‘Don’t compress the chest in traumatic arrest…’ That’s the narrative. But Alan Garner has questions.
Do you do chest compressions in traumatic cardiac arrest (TCA)?
Don’t be dopey, right? Compressions are not important compared with seeking and correcting reversible causes. Indeed you can just omit the compressions altogether and transport the patient without them as they are detrimental in hypovolaemia and obstructive causes of arrest, right?
I would like to work through the logic of this. I think the nidus of an idea got dropped into a super saturated FOAMEd solution and Milton the Monster* precipitated out. The end result might be an approach that got extrapolated way beyond the biologically plausible.
The Starting Point
First let’s try to step slowly through the logic…
In hypovolaemia or obstructive causes of shock that are likely in the trauma patient (tension and tamponade) and where the patient is in PEA (preferably with good cardiac motion on ultrasound) then they might still have output but below the limits of detection except by doppler – certainly not by feeling pulses.
In these patients RV filling is critically impaired and chest compressions during diastole can further impair filling.
This converts the low output state into a no output state, and this has to be bad.
(There is an underlying assumption I think we’d have to agree with here – that a no output state is worse than a low output state. Pretty fair call.)
This all seems to bring you to ‘don’t bother with chest compressions at all as they are detrimental in all TCA patients’. Or, ‘at least don’t do them till you have done everything else and only if you have some spare hands available’.
But I’ve certainly heard those at the extreme end of the spectrum who don’t do compressions at any point and transport the patient without them even if plenty of hands and a LUCAS or two are available.
Let’s get back to each step along the way. The first point was that the patient in PEA (who has an organised rhythm and probably decent contractility on ultrasound) is producing subclinically detectable flow which we do not want to mess with. This is great if the patient is that exact patient with PEA and a sniff of reasonable contractility (this and pupillary reflexes are the two things you really want to see in an arrested trauma patient). No argument from me here. But what if the patient starts in what can only be a no output state – like asystole or VF?
Should we not be trying to convert these clearly non-perfusing, no output rhythms into at least a low output state by performing CPR? My assumption here is that it should work like conventional CPR – stop for as long as you need to perform critical interventions, but absolutely minimise hands off time otherwise.
We have already agreed that a low output state is better than a no output state and indeed this is the nidus of the argument for withholding compressions in TCA. But we are also being asked to assume that a low output state is the most likely situation the patient has ended up in. This is the justification for withholding all compressions, on the basis that PEA might be present, and that there is a theoretical chance that RV filling might be impaired if a compression coincides with diastole.
Again let me emphasise that I am not saying that compressions should be performed at the expense of treating reversible causes like hypoxia or tension pneumothorax. These things absolutely take priority – just with minimal interruption to compressions.
What I am seeing however is intubated patients with bilateral thoracostomies and no tamponade on ultrasound who are in asystole not having CPR performed. No cerebral or cardiac perfusion is even biologically plausible without compressions in these patients who do not have a perfusing rhythm.
To the Library
So for the assumptions that let you say ‘compressions are a waste of time’, you’d need to be pretty sure that PEA is overwhelmingly the most likely thing you’d see.
But it’s not.
The most common rhythm in TCA appears to be asystole and it amounts to over two thirds of patients. These patients are not in a low output state with critical RV filling issues. They are in a no output state. There will be no output unless you do something to encourage it.
Another 7% are in VF. This would be another rhythm where chest compressions are indicated. So that means in total about 75% of TCA patients will have a rhythm where compressions may help. So does omitting compressions for all, to address a ‘sometimes’ thing in the 1 in 4 range, seem like a sensible balancing of probabilities?
It seems more sensible to keep compressions as the default in TCA given that most patients are in a “definitely no output” rhythm. If you find PEA then it’s your call I guess. Personally I will still be doing compressions particularly once hypoxia, tension pneumothorax and tamponade have been excluded. Using an approach like the HOTTT drill this can generally be achieved within a couple of minutes of patient contact and that should hopefully keep ‘no flow time’ to an absolute minimum.
What about studies were compressions are delivered?
Well this recent study noted rising EtCO2 levels with chest compressions in TCA, which suggests some increase in flow. A recent Japanese study of 893 blunt trauma arrests who received closed chest CPR found a 28 day survival rate of 6%. In the context of TCA in blunt trauma that’s a pretty decent number.
There’s also a study from Germany where compressions are the routine which combines data from the national cardiac arrest registry and trauma registries. They looked at ROSC rates in traumatic arrest patients and the outcomes. They found high rates of ROSC with more than a quarter having spontaneous output at hospital admission, though only 7% survived to hospital discharge. CPR was the default and survival rates comparable to the best reported in the literature. It’s not nothing.
So it seems to me we can make the cognitive load of TCA patient management about the same as other forms of arrest. It can sort of just be normal ALS with a big focus on the processes to reverse the reversible, supported by something like the HOTTT drill.
Am I missing something?
* It was an editorial decision to leave the reference to Milton the Monster there because sometimes you just have to let people show their age.
The image of that leaping reptile came from excellent sharing site unsplash.com and was posted by Denny Luan.
OK it’s been a really long time. But it’s here. Dr Alan Garner returns with the wrap from AirMed Day 2, and it is absolutely not hot on the heels of the wrap from day 1. It’s good though.
Unfortunately day 2 of Airmed was notable for a complete absence of Russian designed helicopters. Fortunately there was enough of interest to keep me going. Comments for day 2 have also been supplemented by notes taken by my colleagues Captain Greg Ohlsson and Dr Toby Fogg which helped me with the concurrency conundrum so thanks to both.
Day 2 kicked off bright and early with Klaus Egger from ÖAMTC in Austria with a Safety Analysis of current HEMS accident trends in Europe. He noted that accident databases are very poor and it was difficult for him to quantify exact numbers. However he was able to deduce that the rate of accidents in Euro HEMS is static in real terms, though the fleet has increased by 25-30% so in terms of rate, it can be seen as a reducing trend. This supports the notion of upgrading the fleet to modern IFR capable, glass cockpit twins.
Far and away the most common two accidents were wire strike and Controlled/Uncontrolled Flight into terrain or obstacles (CFIT). There were a few other causes. He railed against the intrusion of voice activated terrain and wire database warnings (EGPWS/HTAWS) pointing out that they are fitted everywhere and still had not changed the accident rate. Indeed they provide significant distraction in the cockpit comms for HEMS crews.
He noted that in the CFIT incidents the collision was almost invariably with terrain that the crew already knew was there. For example an aircraft brushed its main rotor tips against a rock wall during a winch. The crew knew the cliff face was there but drifted just enough to strike it due to distraction. HTAWS provides no additional situational awareness in that situation and actively distracts the crew – which seems to be the common thread in these incidents.
The technology is just not helping us here. It was sobering to note the fatal crash of the Irish Coast Guard helicopter in the last 18 months involved the aircraft striking an island in poor visibility that was not in their mapping system. You could argue in this case that reliance on the technology set up the crash.
Also notable was the fact that not a single accident was caused by engine failure. Huge amounts of money and regulatory effort have gone into mitigating this risk over the last 20 years. It might be time to invest more in the risks that are actually killing crews. Having said that the accident rate is trending down – which shows that modern aircraft are a better safety proposition overall.
Next up was David Lockey from London HEMS on where are we heading in Prehospital Trauma Care. I would describe this as the usual line up of suspects; REBOA, ECMO, POC testing, imaging etc. To echo a point I have made on the Collective previously he concluded by asking whether we should do these things prehospital just because we can, as it seems we can do a lot.
After morning tea the concurrent sessions kicked off. I went off to hear Lionel Lamhaut describe the Paris experience with prehospital ECMO. The logistics of this are not inconsequential. It takes a 6 person team; three to provide high quality CPR while ECMO is established and 3 to do the plumbing/EMCO bit. 7.5% of the time it fails for technical reasons. He also noted that they have moved into REBOA as well and have done their first zone one case, which has since been reported in Resuscitation.
I then had to run to another session to watch our own Captain Greg Ohlsson speaking on our ten years of NVG experience in Australia. Although some European operators such as REGA have been using the technology for much longer than we have CareFlight seems to have developed our own unique approach.
Greg outlined the three components of that; use of lots of white light as we approach the landing zone (we have installed additional lights all around our aircraft), a long slow approach, and what he terms “eye relief”. In this last point he noted that unaided hovering cues are superior as depth perception works, good cues need less cognitive processing and we can light the unaided look around and produce good unaided hover cues (by point one – lots of white light). Eye relief here means setting the goggles further away from the eyes so it is easier to look around them rather than through them.
The end result is that the crews look around the goggles more in the landing zone, depth perception is improved and less cognitive load is imposed on the crew. I doubt I have explained this well but please contact us in the comments section below and we will put you in contact with Greg if you have any questions.
Then it was another run between rooms to hear Herbert Schöchl from Austria on point of care testing, particularly around coagulation. He noted that there were POC devices already on the market that can deliver an INR very rapidly but they were not very useful in trauma coagulopathy as it is clot strength rather than time to clot per se that is the issue.
He then described some testing they had been doing on the TEG 6S that will give a measure of clot strength in 2mins. He sort of implied that it is small and light enough for prehospital (or at least interfacility transport) use, though local types who have seen it in action tell me it’s still a bit of a beast.
They had been testing the effect of vibration on the measurement as they were worried this would produce unacceptable artefact or disrupt the reading, but this turned out not to be an issue. You will understand why they were worried about this when you look at this picture – the device measures clot strength by measuring resonance frequency of the meniscus. It at least looks like we’re getting closer to a more mobile TEG or ROTEM system.
Dispatched to Dispatch
After morning team it was off to a workshop (in my favourite venue behind the bar) on HEMS dispatch.
There were two speakers here that grabbed my interest. The first was Kevin Hutton from the US who gave a very interesting perspective on appropriate utilisation from a cost perspective. Large bills for HEMS transport have been making the news a lot in the US recently with charges to individual patients of up to USD50,000! The reason is the reimbursement system. The HEMS companies cannot recover any funds from about 80% of the patients transported. So they have to recover the costs of these 80% from the other 20% and when you understand this the charges to the insured few start to make sense. Just part of the madness of the system that is healthcare in the US.
