Article Text

Impact of videolaryngoscopy introduction into prehospital emergency medicine practice: a quality improvement project
Free
  1. Alistair Steel1,2,
  2. Charlotte Haldane1,3,
  3. Dan Cody1,4
  1. 1 Magpas Air Ambulance, Huntingdon, Cambridgeshire, UK
  2. 2 Department of Anaesthesia, Queen Elizabeth Hospital NHS Foundation Trust, King's Lynn, UK
  3. 3 North West Air Ambulance, Knowsley, UK
  4. 4 South East Coast Ambulance Service NHS Foundation Trust, Crawley, UK
  1. Correspondence to Dr Alistair Steel, Magpas Air Ambulance, Huntingdon, Cambridgeshire, UK; alistair.steel{at}nhs.net

Abstract

Introduction Advanced airway management is necessary in the prehospital environment and difficult airways occur more commonly in this setting. Failed intubation is closely associated with the most devastating complications of airway management. In an attempt to improve the safety and success of tracheal intubation, we implemented videolaryngoscopy (VL) as our first-line device for tracheal intubation within a UK prehospital emergency medicine (PHEM) setting.

Methods An East of England physician–paramedic PHEM team adopted VL as first line for undertaking all prehospital advanced airway management. The study period was 2016–2020. Statistical process control charts were used to assess whether use of VL altered first-pass intubation success, frequency of intubation-related hypoxia and laryngeal inlet views. A survey was used to collect the team’s views of VL introduction.

Results 919 patients underwent advanced airway management during the study period. The introduction of VL did not improve first-pass intubation success, view of laryngeal inlet or intubation-associated hypoxia. VL improved situational awareness and opportunities for training but performed poorly in some environments.

Conclusion Despite the lack of objective improvement in care, subjective improvements meant that overall PHEM clinicians wanted to retain VL within their practice.

  • airway
  • pre-hospital

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Advanced airway management is necessary in the prehospital environment and difficult airways occur more commonly in this setting.1 2 Failed intubation is closely associated with the most devastating of complications of airway management.3 4 Success rates of prehospital endotracheal intubation vary from 85% to 99%, depending on the nature of the service.5–7 Videolaryngoscopes (VL) may improve the success of tracheal intubation, particularly when the patient has a difficult airway.8 VLs have been advocated as a means of reducing the risk of viral transmission while performing tracheal intubation and are recommended in preference to direct laryngoscopy for this purpose.9

Within our service, our first-attempt success at intubation was approximately 85% and we recognised that subsequent attempts are increasingly likely to be associated with complications.10 11 In an attempt to improve the safety and success of tracheal intubation, we implemented VL as our first-line device for tracheal intubation within a UK Prehospital Emergency Medicine (PHEM) setting.

Methods

Magpas air ambulance provides prehospital emergency medical teams in support of the National Health Service ambulance service. A physician–paramedic team undertake advanced airway management with a modified rapid sequence induction technique using conventional direct laryngoscopy (DL) and a Macintosh blade sized appropriately for the patient. Both paramedic and physician members of the team are competent at undertaking advanced airway management with no one professional group favoured over another within our system. Airway management is undertaken only as a physician–paramedic team. All team members are senior clinicians that have undertaken advanced airway and prehospital emergency anaesthesia training, the nature and content of which was unchanged over the study period. The physicians within the service are predominantly from an emergency medicine background, and most estimate they have performed between 100 and 500 tracheal intubations using DL within their careers. When advanced airway management was required as part of intracardiac arrest or post-return of spontaneous circulation (ROSC) management, a standard operating procedure determined whether drugs, including neuromuscular blocking drugs, were given to assist intubation. When in-line cervical spine immobilisation was required, this was performed manually with cervical collar open or removed. Clinicians are instructed to obtain a good view of the laryngeal inlet prior to attempting intubation. Gum-elastic bougies were used as part of our standard operating procedure for all intubation attempts throughout the study period. There were no other significant changes to our standard operating procedures relating to airway management, or changes to our service characteristics, over the study period. Team configuration, experience, tasking and other items of airway equipment were unchanged.

The change in practice was considered a service improvement project and formal ethics approval was deemed not required, as VL is a well-established device with well-studied characteristics and the study would not be considered research based on Health Research Authority recommendations. Institutional approval was gained. The Magpas Clinical Governance Committee formally approved the project.

