electrode falls off	2
	Recovery staff unfamiliar with operating infusion pump	2
	Loose connection within peripheral nerve locator	2
	CO2 monitor not functioning	1
	Ventilator tubing kinked	1
	Propofol leaking from around i.v. cannula (TIVA)	1
	TCI pump battery exhausted	1
	Air in TCI line	1
	Breathing circuit incorrectly assembled	1
	Difficulty establishing $S{\text{p}}_{{\text{O}}_2}$ trace	1
	Rebreathing CO2	1
	Shoulder surgery attachments did not fit operating table	1
Organizational/personnel (n=11)	Order of operating list changedDiabetic patient not established on insulin as prescribed	11
	Anaesthetist sedated patient without informing other staff present	1
	Premedication not given when prescribed	1
	Anaesthetist misunderstood surgeon	1
	Surgeons pulled patient down table without first informing anaesthetist	1
	Anaesthetist set propofol infusion running without informing patient or other staff	1
	Air pipeline disconnected by theatre staff after being checked by anaesthetist	1
	No anaesthetic assistant present on emergence—absent from theatre	1
	No handover to recovery staff	1
	Request to theatres to check space available in recovery before transferring patient ignored	1
Physical hazards (n=6)	Disconnection of breathing systemAnaesthetist tripped over cable	21
	TIVA infusion lines stretched	1
	Cables and i.v. lines badly tangled	1
	I.V. cannula pulled out inadvertently	1

During the study period we observed two departmental audit meetings and were also able to analyse the minutes of the remaining 10 meetings. Twenty-eight cases and events were discussed, of which 5 (18%) were presented as ‘critical incidents’ (Table 2).

Table 2

Incidents and cases discussed at audit meetings during 12 months of study. Those presented as critical incidents marked^*

Preoperative

(i) *Diabetic patient brought to theatre with insulin infusion running but no dextrose

(ii) Asystole on induction—patient undergoing laparotomy for peritonitis. Excellent preoperative resuscitation and preparation

(iii) Intravascular injection of local anaesthetic during local block for carotid endarterectomy—despite prior aspiration—successfully treated

(iv) *Leaking thiopental syringe during rapid sequence induction for emergency Caesarian section—leak in pre-prepared syringe made up in Pharmacy department

(v) Unexpected awkward intubation during rapid sequence induction for emergency laparotomy, resulting from undiagnosed laryngeal polyp—intubated with 4 mm microlaryngeal tube

(vi) Difficult intubation—eventually performed using fibreoptic laryngoscope through a laryngeal mask airway

(vii) Difficult intubation in patient for thyroidectomy—7 mm tracheal tube would not pass beyond vocal cords, though 6 tube would. Review of chest radiograph revealed tracheal deviation, though not reported by radiologist

(viii) *Diabetic patient with unstable glycaemic control for termination of pregnancy. Admitted to hospital the day before planned operation. Delay in adequate glycaemic management because of the moral objections of trainee gynaecologists to procedure

Intraoperative

(ix) Cardiac arrest after administration of i.v. protamine during aortic stenting. Patient successfully resuscitated. Consensus of meeting was that it was an exaggerated ‘normal’ response to drug

(x) *Tracheal tube severed by surgeon's osteotome during maxillary osteotomy

(xi) Bradycardia during laparoscopy in patient with known latex allergy—unrelated to allergy

(xii) Small pneumothorax, leading rapidly to hypotension, electromechanical dissociation and asystole during thoracoscopic sympathectomy. Patient resuscitated successfully

(xiii) Anaphlyaxis to i.v. antibiotic during awake Caesarian section

(xiv) Sustained low arterial blood pressure during revision of total hip replacement—patient developed postperative renal failure

(xv) Low end-tidal CO2 and heart rate 170 beats min−1 during irrigation of amputation stump abscess with hydrogen peroxide—thought to be air embolism. Successfully treated

Postoperative

(xvi) Two cases of postoperative respiratory depression on surgical wards because of opioids. One from patient-controlled analgesia, one from oral morphine. Both appeared to be a result of enhanced sensitivity rather than absolute overdose

(xvii) Bradycardia and cardiac arrest in the recovery room, thought to be attributable to myocardial ischaemia

(xviii) Postoperative fits in epileptic patient who had received propofol and alfentanil for general anaesthesia. Postoperative questioning revealed she had discontinued her usual phenytoin a few days before the procedure

(xix) Recovery staff not informed on handover that a patient's tooth was loose. Tooth fell out and was retrieved but staff felt it would have been better to have been told in advance

(xx) Horner's syndrome noted after interscalene brachial plexus block (recognized complication of procedure)

(xxi) Cardiac arrest after sedation for dental clearance. Postoperative investigation revealed a low serum potassium concentration. Patient was taking two diuretics and should have had serum electrolytes checked before procedure

(xxii) Sepsis during appendicectomy (see text)