The other presented was from the East of England Ambulance Service. This was a system I had previously worked in and it has moved on a lot from when I was there eight years ago. I thought their immediate dispatch criteria for HEMS were spot on. They have been refining these for many years and Sydney could certainly learn from this.
Field Notes from Toby Fogg
While I was in the back of the bar my colleague Toby Fogg attended a session on medical competencies in HEMS – an area in which he has some interest. Here are his notes, faithfully reproduced for the reader:
“Competencies in the HEMS World – by Akos Soti from Hungarian Air Ambulance.”
He carried out a survey of self-reported competency using this scale:
Competent – proper knowledge/training, not a routine intervention, would/can do if necessary
Partly competent – some knowledge/information but not properly trained; would try as a last resort.
Non-Competent – no/minor knowledge of the procedure; would not perform.
LPR service – 88% consultants, 84% male, 80% age 30-50. 61% anaesthetics, 28% ED – very different from ours in Sydney. Occasional neurologist or trauma surgeon. Their self-reported competence is 1 or 2 on the above scale.
Surgical airway 93% felt competent, 68% have done one.
Chest decompression, 97% competent,
USS 79% for vascular access, 79% for chest, 63% for RUSH and FAST
Thoracotomy 40% , Resuscitative hysterotomy 30% competent and 7% had done one.
HUET 44% competent.
Then Matthais Ruppert from ADAC was up: They carry out 50,000 missions per year and are daytime only. Only 40-60% of patients are transported with 10% interhospital transfers. Drs are on duty 2-4 days per month, much like Sydney and the service is mostly consultants so little turnover. Training involves simulation e.g. inflight deterioration and conflict resolution in the interhospital setting.
Jens Stubager Knusden, an Intensivist from Denmark ACRM instructor presented on the Danish model. Shift patterns: Doctors are on duty 3 days straight, pilot/ACM do one week straight, living on base as a team. The pilots trained to be medical assistants. There is a strong emphasis on Medical/Scenario Debriefing and then ACRM debriefing.
The final speaker was Andreas Kruger from Norwegian Air Ambulance. His theme was that the right skills and right tools are required to deliver excellence. They have a highly efficient tasking system – the medical dispatchers are able to view the telemetry from the road ambulance in order to aid decisions. He also talked of some of his research, looking at physiology mapping for the patient.
“The fact that many medical experts point to a variable, does not make it a good quality indicator – one needs to be able to get a meaningful and useable number. The variable must have values that have a clear ordering from not-so-good to very-good. It should have the potential for sufficient variation.”
He also discussed system wide performance mapping and quality indicators such as the airway registry for which he has published a series of Utstein-like Criteria.
Enough of the Field Notes
At the end of these concurrent sessions it was time for lunch. This was the first time that I recall seeing my colleague Dr Chris Cheeseman. This was odd as the Polish doctors I last saw him with had all been there for the first session at 8am.
I unfortunately had to catch a train for Berlin so missed the final clinical session. The news at the closing ceremony was that Airmed 2021 will be held in Salzburg. The Sound of Music Airmed. Can’t wait.
If you’re into quality indicators, you might like this paper:
There are whole shows set up for glorified travel diaries. Why not have Dr Alan Garner do the same? Except with medical bits also, because that’s what the site is for.
In October 2017 Airmed was due to take place in London. It all fell through though. Probably not because of Brexit. Maybe. There were arrangements with Helitech that fell through and … actually let’s forget that bit. The good news was that 400 delegates from 35 countries turned up to Warsaw on the 13th and 14th of June for a fresh running and a reminders that friends and colleagues all over deal with many of the same issues we do. A few different ones too mind you.
Why just turn up for the meeting though? We arranged a visit with the local air ambulance operator Lotnicze Pogotowie Ratunkowe the day prior. The local system has come a very long way in a relatively short time. It was only 7 years ago that they said goodbye to their last Mi-2. I am going to say I have a strange and irrational love of Russian designed helicopters so seeing this was a real highlight for me.
They now have a fleet of >20 EC135s with Aerolite medical interiors, their own simulator to train their > 100 pilots as well as operating a couple of Piaggio fixed wings for longer distance transports. Times change.
One of their great challenges is coordination and tasking.The entire country is managed from the base we visited near Warsaw for interfaculty transports. Prehospital dispatch is done by one of 40 local dispatch centres around the country. There is a huge amount of variability in the prehospital dispatch accuracy however.
We were told that to become a dispatcher you simply required 5 years prehospital experience as a paramedic. There is a dispatch course but it is not necessary to complete this until you have been doing the dispatch job for three years. That is a lot of time dispatching.
We were fortunate enough to have a presentation by one of their senior pilots where he described the weaknesses of the system. I quote directly from his slide to mention two of them:
“Strange, incomprehensible fears
Unjustified prohibitions of helicopter summon, issued by superiors”.
Sometimes when you start in another language and translate to English you get a slightly different take on an issue that turned out to be very familiar.
A very impressive part of what we saw turned out to be the strides they are taking to standardise their medical approach. This is being led by Tomasz Derkowski, their medical director. Tomasz has previously worked with LifeFlight in Queensland. It is quite a small world it turns out.
There has been a lot of work put into standardising their equipment across the country and introducing checklists for things like intubation. On the governance side there appears to be a much bigger issue as medical governance systems are not privileged by legislation in Poland.
Apparently there are moves to change this but I don’t think real advances can be made till this is in place as it is otherwise really hard to build an open culture.
They have recently (as in, this past week) introduced ultrasound to their system and have a lot of other new things planned but introducing things takes time when you have more than 20 bases to consider.
Day One Gets Going
After a brief bit of time with the Polish Minister for Health, Lukasz Szumowski, the clinical sessions kicked off.
The first of these commenced with the account of the rescue of a 30-year-old kayaker from a really cold lake in Sweden where the water temperature was 2.8 degrees. The story was remarkable for the cross border and interagency cooperation required to effect his rescue and it was told by the three members of the Norwegian helicopter crew involved.
“Norwegians??” you might be thinking. Well the closest Swedish helicopter to the scene did not a have a rescue swimmer available so a Norwegian chopper was also dispatched with the information passing through rescue coordination centres in Sweden and then Norway.
The Swedish helicopter crew located the man in the water floating face down and directed the Norwegian team to effect the rescue by long line from where they lifted him to a small clearing in a forest. The crewman showed amazing situational awareness having determined that the helicopter could not land due to the trees. So he dragged the patient 10-20 metres along the ground to a more open area (still attached to the line) so that the helicopter could land.
The doctor then went to work with the crewman and pilot to resuscitate the patient. He was asystolic with oesophageal temperature of 20 degrees Celsius. A LUCAS device was applied and the patient was intubated. In a great demonstration of the cross-skilling that occurs in these small integrated teams the pilot assisted the doctor in the intubation by performing external laryngeal manipulation.
By this time that Norwegian helicopter did not have enough fuel to transport the patient to the nearest ECMO centre which was back in Norway where it had originated from. The patient was therefore driven by a Swedish ambulance 2 kms to the landing site of the Swedish helicopter where the patient was loaded with the Norwegian doctor and transported. He subsequently made a full recovery, presumably to tell the tale of how impressive the team work within the helicopter team, between the helicopter teams and between the rescue coordination systems of two countries was. Once he is told the story.
There was a whole stream devoted to hypothermia in the afternoon which I did not attend, as this is a big issue for the Northern Europeans who seem to have very well developed systems for rewarming on EMCO. This patient was certainly the ideal candidate as he was young and healthy and had cooled slowly whilst hanging on to the kayak before eventually losing consciousness. The crew were aware of these circumstances and continued aggressive resuscitation over an extended period to get this result.
Then Wolfgang Voelckel from Vienna was up. He spoke on professional networking and mentioned some new data from clinical trials he has conducted on prehospital fibrinogen. More on this later. The session closed with Erik Norman from Norway speaking on improved medical care through aviation. The point that stuck was that aviation had made enormous gains in avionics and autopilot systems. But the regulations are the same as they were 30 years ago in terms of visibility and minima. Perhaps it is time for a change given some aircraft now have autohover systems certified down to 3ft from the ground!
There were a couple of highlights here. The first was a talk by Jostein Hagemo and Even Wøllo from Norway looking at the medical workspace that is a helicopter. They have been keen to apply industrial design principles to improve medical care when airborne. They noted that for the helicopter there is a master alarm when things go wrong. In the back seat though there are multiple different pieces of equipment (ventilator, monitor, syringe drivers etc etc) each of which has its own alarms and nothing is integrated. Perhaps the only way to solve this is to have single device that does everything the patient needs. This seems unlikely for the moment.
They also did a bit of brainstorming about the stroke helicopter of the future…
Hmmm… well the word ‘brainstorm’ doesn’t tell you if it’s a good one, just that it happened.
We then heard from Jaap Hatenboer from the Netherlands on disruptive innovation, particularly around the pilotless aircraft concept. They are setting up a system to transport drugs by drone out to islands off the coasts from the Netherlands. He also mentioned the Zipline system that is being used in Rwanda to transport blood products up to 100km to smaller hospitals. This technology is certainly gaining ground. We have looked at this for our Northern Territory operations. The problem for us is 100km is a very short distance in the NT. We would need something that could fly 1000km round trip for it to start to be useful and those machines don’t yet exist – for civilians at least.
The Post-Lunch Conundrum
After lunch the concurrent sessions commenced and with this the concurrence conundrum of which stream to attend. I went for the ‘Violence in HEMS’ session which was strategically run in the back of the bar in the hotel (so I felt immediately comfortable).
There were no real answers here, just more conundrums. Anne Weaver from London HEMS spoke about the spectrum of violent trauma now seen by their service. One third of Royal London Hospital trauma patients are now penetrating (which according to Donald Trump could be reduced if more Londoners carried guns). The figure that surprised me was the number of corrosive liquid attacks now occurring in London being more than 400 a year. This causes significant disfigurement and appears to be on the rise as means of inter gang violence with perpetrators often quickly escaping on motor cycles. Not a trend I would hope to see Sydney follow.