A number of VLs were assessed for suitability and the McGrath MAC EMS (Medtronic, Watford, UK) was selected as best meeting the service’s needs, as it could be used as a conventional laryngoscope for DL as well as VL with both Macintosh and hyperangulated (‘Difficult Airway’) blades. Additionally, at the time of implementation, it was the only device available that did not require charging, instead using a disposable easily replaced battery. A standard operating procedure detailed that the first attempt at tracheal intubation would be using a Macintosh blade on the McGrath VL, with a hyperangulated blade selected on second attempt if the tracheal tube could not be passed due to an anterior larynx. A second attempt using a VL Macintosh blade of the same or different size is approved if a change (eg, head position) is used to improve the chance of second-attempt success. A conventional laryngoscope was available for DL in the event of VL technical failure. The use of DL as a first-line device was discouraged rather than prohibited, as the team are supported in applying clinical judgement.

The VLs were stored on the ‘ready-to-use’ airway boards. Following use, the disposable blades were discarded and the handle and screen disinfected using antimicrobial wipes (Clinell, GAMA Helathcare, Watford, UK) and returned to use.

A training package consisting of pre-reading, an instructional video, medium-fidelity simulation and in-person assessment of competence by an experienced trainer was provided prior to the implementation and use of VL. Either member of the team, paramedic or physician, are empowered to be the intubator.

The type of laryngoscope and blade used, the laryngeal view and the number of attempts taken to secure the airway were recorded within an electronic database (HEMSBase, Medic One Systems, UK). First-pass success was defined as achieving tracheal intubation within the first insertion of the laryngoscope within the oropharynx—removal and reinsertion of the laryngoscope constituted further attempt(s).

We focused on three clinically important variables—first-attempt intubation success, intubation-associated hypoxia and view of the laryngeal inlet. Statistical process control (SPC) charts were used to identify whether the variation seen within our system was influenced by the introduction of VL into practice.12

An anonymised online survey (SurveyMonkey, London, UK) of all the active team members was conducted via email on 5 August 2018 to ascertain their views on the introduction of VL. A further verbal reminder was given at the Team’s Governance day on 26 September 2018 (Survey given as online supplementary appendix 1).

Supplemental material

Data were collected from January 2016 to February 2020. VL was introduced in October 2017.

SPC charts were used to identify whether the introduction of VL imposed a signal that there had been a change to the system, as such charts allow significant changes to a system to be identified without incorrectly attributing natural variation to a special cause. Upper and lower control limits are calculated based on the data’s mean and moving range (3 x Sigma). Conventional rules to identify special-cause signals were applied.13

Results

There were 919 patients that had advanced airway management undertaken within the data collection period, averaging 18 tracheal intubations per month.

A total of 481 (52%) of these were patients in cardiac arrest or within an ROSC period. Muscle relaxant drugs were used to assist airway management in 50% of peri/arrest patients prior to VL introduction and 57% of peri/arrest patients following VL introduction. DL was selected for intubation in 39 peri/arrest patients (13%) and VL selected in 254 (87%) peri/arrest patients.

The predominant indications for airway management were a failure to maintain an airway (44%), failure to protect an airway (25%), failure to oxygenate/ventilate (15%), cardiac arrest (9%) and anticipated clinical course (7%). The basic characteristics of the two groups are shown in table 1.

Table 1

Age, gender and mechanism of injuries of the two groups are shown

The recorded view and number of attempts for each type of device are given in table 2.

Table 2

Laryngeal inlet views (Cormack-Lehane equivalent) and number of intubation attempts according to device used

DL, not using VL, was performed on 70 occasions after October 2017, typically because of the presence of airway soiling (n=35) and / or as part of ‘crash’ airway management in cardiac arrest (n=40). In cardiac arrest patients, the first pass success using DL was 64% and using VL was 79% (χ2 test, p=0.43).

A hyperangulated VL blade was selected on five occasions throughout the evaluation period.

A pool of 15 paramedics were responsible for 294 (39%) first-intubation attempts, and 46 physicians were responsible for 465 (61%) first-intubation attempts. Paramedics gained more experience with VL than the physicians—by the end of the evaluation period individual paramedics had a mean of 22 VL uses (SD 20) and individual physicians a mean of 12 VL uses (SD 10). Overall paramedics achieved an 81.3% first-attempt intubation success and physicians achieved an 81.9% first-attempt success.