(xxiii) Negative pressure pulmonary oedema after emergency appendicectomy—thought to have occurred when patient was biting on, but trying to breathe through, the tracheal tube before extubation

Miscellaneous

(xxiv) *Oxygen analyser on anaesthetic machine in anaesthetic room read 17% when machine delivering at least 33%. Machine withdrawn from use until repaired. No harm to patients

(xxv) ‘Cautionary tale’—sepsis masquerading as intra-abdominal bleeding—see text

(xxvi) Three cases of central venous line insertion discussed to show recommended tip position

(xxvii) Case report—use of cell saver device to reduce transfusion requirements during emergency repair of ruptured abdominal aortic aneurysm

(xxviii) Brown fluid aspirated before local anaesthetic injection during lumbar sympathectomy. Ultrasound scan revealed 9 cm wide aortic aneurysm sac. No harm resulted to patient; all patients now scanned before procedure performed

We have found only one incident form completed during the study period in the department records. This is not entirely clear, but seems to relate to difficulty estimating the oxygen content in the gas mixture in the recently installed new anaesthetic machines. This appears to relate to an incident discussed at an audit meeting (Table 2, incident xxiv).

The purposes of discussion and reporting

Our data also shed light on practitioners' perceptions of the purposes of discussing and reporting incidents. Critical incidents were perceived by our interview respondents as ‘significant moments’ in the acquisition of their professional expertise. When we asked them about how they had acquired their anaesthetic knowledge, many focused, without prompting, on cases which had not gone according to plan, which had either happened recently or during their training. They seemed to have learned much from such cases, even though there were usually only a small number of them in each individual's career. Within the audit meetings we observed, the section where cases were presented and events were discussed were usually of interest to all present, and triggered lively discussion. This alludes to the first reason for reporting, that is for continuing education within the community of anaesthetists. This may be to alert them to new problems, or simply to remind them of things which are recognized complications of anaesthetic practice, though not often seen.

Our respondents perceived that presenting and discussing adverse events could invite criticism and censure from colleagues. The distinction was made between an incident where a mistake might have been made, and one where an unpredictable chance happening takes place. In this context we report our observation of a case presentation at a departmental audit meeting. The anaesthetist involved, after commenting that it is still possible to be surprised after many years in clinical practice, described how a previously fit woman had become hypotensive after an abdominal operation. This had persisted despite i.v. fluid replacement and he had mistakenly attributed it to continuing occult intra-abdominal bleeding. Only after a further laparotomy, where no abnormality was found, did systemic sepsis suggest itself as the cause of the hypotension. This event could be defined as a critical incident using the RCA definition above but was instead presented under the title ‘A cautionary tale’. As such, it acknowledged tacitly that there had been an error of judgement but the tone of the presentation suggested that this could have happened to anyone. Discussing the case was clearly of educational benefit to the other anaesthetists present, but there is no record of a written incident report for this case.

Discussion and reporting seems also to have been complicated by the proliferation of schemes. One respondent perceived in particular that the two schemes in operation in the hospital served different purposes and this appeared to influence the definitions used (see supplementary data for Appendix 3).

Discussion

Anaesthetic practice has many potential hazards to patient safety. Our data suggest that some of these are discussed and lessons shared, though few are formally reported as critical incidents. Further, although many minor events may be reportable as ‘near misses’, or even constitute violations of official safety notices, many do not seem even to be acknowledged.

Despite the acknowledged importance of incident reporting, it is known that under-reporting is usual.12,13 This has been attributed to design of forms,14 the belief that error can be used as a measure of individual competence15 or unwillingness to report colleagues.16 Further, different schemes may be perceived as having different, sometimes conflicting purposes. The anaesthetist quoted in web Appendix 3 also alluded to this when he distinguished between ‘critical incident reporting’ (as understood by anaesthetists) and the recently introduced Trust incident reporting scheme (termed ‘clinical incident reporting’) and now linked in to the NPSA's National Reporting and Learning System, with its potential for loss of ‘control’ of the incident and its interpretation. In particular, the potential for blame, which may be more easily attributed by those from outside anaesthesia as they lack the contextual understanding of our practice, appears to influence the decision to designate a particular event as critical. This might be expected to discourage reporting through hospital-wide mechanisms, but a decline has also been noted in reporting within one local departmental scheme since the NPSA scheme was introduced.17 Thus, although discussion of events and cases persists—including the sharing of positive experiences, in line with Flanagan's original ‘critical incident technique’18—formal reporting is neglected.

Definitions of criticality within anaesthesia have largely been taken at face value in reviews of the subject.19,20 In the UK, the RCA definition is widely known and the NPSA National Reporting and Learning Scheme definition is similar.4 Other definitions vary20 but typically incorporate elements of undesirability, preventability, severity and include also the potential for harm and actual harm. Why then are potential threats to patient safety, such as the disconnection of the breathing circuit we observed, not reported? We suggest that expertise in anaesthesia brings with it not only knowledge and skill but also ‘definitional power’—the ability to ‘set’ or impose a definition of the situation as routine or critical.