Pål Nesfossen gave an overview of the attacks in Oslo from 2011 involving first of all the bomb near the parliament building as a decoy followed by the shootings on the island at a youth camp. He particularly mentioned the difficulties of determining when a scene is safe and when the incident is over. When do EMS move in? If it is only when it is all declared completely safe this could be many, many hours which is also unacceptable.
This problem had not been clearly resolved in Norway. Like Australia they have many remote communities and it is not always possible for EMS/fire to stand off in violent incidents waiting for police to arrive. They have reached a compromise of sorts where EMS always stand off it is a firearm incident. If not firearm, then the responding personnel (fire and EMS) have some discretion as to whether they enter the scene although the may have to protect themselves or victims using whatever is available, e.g. axes or spades from fire trucks. The Norwegians do not carry stab vests, and part of the thinking here was that it may lower the threshold for responders to enter a scene if they perceive they are protected. This is a very controversial area but one that is increasingly going to be debated in prehospital care conferences in the coming years unfortunately.
The second session after lunch followed this theme with a Terrorist attack stream. Lionel Lamhaut spoke on the Paris attacks from a couple of years ago. The French appear to be at the other end of the spectrum to the Norwegians where they have physicians embedded in their police ant-terrorism units and the fire brigade is a part of the military. Hopefully we will never arrive at a point where we believe that EMS should carry weapons.
And Finally, Some Blood
Last session for the day was a stream on massive bleeding. Dan Hankins gave an overview of the approach used by the Mayo Clinic service in the US for blood products. They have been carrying red cells since 1988 on their service. CareFlight has carried red cells since at least 1987 perhaps making us the first civilian service in the world to routinely carry them but Mayo was way ahead of us on plasma having carried it since the early 1990s. We only started with this product a few months ago. Mayo currently carries an interesting mix of whole blood (1 unit), red cells, plasma and platelets although the exact combination varies a bit depending on availability. Mayo are the only service in the world routinely carrying platelets that I am aware of.
Wolfgang Voelckel who I mentioned earlier spoke about the FINTIC study (Fibrinogen in Trauma-Induced Coagulopathy) they have been conducting in Austria. This study involved randomising hypotensive trauma patients to receive either fibrinogen or placebo prehospital. They were then examining clot strength on arrival in the ED as their end point – it was not sized to assess outcoomes like mortality. They were able to demonstrate increased clot firmness at ED arrival in patients who had received fibrinogen compared with those that received the placebo. Early days yet and studies looking at mortality will need to be conducted but fibrinogen is worth watching out for.
Interestingly he noted that some of the patients that received fibrinogen on the basis of prehospital hypotension did not have bleeding identified later in hospital and he postulated that the hypotension was simply on the basis of over sedation. They are going to have to refine their criteria for inclusion in subsequent studies as fibrinogen has a clear risk of iatrogenic thrombosis (unlike the data on TXA to date) and it should not be thrown around too liberally even without considering the cost.
And that was day 1 for the scientific content. Then it was off to a very lovely dinner by a lake. I beat a tactical retreat when my colleague Chris Cheeseman started doing rounds of vodka with the local LPR doctors. This sort of fits with my broader ‘just because you can, doesn’t mean you should…’ ethos.
If this happens to come the way of any other attendees, your reflections would be greatly appreciated.
Airway management seems like the current flavour. Actually it’s sort of always the flavour. Finally, Dr Alan Garner has something to say about something that isn’t about first pass success – checklists.
At the risk of treading into an area likely to stir up as much passion as first pass success, it’s time to talk about checklists. There’s a new publication out there touching on standardisation and the use of checklists among teams providing prehospital drug assisted intubation that has just been published. You can find it here, although it is not open access unfortunately.
The authors surveyed services that they could identify providing prehospital emergency anaesthesia in the UK and sent them a questionnaire. 43 services participated. There was a spread of helicopter and road-based services in addition to three ED-based teams representing 75% of UK services. That’s a reasonable sample.
The issue that particularly grabbed my interest was the use of checklists. Most reported services used checklists, particularly the busier ones. Many services have a longer checklist they use for drugs assisted intubations and another shorter one they use for crash intubations. But unlike any paper I have seen previously this study gives a lot of detail about the checklists themselves, things like the number of items on the checklist, the wording and formatting.
The thing that caught my eye was the length and complexity of the checklists. To directly quote the study:
“On average, standard checklists contained 169 (range: 52–286) words and 41 (range: 28–70) individual checks.”
That caught my eye because the service I’m working in is a massive outlier when it comes to checklists. Out standard prehospital intubation checklist has 13 items when counted using the methodology from this paper.
This is less than half the items on the shortest checklist reported in the study which had 28 items. That’s startling enough to have another look.
Let’s Look Then…
First thing that is worth saying is that this is the checklist utilised by our rapid response helicopter service in Sydney. This service does just one thing which is prehospital trauma response within a 30 minute flight time of our Sydney base. No inter-facility transfers which have a quite different workflow. In those longer haul operations we use a longer checklist which is this one:
This checklist has 40 items which places it in the middle of the pack compared with this UK study. We use this checklist in our Northern Territory and international jet operations where intubation is a less common event for the teams. Many of the referring sites are small and sometimes have no staff with advanced airway skills. Plenty have no plumbed oxygen systems, relying on bottled gas. In our international jet operations the staff at the referring hospital may speak limited or no English and getting assistance or additional equipment can be very challenging. We therefore take nothing for granted and check everything. This appears to be comparable to the reported checklists in the study.
P is for Prehospital
But the study is about prehospital anaesthesia specifically. Many of the reported services, particularly the HEMS services, are like our Sydney service and conduct only prehospital operations. Our standard prehospital intubation checklist in Sydney is more equivalent to the “crash induction” checklist mentioned in the study both in number of items checked and word count but we use it for all intubations whether time pressured or not.
So why is our standard prehospital checklist such a dramatic outlier and why do we only have a short checklist that we use all the time? Did we sit down, follow the KonMari method and ask if every individual item on the checklist gave us joy? Well, no.
Before we look at this I should say that I’m pretty happy that our success and complication rates are very good compared with the published literature. You can see some of this in previous posts about how we measure quality in intubation practice here and here. So being bad at it and accepting lots of complications is not the explanation.
When your thing does the opposite of what you want it to do
To explain this we need to have a look at how checklists can sometimes hinder what we do. As checklists have been increasingly adopted in medicine and other safety critical industries the potential problems associated with their use are becoming clearer. Some of these are cultural – do the teams actually use the checklists in the way they were intended? Do they use them at all despite an SOP mandating them? Some of the resistance to checklists has been perception that they are just a “tick and flick” exercise for audit purposes but don’t really improve patient safety. That they slow things down and get in the way of patient care. Or that the items on the checklist are not really relevant or the list is too long and onerous. A level of checklist fatigue can result with personnel hurrying through them without really paying attention or omitting them altogether.
At this point I would seriously recommend having a listen to Martin Bromiley, a pilot whose wife died due to human factors issues during a routine operation. He discusses what checklists are and are not.
To mitigate these factors checklists need to be short, and the list needs to have only items that both can be omitted by oversight but at the same time are critical to safety. But I don’t think some of the items on the lists in the study meet these criteria or the issue they attempting to address can be managed in another way by re-engineering the process.
To illustrate what I mean 100% of standard checklists in the study had an item to ensure that an IV line had been placed and was patent. But it is impossible to proceed with a drug assisted intubation without functioning venous access (whether IV or IO). If you attempt to proceed without having checked this item you will rapidly come to halt anyway.
In other words it is not possible to omit this item whether you check it or not. So why check it? You are just wasting time.
An example of engineering out a source of error is the oxygen supply. When I worked in the UK myself years ago we did not routinely carry oxygen to the scene. You had no control over how full the bottle that came with the ground ambulance was and you needed to check every time. And the bottles that were available were only 400L which did not last all that long on a mask on high flow in any case.
Our approach is to carry our own 600L O2 bottle to every case, and use it for the intubation process every time. We checked it either in the morning checks, or after the prior case so we know it is full. So we don’t check it again.
This is another part of shortening the checklist. If you can check it before the phone goes off do so. Our checklist is really focused on the factors that we could not do before we met the patient because we did not know who the patient was and what their issues were.
Our checklist aims at optimising the process for that specific patient in terms of plan, positioning, specific drug selection and getting out the right size equipment. But everything that we could check before hand was already done and we don’t check it again.
You don’t make sure every nut on the helicopter is properly torqued before you depart on a mission because that has already been done, and this should be no different. Most of the equipment items mentioned in table 3 of the study fall into this category. We check our laryngoscope and ventilator every morning and we don’t do it again on the scene. We have had no failures of either over the past 13 years.
The only other things we do are check that we have the suction out and the monitoring on – simply because it is possible to proceed without these being in place and both are critical for patient safety. These are the items that a checklist was really designed for.
Having said this we always carry a copy of the longer checklist that we use in our inter-facility operations. If we are tasked to another case before we can properly redo our checks, or either of the prehospital team members is just not happy for any reason the team reverts to the full check list although in practice this occurs very rarely.
Getting the Team Onboard
I think it is basic human behaviour that compliance with a process will be better when team members can see that it is just what is required without unnecessary steps. That the really critical components are captured which protects both the patient and themselves.
But I think that they also appreciate a carefully designed process that has removed the requirement for additional checks by engineering out the possibility of error in the first place wherever possible. If the whole team is actively involved in process redesign through identifying and eliminating opportunities for error they own the resulting shorter checklist. They follow it because they know if the item is still on the checklist then it both matters and we could not find an alternative to checking it on the scene.
So in the end we have very high success and low complication rates but with a very brief checklist. But maybe this story is more about empowered teams, and the never ending quest for quality.
And the challenge is always there: does the checklist provide what patients and crew need? And is every item there useful, or could you have sorted it earlier so you just do the vital bits to get the job done in the moment?
Feedback is great because we don’t get better without hearing from clever people. So drop a comment. You might be the person who shows us something we could improve.
Somewhere around Sydney at the recent ANZCA Annual Scientific Meeting, Dr Andrew Weatherall had the chance to kick along a discussion about trauma in kids. This is the post version of things covered and things in the chat. This is also cross-posted over at the kids’ anaesthesia site.