Introduction of VL did not result in a change to our first-attempt success rate. Variation is seen, with a monthly expectation, within 3-sigma, that between 50%and 100% of endotracheal intubations will occur on the first attempt. Introduction of VL did not exert a special cause and common causes of variation are determining the first-attempt success rate (figure 1).

Figure 1

SPC chart showing monthly first-attempt success rates. SPC, statistical process control; VL, videolaryngoscopy.

Introduction of VL did not exert a significant change to the occurrence of intubation-associated hypoxic episodes, defined as a reduction in SPO2 of ≥10% from preinduction values over two consecutive measurements, with a monthly expectation of a hypoxia-free intubation of between 82% and 100% (figure 2).

Figure 2

SPC chart showing monthly rates of intubation without hypoxia >10% of preinduction SPO2. SPC, statistical process control; VL, videolaryngoscopy.

There was no significant change in the likelihood of a ‘good view’ of the laryngeal inlet, defined as a Cormack-Lehane view of 1 or 2, or the VL equivalent.14 Common cause variation is determining the view obtained and introduction of VL did not significantly influence this (figure 3).

Figure 3

SPC chart showing monthly rates of ‘good’ laryngeal inlet views obtained during intubation. SPC, statistical process control; VL, videolaryngoscopy.

Those patients predicted to have difficult airways by external assessment were intubated at first attempt 36% using DL and 31% using VL (χ2 test, p=0.55).15

Sixty-one different team members intubated over the study period, 15 paramedics and 46 physicians. At the time of the survey, there were 33 active team members. Twenty (61%) of the team’s paramedics and physicians responded to the survey. As a general email discussion among the team occurred in response to the initial call for survey completion, a second request for completion was deemed unnecessary.

When asked, ‘What do you like most about the introduction of videolaryngoscopy?’, dominant themes relating to situational awareness and glottic view (box 1).

Box 1

Responses to the question ‘What do you like most about the introduction of videolaryngoscopy’ grouped by theme.

Situational awareness and teamwork

  • ‘Improved situational awareness. It makes laryngoscopy a team effort.’

  • ’As the assistant it increases your situational awareness—this allows for the assistant to plan ahead regarding what may be needed. Useful when DL view is challenging’

  • ‘The ability for the assistant±supervisor to share the view on the screen.’

  • ‘Good views. Team member can see what you see.’

  • ‘The ability for both operator and assistant to see the view.’

  • ‘The situational awareness that it brings to the team.’

  • ‘Shared image of the laryngeal inlet/vocal cords. Aids the assistant in laryngeal manipulation.’

  • ‘I like that you have two-person view of the anatomy being seen.’

Improved glottic view

  • ‘It appeared useful whilstwhile intubating an obese patient.’

  • ‘Where a DL view is suboptimal but a sufficient videolaryngoscopy (VL) view is present to allow 1first pass, for example, cardiac arrests.’

  • ’As an assistant, where the operator has had a poor view, I was able to move the larynx into a better position without any verbal comms required due to the fact that I could see the screen.’

  • ‘Normally get a good view on the screen, which is good for teaching (especially ambulance staff).’

  • ‘The VL is gold standard in terms of ensuring safe ET insertion and I support anything that will make this process safer.’

Poor brightness/insufficient contrast in daylight

  • Poorer illumination than direct laryngoscopy, which may lead to a poorer view in high ambient light.’

  • ‘Sometimes difficult to pass the endotracheal tube when you have a good indirect view. Difficult to see in bright sunlight.’

  • ‘Screen sometimes a challenge in outside environment.’

  • ‘The light is dim. The screen is not good in bright light.’

  • ‘Not so useful in bright environments.’

  • ‘Difficult to use in bright light—unable to use screen even when not direct sunlight if it’s a bright day.’

  • ‘I feel the light is not bright enough and it is difficult to see in outdoor environments.’

Difficulty instrumenting trachea

  • ‘I have found it very difficult to get a direct view with the videolaryngoscopy (VL). I feel that it has also made passing the tube harder. I don't feel that the VL has let me down but I believe it has made most of the intubations harder and added an extra element of stress to the procedure.’