We have previously argued that the ability to distinguish the wide range of ‘normal’ responses and conditions during anaesthesia from abnormal is one of the hallmarks of expert anaesthetic practice,6 and hence it is logical that it also defines the threshold for criticality. We have reported two similar cases of abdominal sepsis causing diagnostic confusion. However, the trainee's was reported as a critical incident, but the consultant's as a ‘cautionary tale’. Thus, although the official definitions of criticality are apparently clear, they are ‘operationalized’ within a specific ‘community of practice’.21 Thus in Tamuz and colleagues'22 study of hospital pharmacy personnel in the United States, 23 how errors were classified dictated how they were handled within the hospital organization. In Barach and Small's1 survey of reporting system experts, there was agreement on the importance of accepted definitions, but the problem of interpretation in practice was not addressed. Whilst, in theory at least, broad, ambiguous definitions of potential dangers might aid the discovery of risks that escape existing definitions,24 this has not been borne out by experience in other industries.

Our observational data captured many minor events which may not in themselves be significant. However, Boëlle and colleagues25 demonstrated an association between ‘undesirable’ events in anaesthesia (not in themselves harmful) and subsequent critical outcome events. If this relationship is robust, events hitherto judged too trivial to scrutinize may be worthy of closer analysis.26 We would not, however, advocate formal reporting of such events until their proper role is established, though they may be useful in internal departmental quality control.

We also witnessed anaesthetists intentionally deviating from protocol. For instance, re-connecting partially used bags of i.v. fluid is in direct contravention of a UK Medical Devices Agency Hazard Notice27 of which at least some of the anaesthetists we studied were certainly aware. The Notice was issued after two patients died from air embolism when reconnected bags were pressurized for rapid infusion. What our anaesthetists did not do was pressurize the bags. We suggest that they felt that their knowledge of the circumstances of the fatal incidents prevented them from repeating them. In doing so, they exercised their expert judgement to keep things safe but this could still be construed as a ‘violation’ of defensible practice. It is of note, however, that the arbiter of standards, the Medical Devices Agency, falls outside the anaesthetic ‘community of practice’. In this context, we note the findings of Beatty and Beatty,28 whose scenario-based study suggested that anaesthetists' intentions to violate safety guidelines are most strongly influenced by the opinion, they believe, a group of peers and significant other people would hold about the violation. We interpret this as support for our notion that expertise in anaesthesia brings with it the authority to define the boundaries between routine and critical but also between acceptable and unacceptable practice. However, we suggest that such variability in what is considered critical, reportable and acceptable is a product of the culture of medicine. In other safety-critical industries, professional experience and judgement are not allowed to dictate reporting behaviour. In aviation, for instance, all pilots, regardless of rank or experience, are expected and required to describe and report even the most subtle and minor events, not just those deemed critical or serious by individual pilots. Such attitudes to open discussion of error and potential error are intentionally fostered by the behavioural strategies of crew resource management training.29 In medicine, on the other hand, susceptibility to error is not universally acknowledged by medical staff and error is often not handled appropriately.5

The tremendous educational value of noteworthy events, both in the development of trainees' knowledge and in the maintenance of collective wariness within the profession of anaesthesia as whole, is widely recognized in anaesthesia. In an era where evidence-based practice is favoured, it is important to remember that the educational power of ‘narratives’ of individual cases. However, confusion over how official definitions should be interpreted, mistrust of newer reporting systems and the perception of loss of ‘control’ over incidents reported outside the anaesthetic ‘community’, appear to have discouraged the formal documentation of incidents. Large-scale critical incident reporting schemes are destined to fail if they do not take these professional issues into account.

Further, although the specialty of anaesthesia led the way in ‘systems thinking’ in healthcare safety through the collection and analysis of critical incidents, there appears to be room for further improvement. Within other safety-critical industries, such individual interpretation of official directives such as the re-pressurizing of i.v. fluids we observed would be unacceptable. To us, this is the paradox of anaesthetic expertise. On the one hand it brings the ability to recognize and react flexibly to (sometimes previously unknown) threats to patient safety, but on the other it also militates against the adoption of highly standardized operating procedures as in other industries. Full adoption of the ‘industrial’ paradigm of patient safety would imply an end to the exercising of such professional discretion in most routine practice conditions.

Further research in this area could usefully define the limits, if any, of transferability of such principles into anaesthesia.30 It would be useful, too, to explore more systematically, perhaps through scenario-based work, differences in classification with experience. A clearer understanding of the significance of minor events would also be of interest. Lastly, given the confusion over what is reportable in anaesthesia, it is reasonable to ask if incident schemes in ‘high-risk’ areas need to be different in some way from those in clinical areas where there may be less potential for error.

Supplementary data

The appendices can be found as supplementary data in British Journal of Anaesthesia online.

Funding: NHS North West Regional R&D Fund Project no RDO 28/3/05.