Let’s start by keeping in mind a very, very important point: it’s probably not possible to find anyone near a conference meeting room in Sydney on a Thursday who is likely to be a true expert in paediatric trauma, particularly in anaesthesia. True paediatric trauma experts, the ones who know the literature backwards and have an amazing array of personal experiences that have refined their approach, are a rare, perhaps even non-existent, species.
That’s not a statement trying to offer up an excuse or throwing shade anywhere else. It’s just stats. If you look at the most recent Trauma Registry report out of NSW, our most populous state in Oz, you’ll get a chance to look at the 2015 collated serious trauma stats. For the whole of that year, across the whole of the state, there were 225 kids who got to hospital with serious injuries. 225 across the three kids’ trauma centres. Now spread that across all the people who work there and ponder how many people are likely to get the sort of exposure to get really good.
There just can’t be that much exposure. And if people tell you they see heaps, well, I reckon they probably don’t.
Which I guess means that everything that follows here should be held up to really serious scrutiny. Check the references. Size it up. See if it holds water. Add another cliché here.
The attendees at this session came from a variety of anaesthetic backgrounds from the level of student to very experienced. For most of them the main theme seemed to be ‘I don’t really feel comfortable with kids’ trauma [“Phew,” I thought, “me too”] and I don’t really get to see it much. But when I do it’s usually bad.’
This is common in lots of places. In NSW, prehospital organisations are directed to drive past hospitals and go on to a designated kids hospital with an injured child they’ve picked up unless they genuinely think that child is about to die. So if they pull up at your joint, it’s bad.
The aim here is to start with a story. In that story we’ll get to cover a range of things about kids’ trauma. It probably won’t be earth shattering. It should be practical.
So let’s get to it.
Let’s start with a standard day at your local anaesthetic joint. It’s your favourite hospital at Mt Anywhere. Like most Australian “mountains” it is, in fact, a very poor excuse for a mountain and actually “Anywhere” is really “somewhere”. I’m just being vague about the somewhere.
Let’s say it’s a solid-sized place on the edge of a metropolitan area. There is plenty of adult surgery, the occasional elective paediatric list of some sort. The place has a neurosurgeon but not necessarily continuous coverage and big kids’ stuff goes elsewhere.
You get a call from the ED because they have received a call just a few moments ago. A prehospital crew out there somewhere near Mt Anywhere have picked up a kid. This kid is 6 years old and thought to be about 26 kg. They have had an altercation with a dump truck. Ouch.
The initial assessment is that this kid is pretty unconscious with a GCS of 6, which seems not that surprising because there is a fair bit of swelling around the left eye like they took a hit. Their heart rate is 128/min, they have a blood pressure of 95 mmHg systolic. Happily when they checked peripheral saturations they were in the high 90s and they can’t find anything on the chest. They added oxygen anyway. They also placed an intraosseous needle. They are on their way. You have 10 minutes.
Big Question Number 1
So at this point the question I asked was “What are you worried about?”
I think the response was “It’s a kid. Everything.”
And then more seriously:
There were worries about the injuries themselves. Head injury was thought to be likely. The heart rate might point to bleeding somewhere and kids can compensate for a bit before they fall off a cliff.
There were some who were worried about their ability to do technical things in kids. Challenging at the best of times if you’re not doing it regularly, everyone was pretty unanimous that the situation was unlikely to elevate their performance.
What can we do here?
This last one was an excellent point. A kid with big injuries should ideally be going somewhere dealing with critically ill kids all the time. If you think there’s a good chance they’ll have to go elsewhere there should be absolutely no one in the system who would mind if you called retrieval before the patient even arrives so they can start thinking about plans. You might even find they have useful ways of supporting you and they can get things rolling if retrieval will be needed.
The patient turns up and they are basically as advertised. The obs are the same. The left upper arm looks wrong enough that you’re thinking “that’s a fracture”. The patient is a bit exposed and there’s some bruises down the left side of their abdomen.
Question 2 is pretty obvious; “what first?”
Or perhaps the better way to phrase it is “What next (and how is it different because it’s a kid)?”
The discussion pretty much came down to the following (there’s a bit of abridging here):
‘I’d use the team to assess and treat with an aim to get as much done at the same time as possible.’
‘I’d assess the airway and maintain C-spine precautions.’
‘I’d assess breathing and treat as I needed to.’
‘I’d get onto circulation, try to get access, and if I needed fluids try and make it blood products early rather than lots of crystalloids.’
‘I’d make sure we complete the primary survey and check all over…’
Now, you probably noticed that all of these things are just the same things as everyone would say for adults. Maybe it turns out they are just litt… wait, I’m not supposed to say that.
There’s a point worth noting though. If you are going to have to face up to kids’ trauma and there are things that worry you, it’s also worth noting the stuff that is close to what you are more comfortable with. There will always be basics you can return to.
Now the discussion did touch on things around the topic of how you’d go about induction of anaesthesia and intubation. There were no surprises there with a variety of descriptions of RSI with agents that people felt they were excellent at using. A whole thing on that seems like too much to go with here but you could have a read about RSI in kids at this previous post.
Likewise THRIVE (and other forms of high flow nasal prong work) was mentioned. That’s probably beyond the scope of this post too if it’s going to stay under a bazillion words but it’s worth pointing out a couple of things that are also in this thing here and here. One is that the research that has been done that’s kind of relevant to extending apnoeic oxygenation hasn’t been done in an RSI set up and the nasal prongs aren’t generally applied during the actual preoxygenation bit.
Where to from here?
Now it’s probably time to move this along so let’s say that heart rate has improved a little to 115/minute, the blood pressure is about the same and you’ve assessed all those injuries and think facial fractures are on the cards, plus a fractured left humerus.
Oh, I should have mentioned that left pupil. The one that’s big and not doing much. The one I deliberately didn’t mention until now because I didn’t want the thing to move too quickly.
This brings us to a crucial and very deliberately placed point – what sort of imaging are we going to do?
We’re going to bench FAST as a super useful option here because the negative predictive value is somewhere around 50-63% (from the Royal College of Radiologists document) and we’re moving to a cashless society so coin tosses seem old school.
Let’s assume we’re heading to the CT scanner because there is no neurosurgeon around who doesn’t want a scan to make a plan. So how much do we scan?
I threw this to the room and there was a variety of options offered. The classic Pan Scan was mentioned. Or just the head. Or maybe head and neck. Or head and neck and abdomen but maybe not chest.
Finally we get to something that really is different in kids then. In kids the threshold for exposing the patient to radiation is a bit higher than in adults. This is because the risks of dosing kids with radiation during scans are far more significant than for adults. The ALARA principle (“As Low As Reasonably Achievable”) comes very much into play here. You can find a bit more description about this here or you can look at the Royal College of Radiology guidelines.
The headline things to remember are that if you expose a kid to 2-3 head CTs before they hit the age of 15 it looks like it might almost triple the risk of brain tumours. Make it 5-10 and that’s triple the risk of leukaemia. Abdominal and pelvic CTs give you a higher dose of radiation.
So in this context in kids there is a real second thought about what scanning to do. On top of that for things like abdominal trauma it’s much more likely in kids that the surgeons will pursue non-operative management. And while there are probably better places to delve into the minds of surgeons it’s worth spending a moment with the flowchart from the ATOMAC guidelines to try and get a sense of their thinking. Or if you look at it long enough I think it works like one of those 3D eye pictures.
What is definitely the case is that treating abdominal injuries on the basis of the grade of injury as demonstrated on scanning (for spleen and liver injuries particularly) isn’t really a thing. Early decisions are based very much on haemodynamics and clinical assessment.
So in our patient where there isn’t current clinical evidence of intra-abdominal pathology (just trust me, there isn’t) and the haemodynamics aren’t suggesting hidden pathology, then the scanning is probably just going to be looking at the head and maybe cervical spine. Plus this patient is going to start with a chest X-ray (particularly after intubation).
Lo and behold, the CT head shows a left subdural haematoma with a bit of midline shift. Time to go here…
Off to theatres then and I guess the next question is:
What are the priorities for the anaesthetist here?
Everyone pretty much jumped on two:
Get on with it – meaning the thing that needs to happen to protect brain tissue is the surgeons need to do a thing. There’s not much the anaesthetists can do that will help brain tissue as much as the drilling bit in this context. Delaying for things that’d be ideal (say, an arterial line) is not really what the patient would ask for. So ‘hop to it’ was a universal endorsement.
Make sure you are giving the brain the best odds of scoring blood supply.
There was passing discussion on agents, where to have the CO2 levels, hypertonic solutions and things like that but really most of those are as per adults so people zeroed in on perfusion targets.
In kids this is a bit of a problem because there is even less good evidence compared to the adult population. This is particularly the case for blood pressures before you have access to intracranial pressure monitoring and can therefore figure out the cerebral perfusion pressure (CPP). On top of that the Brain Trauma Foundation TBI guidelines have recently been updated, but not for kids. That document still lives on from 2012 (at least for now).
When I went to check on the targets listed at The Children’s Hospital at Westmead, their CPP targets went like this:
> 10 years old aim for 60 mmHg CPP or above.
In the 1-10 year old age range aim for CPP 50 mmHg or above.
In the under 1s aim for 45 mmHg or above.
The thing is, at least when you start you probably won’t have access to intracranial pressure (ICP) to do the CPP = MAP – ICP (or CVP if that’s higher) calculations. Hence this suggestion that you should treat for a bad case scenario where ICP is assumed to be 20 mmHg because that’s when you’d step in and do something about it.
In this case you need to add 20 to your mean arterial pressure (MAP) and aim for that target. What would be kind of nice of course is having a systolic BP target. Unfortunately we don’t get that until the age of 15, where the new TBI guidelines suggest you should keep SBP above 110 mmHg.
As an aside I have some reservations about the ‘let’s just assume ICP is bad’ because assumptions seem like not the best basis for manipulating physiology. They seem even worse when you’re making a lot of assumptions about how pathophysiology will play out.
Given that TBI is associated with disruptions to the blood brain barrier and a variety of other stresses, assuming that raising MAP won’t just result in swelling, bleeding into vulnerable areas or other causes of general badness seems … fraught.
For now it’s all we’ve got though so there it is.