  • ‘Despite obtaining a view more easily, the angle makes intubation with a bougie/without a stylet more challenging.’

  • ‘Can have a good view on the screen but still find it difficult to pass tube due to the angle.’

  • ‘Often need to use a stylet instead—this often leads to multiple attempts even in patients that are predicted to be easy intubations and has on occasion lead to more stressful situations than were necessary.’

Other limitations

  • ‘The ‘feel’ is slightly different, so a good view using direct laryngoscopy is less easily achieved.’

  • ‘If the airway is full of blood or vomit, it makes VL impossible to use as the video image is blocked.’

  • ‘There have been a number of occasions when the assistant has been distracted by the VL and helping the intubator get a view instead of monitoring the oxygen saturations of the patient.’

  • ‘Distraction of assistant looking at the screen puts all the focus on this rather than observations.’

When asked, ‘What do you dislike most about the introduction of videolaryngoscopy?’, the dominant themes included its poor performance in bright ambient light and difficulty in instrumenting the trachea despite a good glottic view (box 2).

Box 2

Responses to the question ‘What do you dislike most about the introduction of videolaryngoscopy’ grouped by theme.

Situational awareness and teamwork

  • ‘Improved situational awareness. It makes laryngoscopy a team effort.’

  • ’As the assistant it increases your situational awareness—this allows for the assistant to plan ahead regarding what may be needed. Useful when DL view is challenging’

  • ‘The ability for the assistant±supervisor to share the view on the screen.’

  • ‘Good views. Team member can see what you see.’

  • ‘The ability for both operator and assistant to see the view.’

  • ‘The situational awareness that it brings to the team.’

  • ‘Shared image of the laryngeal inlet/vocal cords. Aids the assistant in laryngeal manipulation.’

  • ‘I like that you have two-person view of the anatomy being seen.’

Improved glottic view

  • ‘It appeared useful whilstwhile intubating an obese patient.’

  • ‘Where a DL view is suboptimal but a sufficient videolaryngoscopy (VL) view is present to allow 1first pass, for example, cardiac arrests.’

  • ’As an assistant, where the operator has had a poor view, I was able to move the larynx into a better position without any verbal comms required due to the fact that I could see the screen.’

  • ‘Normally get a good view on the screen, which is good for teaching (especially ambulance staff).’

  • ‘The VL is gold standard in terms of ensuring safe ET insertion and I support anything that will make this process safer.’

Poor brightness/insufficient contrast in daylight

  • Poorer illumination than direct laryngoscopy, which may lead to a poorer view in high ambient light.’

  • ‘Sometimes difficult to pass the endotracheal tube when you have a good indirect view. Difficult to see in bright sunlight.’

  • ‘Screen sometimes a challenge in outside environment.’

  • ‘The light is dim. The screen is not good in bright light.’

  • ‘Not so useful in bright environments.’

  • ‘Difficult to use in bright light—unable to use screen even when not direct sunlight if it’s a bright day.’

  • ‘I feel the light is not bright enough and it is difficult to see in outdoor environments.’

Difficulty instrumenting trachea

  • ‘I have found it very difficult to get a direct view with the videolaryngoscopy (VL). I feel that it has also made passing the tube harder. I don't feel that the VL has let me down but I believe it has made most of the intubations harder and added an extra element of stress to the procedure.’

  • ‘Despite obtaining a view more easily, the angle makes intubation with a bougie/without a stylet more challenging.’

  • ‘Can have a good view on the screen but still find it difficult to pass tube due to the angle.’

  • ‘Often need to use a stylet instead—this often leads to multiple attempts even in patients that are predicted to be easy intubations and has on occasion lead to more stressful situations than were necessary.’

Other limitations

  • ‘The ‘feel’ is slightly different, so a good view using direct laryngoscopy is less easily achieved.’

  • ‘If the airway is full of blood or vomit, it makes VL impossible to use as the video image is blocked.’

  • ‘There have been a number of occasions when the assistant has been distracted by the VL and helping the intubator get a view instead of monitoring the oxygen saturations of the patient.’

  • ‘Distraction of assistant looking at the screen puts all the focus on this rather than observations.’