The Red Stuff
The surgeons do their thing of course and that means (particularly when you have certain topics to cover in a conference session) lots of bleeding. There are bigger places to go into massive transfusions in kids here, but it’s worth noting a couple of key tips:
Massive Transfusion Protocols help and emphasise the need for not just the red stuff but good amounts of a fibrinogen source (locally that’s cryoprecipitate rather than fibrinogen complex concentrates, platelets and FFP. A quick Google search will find the guideline used at The Children’s Hospital at Westmead and the breakdown of what comes first…
and what comes next…
The number for pretty much all of the units (at least to start with) is 10 mL/kg. Quickly figuring out how much 10 mL/kg is for the patient in advance makes the calculations a lot quicker.
Of course the one different one is cryoprecipitate which is around 1 unit per 5 kg (up to 10 units).
Calcium replacement shouldn’t be underestimated as an ally (or even necessity). Perhaps me ending up mostly looking after kids just coincided with everyone getting interested in calcium, but I lean on this way more than I used to, particularly as the things that are supposed to help you clot go in.
Of course you’re not allowed to talk about trauma without mentioning tranexamic acid (TXA) because we’d all like to make sure there’s at least a little less bleeding if there’s a way we can influence it. So we want to get it there and get it here quickly.
The main question then is how much should we be giving?
I can sort of see why because there’s not a huge amount of evidence out there for ideal dosing in kids, particularly in trauma. What we end up with is evidence from other settings where traumatic damage is inflicted on tissues (i.e. big surgery).
If you go to any of those settings, like craniosynostosis surgery or scoliosis surgery or cardiac surgery, you’ll see a dizzying array of dosing regimes too. Loading doses of 10, 20, 30, 50 and 100 mg/kg with infusions any of 2, 5, 10 and 20 mg/kg/hr. This only makes figuring out what to do an awful lot harder.
So when they came up with that “pragmatic dosing” they went for a pretty cautious option. That’s partly because they’re not super sure about risks of thrombosis and there’s lots of concern about seizures with TXA loading. The theory goes that with higher doses you get higher levels of TXA in the CSF and that leads to inhibition of inhibitory glycine and GABA receptors (because they have those crucial lysine binding sites). It’s not everything but there’s at least some cohort research suggesting there’s not much association. In a retrospective study looking at craniosynostosis surgery with 1638 records examined the rate of seizures was the same across groups at around 0.6%.
The problem with that dosing option is there’s enough evidence to suggest that 15/2 is just not going to cut it. You might as well get a mosquito that bit a person who once had TXA and get them to sneeze on your patient. Bigger doses seem likely to work better.
So was there a difference? Well the lower dose crew lost an average of 968 mL and needed 0.9 units of red cells on average. The higher dose crew ended up losing about 695 mL and receiving 0.3 units of red cells on average. Unfortunately there was only 72 patients in the lower dose group and 44 in the higher dose group. So we’re left with not much.
There’s enough to suggest though that higher doses are probably required to actually influence the fibrinolytic pathway. A dose of 20-30 mg/kg to start with is much more like what I’d do (without exceeding 1 g) followed by an infusion of 10 mg/kg/hour.
The Next Bit
Look, don’t you think this has gone on long enough? Everyone did great, the surgeons operated really well and everyone got through a tough day pretty well and gave our imaginary patient the best shot possible.
There were of course other things we chatted about. Things like tricks for getting that IV access (if you remember the name Seldinger and that a 0.018” wire will fit up a 24 gauge cannula you’re in good shape). Then the challenges of spine immobilization and the role of options other than a hard cervical collar. Then of course the importance of considering the impact on ourselves when we look after these kids.
None of those deserve short change though so that can wait for some other time. Or maybe there’s an expert out there for that.
The things on radiation risks in kids to look at would be this one:
Having advanced teams is no damn good if you’re not doing your best to get them to the patient where they might add value. Dr Alan Garner returns with reflections on recent publications on this coming out of Europe. It’s a bit of a passion of his.
Advanced capability prehospital medical teams may well be great, but they are an expensive resource that needs to be matched to the patients that are most likely to benefit. But these patients represent a very small percentage of the total numbers of calls to ambulance services so a way of accurately filtering the calls is critical to optimising the utilisation of such teams.
This of course means an accurate case identification system is required to dispatch these teams to the right patients, and preferably only the right patients. This is a kind of ‘Where’s (severely injured) Wally?’, all day, every day. It is really hard to pick which of the red striped shirts is blood and the signal to noise ratio about the same as a Where’s Wally picture. In NSW only about one in every 250 calls to Ambulance is a severely injured patient.
When it comes to dispatch of advanced capability medical teams (rather than which patient should go direct to a trauma centre – the two things are not necessarily the same) there is not a huge amount of literature out there yet. This has been identified as a priority research area for HEMS.
What the Scots did was move one of their clinicians who work on the service, either a paramedic or retrieval practitioner, into the control centre to look for cases that might benefit from an advance medical team response. Prior to this move the case identification was done by non-clinical dispatchers with some oversight from paramedics and nurses who were not involved in provision of the Scottish retrieval service. They used a simple before and after methodology to see how many of the severe trauma cases that occurred in Scotland were identified by the new system including the EMRS clinician versus the old system with the control room staff only.
The result was an improvement in sensitivity for major trauma that increased from 11.3% to 25.9%. Although 25.9% does not sound great it is possible that the new system identified almost all the severe trauma that was in the areas that the EMRS responds to. EMRS are not dispatched to urban areas in close proximity to major hospitals but injuries occurring in these areas were not excluded from the analysis. Since Scotland has a predominantly urban population it is very likely that most trauma occurs in urban areas (like it does in NSW) so if they are identifying a quarter of all the severe trauma cases this may represent almost all the severe trauma that is in their response area. Unfortunately this is not examined in the paper (I mean you can’t always cover everything) and more studies will be needed to clarify this.
Regardless of this methodological issue they more than doubled their case identification rate by putting a member of the EMRS team in control of identifying the cases. Such systems are common in the UK. As far as I am aware the first to publish this were London HEMS in early 90s where they were able to demonstrate a really dramatic improvement in sensitivity when they put one of the flight paramedics into the central control room. So this is not a new bit of learning. It’s reinforcing what we should know.
Stories we have heard before…
When we started the Head Injury Retrieval trial 13 years ago we had something like the London system in mind. By accident we stumbled upon an improvement though. Between 1989 when London HEMS set up their system and 2004 when we were planning the trial the internet had arrived. We were able to build a system where the crew at the helicopter base was able to screen the calls and identify the cases directly from the Ambulance computer system rather than putting a flight paramedic into the control centre.
This contributed to the trial HEMS system being the fastest reported to date in the world medical literature. We could get airborne about 3 minutes faster than the reports out of London and we’re pretty confident this was related to the ability to do multiple things simultaneously because it was all happening on base. At the same time as a clinically involved crew member was looking at the details of the case, aviators could start identifying potential landing sites and making plans. The pilot could head to the helo and start the checks even as that was happening. An experienced crew of 4 looking at cases also allows plenty of bouncing things off each other. There’s just a bunch of seemingly little stuff you can start working on that adds up to a significant bit of time-saving.
The trial system was however shut down at the end of the trial in 2011, and dispatch in NSW is now done by control room staff who are not involved in service provision.
It is worth noting that the system used to identify severely injured children when HIRT was recruiting was associated with zero safety incidents of even a minor nature, no unintended dual responses by physician teams and zero cost. But it doubled the rate of identification of severely injured children for physician team response and decreased time to a paediatric trauma centre by half an hour.
It is notable that other services are now commencing the direct screening by HEMS crew case identification system. Have a look at this article on the Great North Air Ambulance in the north of England. It sounds like exactly the process we used during the Head Injury Trial to identify severely injured children in Sydney more than a decade ago.
This comment from Andy Mawson, Operations Manager for Great North Air Ambulance is central to the whole thing:
“It’s an extra set of eyes to make sure we are getting to the right patients in the fastest possible time. Essentially we’re working in support of the teams within the NWAS control centre, it’s a great example of collaborative working.”
The system used to identify severely injured children during the head injury trial in Sydney was collaborative too, not competitive. There were extra eyes looking for the same cases rather than one set of eyes looking at the whole state of NSW trying to find severely injured Wally across four different control centres. How can this not be better?
But this is a refinement of the system the Scots studied; putting case identification into the hands of clinicians that provide the services is the core of the system. Why does this work? I think the clinicians directly involved in provision of advanced prehospital teams just understand the services they are dispatching better. They understand the actual logistics of response and the capability that can be delivered on scene. It seems that you need more than a set of guidelines to do this sorting quickly and effectively. The “gestalt” that comes from actual provision of these services is required to lift the dispatch system to the next level.
Sydney previously had a case identification system that was world leading and is now being replicated in other places like the north of England. Sydney also had the fastest physician staffed HEMS yet reported in the world literature to go with it.
But the case identification system was switched off and the effective service area of the HEMS halved. Missed cases and delayed activations occur frequently.
So after all this, the same questions from my last post on this topic in August 2016 still apply:
“You might ask why such a hugely effective system was discontinued in the first place.
You might ask why it has not been reinstated given the subsequent evidence.
And they would be very good questions.”
[Sound of silence].
We are always interested in people’s clever thoughts on tricky topics. So hit those comments if you have things to say.
Now, the first of those papers on the paediatric tasking happening with the crew watching the screens is this one:
Nasal prongs seem pretty popular for lots of things these days. So how about their use in kids. There’s a couple of papers out there on its use in the paralysed patient and Dr Andrew Weatherall is here to splice them together.
Isn’t it supposed to be the kids who stuff things up their noses? Have we just seen them do it so often we started wondering about the possibilities ourselves?
Let’s assume not. It’s more a case of people finally getting around to testing things out on kids when they’ve been running with them in adults for quite some time. This time it is THRIVE and that ever so desirable feature of endless maintenance of oxygen saturations while we get around to the ensnorkelling we’ve planned.
In principle that makes plenty of sense. The normal kid is more likely to rapidly desaturate than the normal adult. Physiology is pretty insistent on that. Plus we know people find paediatric intubation tricky so dropping the stress by avoiding the slide of the plethysmograph tone down that digital scale is probably a worthy pursuit.