When asked whether to keep VL within the service, 95% respondents stated they wanted to keep VL in use.

There were no complications recorded directly attributable to the use of VL. There were no technical failures of the VL.

Discussion

Use of VL is well established within hospital practice, though remains predominantly used as a second-line device if a difficult airway is encountered.16 It is known that airway management in emergency environments is associated with a greater difficulty in airway management and an increased likelihood of a failure to intubate.4 Difficult airway management in the prehospital environment is in the order of 2%–5% of encounters.2 17 18 To our knowledge, this is the first paper to describe the impact of introducing VL into an established UK PHEM service.

Systematic reviews of the use of VL in prehospital settings have concluded that prehospital intubation is ‘worsened’ by the use of VL when performed by experienced operators.19 20 A major weakness of these analyses is the heterogeneity of the types of VL used. In particular, most of the included studies used either a channelled blade or a hyperangulated blade. The technique for using these blade types is far removed from that used by a VL with a conventional Macintosh blade, the latter of which allows the use of the more familiar DL insertion technique and therefore require less familiarisation and training. The VL chosen should be selected according to the indication—for example, a hyperangulated blade is best suited to intubating patients with a difficult airway due to an anterior larynx and using a hyperangulated ‘difficult airway blade’ on a patient with ‘normal anatomy’ will likely make what would have otherwise been easy airway management, more difficult.21 We selected a VL type that on first use would be inserted as for DL using a conventional Macintosh blade for these reasons.

The service agreed to adopt the routine use of VL given the evidence that emergency airway management is challenging and that the routine use of VL is likely to improve the skills required when a difficult airway is encountered.22 Furthermore, VL requires less force and repositioning, which may lead to less cervical spine motion during intubation an advantage of some significance in trauma patients.23 24 We selected a VL that could be used for DL, as we believed this would minimise the training burden from introducing new equipment.

Three metrics were selected—first-pass success at intubation, view of laryngeal inlet and peri-intubation hypoxia. These were selected because they are objectively documented, within our control and clinically important.

Critically ill and injured patients are vulnerable to secondary injury, particularly those with traumatic brain injury. Repeated attempts at tracheal intubation are associated with increased complications.25 26 The introduction of VL into our service aimed to improve our first-attempt intubation success and to reduce the risk of complications associated with advanced airway management. Glottic view was chosen as it is a key step in process of intubating the trachea, with a good glottic view being associated with successful intubation. Hypoxia was selected as a balancing measure. These data are routinely recorded within our system.

Our data did not support our expectation that VL would increase our rate of first-attempt success, reduce intubation-associated hypoxia or consistently provide the team with improved views of the laryngeal inlet. Given that no other significant changes were made to the system during the study period, we believe that the implementation of VL had no overall benefit on our primary outcome measures.

We believe, based on the feedback obtained, that this is because although a good view of the laryngeal inlet was generally easily obtained, difficulty was experienced in passing the bougie through the vocal cords. The image of the laryngeal inlet is intuitively centred in the middle of the VL screen, which leaves the alignment of the airway axes unoptimised for instrumenting the airway—having the inlet positioned at the top of the VL screen gives better alignment and allows for easier instrumentation. We hypothesise that the easily obtained laryngeal inlet view on the VL screen reduced the efforts made to lift the pharyngeal structures anteriorly, resulting in a larynx remaining relatively anterior and the airway axes unaligned.27 This, in turn, may result in the increased difficulty at passing the bougie and tracheal tube, accounting for the lack of improvement in first-pass success.

Given this relatively anterior nature of the laryngeal inlet on using VL, the bougie was not always able to hold the desired acute angle and we experienced an increase in the use of stylets in order to facilitate the passage of the tracheal tube. We later changed from gum-elastic bougies to a more rigid variety which held the desired shape for longer and additionally resulted in a significant cost saving (£40 vs £5).

The McGrath VL cost approximately £2000. The disposable Macintosh-curved blades are approximately £5, and the hyperangulated blades approximately £25. In comparison, the DL disposable Macintosh blades are approximately £5.

The main criticisms of VL were its poor performance, particularly in bright, even cloudy, conditions. This is in part due to the VL screen but also the light source itself was notably less bright than the DL equivalent. Additionally, teammates state that VL ‘felt’ different, citing a perceived difference in length or shape of blade, despite the two being near-identical.