So how about we look at two papers examining just this issue – does THRIVE employed in the little people stop those saturations from … not thriving??
First up is this paper published by Humphreys et al, who work out of Brisbane. They did a small RCT on well kids with 24 in the control arm and 24 receiving 100% THRIVE. The kids fell between the ages of 0 and 10 years of age and are reported in the age groups 0-6 months, 6-24 months, 2-5 years and 6-10 years (with a total of 12 in each age range, meaning 6 in the controls and 6 in the THRIVE group within each age group – got it?)
The routine went something like *induction of anaesthesia* –> pre-oxygenation by doing that whole bag-mask ventilation bit –> the mask disappeared and THRIVE was added or nothing was added –> start the stopwatch.
You’ll note that, like the other paper we’ll mention, this is not about patients who are spontaneously ventilating. That’s a completely different thing.
In this group though the period of the saturations staying up was longer. Across the age groups the extension in apnoeic time was 86.8 seconds (0-6 months), 88.7 seconds (6-24 months), 129.5 seconds (2-5 yeas) and 169.2 seconds (6-10 years).
Right, lock it up. Everyone should have nasal prongs. All the time. It’d stop peanuts ending up there too.
Except there’s more pesky nuance in this paper. Like:
1. It’s not for pre-oxygenation
It’s worth noting that the preoxygenation here was all about face-mask ventilation with a good seal. They added THRIVE after that bit and started the clock. This is not entirely surprising because we know that nasal prongs compromise seals in adults and that’s only more likely with kids.
So if you were thinking that you should set up those nasal prongs from the before time zero, you need to think again. THRIVE for preoxygenation is not something tested here, and you shouldn’t assume it’d be better than good face-mask technique.
2. They didn’t test the duration that it worked for apnoea
All they said was it’s ‘more’. ‘Wait,’ you might say, ‘you mean they didn’t test the thing that was the point of study?’
Well not really because the cut-off was ‘twice the previously noted time to desaturation’. So they tested that they could reach the ‘double or nothing’ limit, but didn’t test the full extension. In the THRIVE groups the average saturation when they stopped the clock was 99.6%.
So I guess be reassured that it was likely to be really a heck of a lot of time.
3. Basic things were part of the procedure
For this study there was a lot of basics being done well. Throughout apnoeic oxygenation they weren’t doing things like airway instrumentation, suction, intubation or, I assume, anything much beyond chatting about the weekend and watching the clock. They did jaw thrust, a basic manoeuvre likely to optimise the impact of THRIVE. So maybe we should remember that all those things we are also interested in were not part of the picture.
And Now an Update from the Swiss
What if you didn’t make your cut-off ‘2 times the other cut-off we knew about’? How long could you go?
Well a Swiss crew with no interest in being neutral on the topic I guess have done a study comparing low flow nasal oxygen (0.2 L/kg/min) with THRIVE at either 100% or 30% FiO2 with 20 in each group. And they found … (wait for it…..) 100% THRIVE prolongs apnoea time.
OK there wasn’t much suspense there really.
Except again it was more subtle, and again cut-off matters. They had a cut-off to terminate on the basis of desaturation, but another at 10 minutes (as in ‘it’s 10 minutes and I’m bored let’s stop because those saturations are still great’) and the 3rd cut-off was if the transcutaneous CO2 hit 65 mmHg.
In the THRIVE 100% group no one desaturated, 4 hit 10 minutes and the other 16 had their nasal prongs ditched when they breached the CO2 target. This actually accords with the other paper where they also found that THRIVE doesn’t achieve ventilation and removal of CO2 in kids.
But at the end of this paper you still can’t say how long apnoea might be extended, at least when it comes to those saturations staying up.
Oh, and a couple of other points:
1. Pre-oxygenation was with face-mask ventilation. Again.
Again the nasal option had no role in the preoxygenation phase. They went with face-mask ventilation until the expired oxygen was 90% or above. Then they started the clock with the chosen nasal prong option going.
2. The other airway things done at the time were … none.
Yep. Once again this was just about the oxygen and the stopwatch. Nothing else was going on.
Let’s Think Clinically
So let’s imagine that we’ve actually got that paediatric patient in front of us. Maybe one who needs to get intubated before we get them out of wherever ‘in front of us’ is.
Let’s agree that maintaining oxygenation throughout is a good and noble goal. It’s not the only goal of course. We’d also like to make sure we make good choices around number of attempts, and for some patients (say the patient with intracranial pathology) we need to think about ventilation.
And we don’t have evidence that pre-oxygenation is aided by having THRIVE in place.
So assuming we’re going to do things standard to modern paediatric RSI like face-mask ventilation for a bit before we get going. There is at least a bit of a question about whether THRIVE adds a huge amount.
What it undoubtedly adds is the confidence that saturations will stay up. That is something that lots of practitioners, particularly those not regularly intubating kids would find immensely reassuring.
There is a couple of caveats to keep in mind though.
There’s a risk to be aware of with THRIVE that those saturations staying interminably up might encourage tunnel vision on persisting with intubation when it’s not working out. It’s not too hard to imagine the scenario where the tube hasn’t passed straight down, but those saturations are OK so you persist a bit longer, and a bit longer, and now long enough that the airway is becoming traumatised and suddenly you’ve created a problem.
So this might be a cognitive challenge to have planned for in advance – how do you keep yourself to a limited number of attempts before re-evaluating and going to plan B (or C)? Do you make it a personal process or have others in the crew hold you to a maximum number of attempts or maximum duration of looking?
After all, THRIVE is going to get you to 10 minutes probably. But if you’re still conducting open negotiations with the glottic structures at 10 minutes, oxygenation is not the airway problem that should still be at the front of your mind. While you’re there, you might have to think about re-dosing anaesthetic agents too.
And the other key patient group is that one where intracranial pathology is an issue. Letting the CO2 rise for some patients is not a good plan because your TBI patient (as just one example) doesn’t need those cerebral vessels dilating and the intracranial pressure going up. For those patients, a step back to face-mask ventilation, or potentially placing a supraglottic airway, to re-establish an ability to exchange CO2 is probably a better option.
So THRIVE might be great for some things. But whether it’s clinically better than an approach to the airway where really excellent pre-oxygenation is routine and good practices around face-mask ventilation are established seems like a line ball call.
I mean it’s still way better than a piece of Lego up the nose. But it remains an adjunct to the basic stuff, not a replacement.
With a couple of new papers landing that touch on the issue of how you provide and measure quality care around airway management, Dr Alan Garner returns to point at big animals that are bad at hiding.
Two new airway papers have come across my desk in the last couple of weeks and I now wish I had waited a bit longer before putting up the last post on first look intubation as a quality measure.
So where to start? Well how about a place where everything is apparently big? Yes, there’s a bit of work just out of Texas which sheds further light on that first look intubation story so that’s where we’ll land.
It sounds like they have used RSI for a while but undertook a quality improvement project to try and reduce their peri-intubation hypoxia rate. The project involved introducing a bundle of interventions described in the paper as “patient positioning, apn[o]eic oxygenation, delayed sequence intubation, and goal-directed preoxygenation”.
The paper provides copies of the protocol for intubation pre- and post-bundle intervention in the on-line appendices so I might just go through them here to see what they did differently.
The first thing is there was an emphasis on positioning in the bundle, specifically head up a bit and ear-sternum positioning. Lots of goodness here that I strongly support.
The second measure they mention was apnoeic oxygenation. However looking at the pre- and post-bundle policies it is evident that they used it in both time periods. In the before period it ran at 6L/min till the sedation was given then it was turned up to 15L/min. In the post period however it was run at “MAX regulator flow” after the ketamine was administered. I don’t know about the O2 regulators in Texas but to me this does not sound like they changed anything significant. I will come back to apnoeic oxygenation later.
For pre-oxygenation in the pre- bundle period they used a NRB mask (with nasal prong O2 as above) in spontaneously ventilating patients (and arrested patients were excluded) but in the post- period the pre-oxygenation had to be by BVM with two handed technique to ensure a tight seal plus PEEP. More goodness here that warms my heart.
Delayed sequence intubation in this study refers to administering 2mg/kg of ketamine then maximising preoxygenation for at least 3mins prior to administration of the muscle relaxant. I don’t think this is necessary in all patients but this was the policy in the bundle.
The last thing they did was “goal-directed preoxygenation”. This refers to having a SpO2 target >93% for at least 3 minutes during the pre-oxygenation phase after the ketamine had been administered. If they could not achieve >93% the patient was managed with an LMA or BVM and transported. I think this represents sensible patient selection in that it removes the high risk of desaturation patients from the process. When you look at the results you need to keep this patient selection in mind. However I agree that in their system this is a reasonable approach to ensure patient safety for which the managers should be applauded.
Show Me The Money
Yes let’s get to that money shot:
I have been banging on about peri-intubation hypoxia being far more important than first look intubation rate for a while now and this data shows really clearly why.
There is no significant difference in this study in either first look or overall success rates pre and post the bundle but the hypoxia rate fell by a massive absolute 41%! The 16% decrease in bradycardia emphasises just how much difference they made. The managers of this system and their staff alike both need to be congratulated for this achievement as this is something that really matters. And the first pass and overall success rates give no clue!
It really is time to drop first look as a quality measure and move on. You could look at this paper and start wondering if it might even be worth dropping overall success rate too, which is an interesting thought. Their policy favoured patient safety over procedural success rates by abandoning the attempt if the pre-oxygenation saturations could not be raised above 93%. It looks like it is working out well for the patients.
Oh, Back to Oxygenation
I promised I would come back to the apnoeic oxygenation issue. I know the authors state that it was part of their bundle, but it was used in the pre- bundle period as well. Hence there is no data here to support it’s use.
All three randomised controlled trials of apoeic oxygenation in the ED and ICU contexts (see the notes at the end) have now failed to find even a suggestion that it helps (check those notes at the end for links) and there are no prehospital RCTs. My take is that it is time to move on from this one too and simply emphasise good pre-oxygenation and good process when the sats start to fall – or never rise in the first place like this group did so well.
Overall a big well done to the Williamson County EMS folks and thanks for sharing your journey with us.
Moving Right Along
The other paper comes out of London, where the ever-industrious HEMS group have published a retrospective review of their database over a 5 year period (from 2009-2014). They were looking for adult trauma patients they reached with an initial noninvasive systolic blood pressure of 90 mmHg or less (or where a definite reading wasn’t there, those with a central pulse only) and with a GCS of 13-15.
This gave them a total of 265 patients (out of a potential 9480 they attended). 118 of those underwent induction of anaesthesia out there beyond the hospital doors (though with exclusions in analysis they end up with 101 to look at) and the other 147 (that number dropped to 135 on the analysis) got to hospital without that happening.
Now the stated indications for anaesthesia listed are actual or impending airway compromise, ventilatory failure, unconsciousness, humanitarian need, patients unmanageable or severely agitated after head injury, and anticipated clinical course.
Now given that the inclusion criteria includes patients having a GCS of 13-15, it seems like both unconsciousness and those really impossible to handle after head injury are likely to be pretty small numbers in that 101. Even airway compromise, ventilatory failure and humanitarian need seem like they’d be not the commonest indications in that list that would apply to this patient group, though they’d account for some.
I guess it’s possible the patients were all initially GCS 13-15 on the team’s arrival but deteriorated en route, though I just can’t sift that out from the paper. Plus if that was the case it seems like you’d say that.
In their 236 study patients, 21 died and 15 of those were in the ‘received an anaesthetic’ group. The unadjusted odds ratio for death was 3.73 (1.3-12.21; P = 0.01). When adjusted for age, injury mechanics, heart rate and hypovolaemia the odds ratio remained at 3.07 (1.03-9.14; p = 0.04).
Yikes, sort of.
What To Make of That?
I guess we should make of it that … things you’d expect to happen, happen? Intubating hypotensive patients and then adding positive pressure ventilation in the prehospital setting is potentially risky for patients for a variety of known pharmacological and physiological reasons that the authors actually go into.
So the question is why embark on such a procedure where you know the dangers in detail? You’ve have to really believe in it to end up wiht 101 cases to follow up.
It feels like there’s an elephant in the room to try and address by name. I wonder if it has something to do with a practice I observed while working in the south-east of England 8 years ago. It relates to that last category “anticipated clinical course”.
The concept here is that if you figure the patient is going to be intubated later on in the hospital, you might as well get on and do it. Except the data here suggests that, much like you’d expect, you probably shouldn’t get on and channel your inner Nike marketing script.
Just because you can does not mean you should. This paper really drives this home though it doesn’t really seem to come straight out and say it. It does pass the comment that “Emergency anaesthesia performed in-hospital for patients with cardiovascular compromise is often delayed until the patient is in theatre and the surgeon is ready to proceed.” Perhaps the problem isn’t using the phrase “anticipated clinical course”. It might be that you just have to remember that the anticipated course might best contain ‘risky things should probably happen in the safest spot’ in the script.
Compare and Contrast
The process of undertaking emergency anaesthesia because later the patient might require emergency anaesthesia is pretty much the complete opposite of the approach from the Williamson County EMS folks. They erred on the side of patient safety and withheld intubation if it was associated with unacceptable risk.
This paper demonstrates that emergency anaesthesia in patients with a high GCS but haemodynamic instability is associated with higher mortality. We should probably be glad the authors have made this so apparent, because this is probably as good as we’re going to get. We’re not going to get a randomised controlled trial to compare groups. No one is allowing that randomisation any time soon making this another example of needing to accept non-RCT research as the best we’ll get to inform our thinking.
Patients with hypovolaemia due to bleeding need haemorrhage control. The highest priority in patients with that sort of hypovolaemia would seem to be getting them to the point of haemorrhage control quicker. And delaying access to haemorrhage control (because the prehospital anaesthesia bit does add time in the prehospital setting) when the patient has a GCS of 13-15 doesn’t seem to prioritise patient safety enough. Patients probably need us to adjust our thinking on this one.
That seems like common sense. The retrospective look back tells us pretty conclusively it’s a worse option for patients. And now it’s up to us to look forwards to how we’ll view those indications for our next patients. And “anticipated clinical course” probably just doesn’t cut it.
That hovering elephant head was posted by James Hammond in a Creative Commons-like fashion on unsplash.com and is unchanged here.
How about all those things that got a mention above that you should really go and read for yourself?
Here’s that whole bundle of care paper out of Texas:
Greg Brown returns to look at an important thing relevant to first responders (and lots of other people really) – the sucking chest wound.
We’ve all been there – sitting through some kind of “first aid” training and having some kind of “first aid trainer” speaking authoritatively on some kind of “first aid style” topic. If you are like me you’ve used your time productively over the years and perfected what my wife refers to as “screen-saver mode” – it’s that look on your face that tells the instructor that you are listening intently, often supplemented by the insertion of “knowing nods” or head-tilts, but in actual fact you are asking yourself “if I was able to collect all of my belly button lint over a 12 month period and spin it into yarn, I wonder if I could make enough to abseil off London Bridge?”
Don’t get me wrong – I reckon effective and accurate first aid training should be a mandatory part of having a car / bike / truck / bus licence. More appropriately trained people should mean faster recovery rates for most injured people (and less work for overstretched first responders).
It’s just that sometimes first aid trainers teach stuff based on ‘we reckon’ or ‘that’s how we’ve always done it’ rather than evidence or knowing it works in the real world. This post is about one of those things.
“What is a sucking chest wound?”
In the Army questions come in a few different shapes and sizes. A popular one is “there is only one obscure answer you should have guessed I wanted”. Trust me, the muzzle velocity of your primary weapon is 970 metres per second.
Another popular one is “the question that should be about one thing, but is actually to demonstrate a quite tangential point”. Like,
“What is a sucking chest wound?”
For an army instructor the answer is not what you are thinking right now. It is “Nature’s way of telling you that your field craft sucks and everyone can see you and now you got shot”.
Let’s Go With the Medical One
We’re going to go with the alternative, more medical one. A sucking chest wound is defined as air entering the thorax via a communicating wound that entrains air into the space between the lungs and ribs more readily than the lungs can expand via inspiration through the trachea.
This is about pressure differentials – in order to inhale, the lungs must generate a relative negative pressure such that air can be sucked into them via the trachea. But if you make a big communicating hole in the trachea, that might become a pretty big highway for air to enter the space with the negative pressure.
The communicating hole does need to be pretty big. Depending upon which textbook you read, this hole needs to be a minimum of a half to three quarters the diameter of the trachea. Also, the patient needs to be undergoing relative negative pressure ventilation (or, in simple terms, breathing spontaneously). If they are being artificially ventilated (which requires positive pressure) then the pressure inside the lungs will be higher than the pressure on the outside of the body; the result is that air will be forced out of the intra-pleural space (or thorax) by the expanding lung (as opposed to being entrained into the thorax via the hole in the chest).
Are sucking chest wounds really that bad?
Well, yes. They suck in fact.
A sucking chest wound creates what is known as an open pneumothorax. Let’s consider the option where that hole does not seal on expiration. We’ll get onto the also very annoying sealing with a flap version in a bit.
In this slightly not so annoying case, the patient will have a ‘tidalling’ of air in and out of this communicating hole. The effect? Respiratory compromise, increased cardiovascular effort and reduced oxygen saturations. Patient satisfaction? No, not really. Death? Maybe – depends on what other injuries exist and the ability of the individual to compensate. See Arnaud et al (2016) for more details.
But if this communicating hole were to seal itself on expiration then you now have an open tension pneumothorax. Sounds bad; IS bad.
In such a case, each time the patient breathes in they will entrain air through the communicating hole in the chest wall (that whole “negative pressure” thing in action). But when they breathe out, instead of having that additional intra-pleural air tidal outwards, the flap will seal it in place; each time they breathe in, the volume of trapped air will increase and you’ll end up with the tension bit.
How much air is required? Well a randomised, prospective, unblinded laboratory animal (porcine) trial conducted by Kotora et al (2013) found that as little as 17.5mL/kg of air injected into the intra pleural space resulted in a life-threatening tension effect.
Actually, that’s a fair bit of air…for those of you who are lazy and don’t want to do the math, that’s 1400mL for an 80kg person. But remember, any tension pneumothorax (open or closed) is progressive – each time you breathe, more air is trapped; therefore, it doesn’t take long to reach crisis levels.
“But are they common enough for us to be worried about?”, I hear you asking. The short answer is yes – in fact, the long answer is also yes.
Kotora et al (2013) reviewed the statistics from the Joint Theater Trauma Registry regarding contemporary combat casualties with tension pneumothorax and found that they accounted for 3 – 4% of all casualties, but 5 – 7% as the cause of lethal injury.
“Yes, but I don’t live in a combat zone…”, I hear you say. I have two responses:
Good for you; but also,
According to Littlejohn (2017), thoracic injury accounts for 25% of all trauma mortality. And sure that stat is for all forms of thoracic injury and a sucking chest wound is but one of those but there’s a neat article by Shahani which sums up the incidence nicely and it turns out you should give this some thought.
So, your field craft sucks – now what?
Now that we know that sucking chest wounds are both possible and bad, we should probably discuss treatment.
Back in the mid 1990’s, Army instructors were very big on rigging up a three-sided dressing. Unwrap a shell dressing, turn the rubbery-plastic wrapper into a sheet and tape three sides down with the open bit facing the feet to allow blood drainage.
And, in an astonishing turn of events, everyone I’ve met who tried this confirmed it didn’t really work that well.
In that Littlejohn paper they make reference to the fact that by the 2004 ATLS guidelines (which are not usually that quick moving), it was being written unblock and white that there was no evidence for or against the three-sided dressing option. It was done because it sounded good in theory, but the evidence wasn’t there.
Now to the New
Actually, not that new. Chest seals already existed.
These chest seals (at that time the Bolin produced by H & H Medical, and the Asherman produced by Teleflex medical) included one-way valves to allow for the forced escape of trapped intrathoracic air and blood. basically they took the impromptu three-sided dressing and made it a ready-made device in the form of an occlusive dressing with an integral vent.
But did they work?
Yes and no.
On a perfectly healthy (albeit with a surgically created open pneumothorax) porcine model with cleaned, shaved, dry skin they sealed well and vented air adequately.