VL is recognised to be a useful tool for training and supervision.23 28 Despite the absence of improvement in first pass-success, the team were keen to maintain the use of VL within the system, with improved overall situational awareness and improved opportunities for training and supervision frequently cited. However, VL introduced an unanticipated source of distraction, for example, at times both teammates have complete focus on the VL video screen at the same time and the awareness of the patient’s oxygen saturations was momentarily lost.

While our standard operating procedure was to select VL as our first line device, the team were supported in their choice of DL if they felt this was clinically advantageous. The team described favouring a brighter device in ‘crash airway’ scenarios and one that was perceived superior in the context of airway soiling. Some team members described having greater confidence using DL and described being inclined to reach for devices that they were more familiar with at times of significant stress. For these reasons, the use of DL was most noticeably seen in the management of periarrest patients.

There was discussion with the team about what role VL should have within the given our findings, for example, whether VL remain a first-line device or just be used in the event that of a difficult airway encountered on DL. The consensus was that in order to maximise familiarity, confidence and technical success at managing difficult airways that the team should use VL first-line for all airway management. DL would remain in service to provide redundancy within the system. No changes to the standard operating procedures were made following the evaluation.

Given the complexity of factors involved in airway management, we used SPC charts to assess whether the introduction of VL influenced our system. SPCs charts demonstrate whether variation seen within a system can be attributed to special causes (in this instance, the introduction of VL) or whether instead it was due to a number of common causes without a single dominant cause.13 When a process displays predictable variation (well within the calculated upper and lower process control limits), the variation can be thought of as a result of many different cause-and-effect relationships where no one cause is dominant over the others (‘common cause variation’) and conversely when values fall outside control limits the variation must be thought of as consisting of the sum of the common cause variation plus some additional cause-and-effect relationship (‘special cause’).29 The control limits are calculated statistically, and a number of simple rules can alert the organisation to a change within the system which is statistically unusual.30

One of the advantages of SPC charts is that they can help reduce over-reacting to data points that may appear significant (ie, attributable to a special cause, such as introduction of VL) but are likely due to common cause variation (all the elements that make up the system). For example, following the introduction of VL there were a few months during which the first-pass success rate reduced to below 60%—while this would appear a worrying development and one might be attribute to a specific change in the system, the SPC tells us that these numbers should be expected. Without further changes to the system, next month’s first-pass success can be expected to be anywhere between 50% and 100% (these are the statistical lower and upper control limits), and continue this way every month until the system changes.

The introduction of VL did not create a special-cause signal within the SPC chart. This suggests that the type of laryngoscope used is not a dominant factor in the success of first pass attempts at intubation nor a means of reducing intubation-associated hypoxia. Common-cause variation is demonstrated, meaning there are other factors aside from which laryngoscope is used that is determining the first-attempt success and intubation-associated hypoxia. This knowledge allows us to focus on other aspects of intubating in order to improve success rates, such as technical and team skills.

A limitation of this study is the lack of control for team members rotating through the service. For example, every six months new physicians join the service and while the criteria for joining remained the same over the study period we were unable to control for the experience newly joining team members had. In particular, the base specialty of the PHEM physician is predominantly emergency medicine while one might expect an anaesthetist to have significantly more experience of using both DL and VL, which could conceivably influence our outcomes.

It is possible that a different make or model of VL may lead to improved performance, eg some makes may provide brighter glottic illumination or video screens) that perform better in sunlight, though these elements were not anticipated and so not sufficiently tested prior to selecting our preferred VL.

We believe that our experience is generalisable to other PHEM organisations within the UK, as the models of care and service provision are increasingly aligned in order to provide national training programmes that meet GMC criteria for PHEM specialty training (www.ibtphem.org.uk).

Conclusion

The introduction of VL in to PHEM practice did not reduce variation of first-pass intubation success, view of laryngeal inlet or intubation-associated hypoxia. VL was perceived to improved situational awareness and opportunities for training but performed poorly in some environments. Overall PHEM clinicians wanted to retain VL as a first-line device for advanced airway management within their practice.

Ethics statements

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • Handling editor Caroline Leech

  • Twitter @alistairsteel

  • Contributors All authors assisted with the project and in preparation of manuscript.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.