However, once the skin was contaminated (dry blood, dirt, hair etc) the Bolin sealed much better than the Asherman. And if there was active blood drainage too (such as in an open haemo-pneumothorax) then all bets were off. Both vents clogged with blood and ceased to work. Sure, you could manually peel the seal back and physically burp the chest but if you did so the Bolin became an un-vented seal and the Asherman was as good as finished (i.e. it wouldn’t reseal). But hey, at least you had sealed the communicating hole and in doing so stopped entraining air.
“Is this the best you can do?” you may be asking. Well to be honest, since the vents didn’t work for more than a breath or two most people decided that the vents were pointless. The outcome was that we all decided to forget about the vents and just seal the wound. That way, assuming that there was no perforation to the lung, this open tension pneumothorax (aka sucking chest wound) became a routine, run of the mill, plain old pneumothorax. And if there were signs of tensioning (e.g. increasing respiratory distress, hypotension, tachycardia….) one just needed to peel back the seal and manually burp the communicating hole thus relieving the pressure. Use a defib pad – those bad boys stick to anything! Problem solved….
Or how about a newer idea + research?
In 2012 the Committee on Tactical Combat Casualty Care (CoTCCC) started questioning the efficacy of contemporary practices regarding the placement of chest seals on sucking chest wounds. It had already been accepted that the current vented chest seals had ineffective vents, so practice had changed from using a chest seal with an ineffective vent to simple, “soldier proof” unvented seals and burping them as required. Surely there had to be a better way…?
Kotora et al (2013) decided to test three of the most readily available vented chest seals in their aforementioned randomised, prospective, un-blinded laboratory animal (porcine) trial: enter the Hyfin, Sentinal and SAM vented chest seals.
What they found was that all three were effective in sealing around the surgically inflicted wounds and in evacuating both air and blood. Thus, in 2013, CoTCCC changed their recommendations back to the use of vented chest seals.
But there were still some questions:
Once life gets in its messy way, do they seal (or at least stick to skin)?
Are all vent designs equal?
To answer question 1, Arnaud et al (2016) decided to evaluate the adhesiveness of the 5 most common chest seals used in the US military using porcine models. What they found was that the Russell, Fast Breathe, Hyfin and SAM all had similar adherence scores for peeling (> 90%) and detachment (< 25%) when tested at ambient temperatures and after storage in high temperature areas when compared to the Bolin. The researchers admitted, though, that further testing was required to assess the efficiency of the seals in the presence of an open tension haemo-pneumothorax.
In response to question 2, Kheirabadi et al (2017) tested the effectiveness of 5 common chest seals in the presence of an open tension haemo-pneumothorax (again, on porcine models). Essentially, there are two types of vent: (i) ones with one-way valves (like in the Bolin and Sam Chest Seals), and (ii) ones with laminar valves (like in the Russell and Hyfin Chest Seals). Their question was: do they both work the same?
What they found was that when the wound is oozing blood and air then seal design mattered. They found that the seals with one-way valves (specifically the SAM and Bolin) had unacceptably low success rates (25% and 0% respectively) because the build-up of blood either clogged the valve or detached the seal. By contrast, seals with laminar venting channels had much higher success rates – 100% for the Sentinel and Russell, and 67% for the Hyfin.
Sucking chest wounds are bad for your health.
Sealing the wound is good.
If the seal consistently allows for the outflow of accumulated air and blood, then that’s even better.
Therefore, now that we know all of this, one’s choice of chest seal is important. At CareFlight we use the Russell Chest Seal by Prometheus Medical (and no, we’re not paid to mention them we’re just sharing what we do). Why? Because it works – consistently. Both for us and in all the aforementioned trials.
The premise of this addition to the Collective is that you’re a first responder. That being the case, use an appropriate vented chest seal on a sucking chest wound.
However, you still need to recognise that the placement of the seal does not automatically qualify you for flowers and chocolates at each anniversary of the patient’s survival – you still need to monitor for and treat deterioration. Such deterioration is likely to include a tension pneumothorax for which the treatment is outside of the scope of most first responders (other than burping the wound).
If you are a more advanced provider then your treatments might include the performance of a needle thoracocentesis, or perhaps intubation with positive pressure ventilation and a thoracostomy (finger or tube).
In essence, know the signs and symptoms then master the treatments that are inside your scope of practice. (Or you could enrol in a course…such as CareFlight’s Pre-Hospital Trauma Course or even THREAT… OK that was pretty shameless.)
Meanwhile we’d love to hear:
What chest seal do you use?
How does it go?
Or you could just tell us what other things you think suck.
We’re not kidding about hearing back from you. Chip in. It only helps to hear other takes.
You could also consider sharing this around. Or even following along. The signup email thing is around here somewhere.
That image disparaging all things Kale (or kale) is off the Creative Commons-type site unsplash.com and comes via Charles Deluvio without any alterations.
Now, here are the articles for your own leisurely interrogation.
Dr Alan Garner has been here before, asking whether we’re asking the wrong questions when we try to measure quality advanced airway care. Here’s a fresh bit of research that adds to the discussion.
Unintended consequences would hardly be a new thing in medicine or in any other endeavour. Here is one of my favourite examples taken from Wikipedia (look we all go there from time to time):
“The British government, concerned about the number of venomous cobra snakes in Delhi, offered a bounty for every dead cobra. This was a successful strategy as large numbers of snakes were killed for the reward, but eventually enterprising people began to breed cobras for the income. When the government became aware of this, they scrapped the reward program, causing the cobra breeders to set the now-worthless snakes free. As a result, the wild cobra population further increased. The apparent solution for the problem made the situation even worse, becoming known as the Cobra effect.”
Avid or maybe even occasional readers who chanced to come back at exactly the right moment might recognise that I have previously expressed my doubts about reporting the first look intubation rate as a quality measure for intubation. Have a look here for the previous post.
Now where might you go to find a basket of cobras these days? Well I have just spotted a new paper published in Prehospital Emergency Care which fits the bill. You can find the full text here. I guess we’d better start picking up the snakes.
Let’s Start with the Headlines
This paper is a look at a ground paramedic system in a small US city (Spokane in Washington State) where the paramedics have used muscle relaxants for more than 20 years i.e. you would have to consider this a mature system. It appears to be a well supervised system and paramedics have a minimum number of intubations they must successfully perform each three-year certification cycle in addition to a well-structured training regime.
Superficially the system appears to be working well. They had a 95% success rate and 82% first look success. Although 95% overall success rate is below par compared with other systems world-wide, all patients not successfully intubated were successfully managed with a supra-glottic device. That should be OK, right? That probably means the primary focus is on managing the airway to achieve the goal that really counts – oxygenation. And that first look rate of 82% seems quite respectable compared with reports from other systems. So not a star system but safe enough if these were the only quality measures you were looking at.
Let’s Get Our Hands Right Amongst the Snakes
The thing is the paper also reports physiological data captured by the patient monitor during the peri-intubation period and this tells a very different story. Much of the data is not that surprising. Desaturations were more common when patients were being intubated for respiratory pathology and were also related to the highest SpO2 achieved at the end of pre-oxygenation.
How about we look at some oximetry data highlights?
Oximetry data was available in 110 cases. Peri-intubation desaturation occurred in 47 cases (43%) and in 32 (68% of the desaturations) it was severe (<80%).
The median nadir was 71% and median duration was 2 minutes. Among cases with any desaturation, the time in the unhappy valley was at least 2 minutes in 46% of cases with first-attempt success and in 100% of cases requiring multiple attempts.
Although the frequency of desaturation was significantly higher in cases requiring multiple laryngoscopic attempts versus a single attempt (70% vs. 37%; p = 0.01), 70% of all desaturations occurred on first attempt intubation success. Only 11% of desaturations were reflected in the EMS patient care report.
Heart rate changes
13% became bradycardic, 7% profoundly. The median SpO2 nadir during bradycardic episodes was 30% with median duration of nearly 5.5 mins.
Sixty percent of bradycardia events occurred on first-attempt intubation success.
Yes in the multiple attempt cases the desaturations were worse than cases requiring a single attempt. But given the very high rate of desaturation events in this study is reporting the first pass success rate providing any meaningful quality data? Is there subtle pressure placed on the paramedics in this system to achieve first pass intubation at the potential expense of desaturation events, by the very fact that first pass rate is being reported?
We can’t be sure and I’ll put my hand up and say “yes, I’m inferring a little bit from what we can see in the paper”. But clearly the overall success and first pass success rates provide no real indication of process safety in this particular EMS system. It is only in reporting of clinically meaningful quality data like desaturation that we see the real safety performance.
Who Else Thinks This?
To quote the paper itself “What may be obscured by this focus on the risks associated with multiple intubation attempts is the large absolute number of physiologic derangements occurring on first-attempt success. In our study, 70% of all desaturations, 60% of bradycardia episodes, 63% of hypotension episodes, and one of the two cardiac arrests occurred on first-attempt success.” That’s really the nub of it and it’s excellent work by the authors to make sure that’s right up there in the discussion.
The authors conclude that first attempt success “is not a reliable indicator of patient safety.” The authors specifically note that prolonged duration of first pass attempts is a contributor to the desaturation rate and that prolonged attempts might be “a consequence of lack of awareness of the passage of time during an intubation attempt, or lack of awareness of the occurrence of desaturation”.
But is the very fact of reporting first pass success rate a subtle psychological contributor too? The authors clearly agree with me here when they comment “prolonged desaturations on first attempt success could be an unintended consequence of the focus on first-attempt success itself and the common use of first-attempt success as a primary measure of intubation quality.”
Maybe it’s an example of the Cobra effect.
The Take Home Bit
Prospectively it is right to set yourself up to get the ETT in the right place on the first attempt and with minimal complications. However once the intubation attempt commences the emphasis needs to shift to prevention of complications by reacting to physiological changes as they occur.
We want to encourage this. I want my teams obsessed with preventing complications, not first pass success. Why are we reporting a process measure as a quality indicator when it might well be having the perverse effect of encouraging those very complications we were trying to remove? The system I work in here in NSW requires us to report first pass success. I remain hesitant to do this as I don’t want to signal to my teams that this is actually something that matters. I would much rather them be proud of the 0% desaturation rate that we have for intubation over the last 9 months – that is really impressive.