Thursday, April 19, 2012

Helicopters & trauma patients: JAMA paper deconstructed

Twitter and the media were abuzz with reports of the paper in this week’s JAMA that found a 1.5% survival advantage if trauma patients were transported by helicopters rather than ground ambulances. The headlines were sensational and misleading.

LA Times: Trauma patients more likely to survive if rescued by helicopter
MSNBC: Helicopter beats ambulance for trauma patients
US News: Helicopter Beats Ground Transport for Trauma Victims: Study
Scientific American: Helicopters Save Lives for Serious Trauma Victims

What I am going to do is point out the many flaws in the paper and give you some information about the other side of the story. As is true of nearly all papers, you should read the full text, not just the abstract.

The paper looked at outcomes for 61,909 trauma patients transported by helicopter and 161,566 patients were transported by ground ambulances. Although the unadjusted results showed a higher mortality rate for those transported by helicopter, they were also more seriously injured. The adjusted mortality rates favored the helicopter by a statistically significant but small 1.5%. The number needed to treat was 65 to save 1 life and the cost for saving 1 life was estimated at $325,000, the same figure as heart surgery.

The authors used the National Trauma Data Bank [NTDB] as their source. More than 900 of the 6000 or so hospitals in the US voluntarily contribute data. Although the NTDB is said to contain more complete data than in prior years, some issues remain.

For example, the paper started with records of more than 1.8 million patients. Various appropriate exclusions, such as restricting the analysis to seriously injured patients, whittled the number down to 978,000. Over 115,000 records were then excluded because information about their method of transportation was not reported. More than 102,000 patients not treated at Level I or II trauma centers were omitted too. Of the patients included, some 38% were missing more than 40% of the data that was supposed to have been submitted. Remember, this is the so-called “improved quality” NTDB data.

This was a retrospective study and the authors point out that no accepted nationwide guidelines for first responders on the ground to call in a helicopter exist.

The cost analysis was done using information from the state of Maryland, which uses state police helicopters. At $5000 per flight, it is one of the least expensive medical helicopter services in the nation. Other cost estimates range from $5000 to well over $15,000 per flight. And some studies have shown that nearly half of all patients transported by helicopter are not even admitted to the hospital. The costs of those flights, had they to be included in the calculation, would certainly increase the cost per life saved.


There is also the issue of medical helicopter safety. Many crashes occur every year with several resulting in the deaths of patients and crew. [National Transportation Safety Board Report, page 32].

Should the safety record and the number of patients unnecessarily transported by helicopter be considered in the number needed to treat?

The authors mentioned several limitations of their study including the following:

They imputed [translation: made up] data for variables where data was missing in no more than 20% of the records. There is selection bias because hospitals submitting data to the NTDB are likely doing better than those which do not submit data. It is unknown whether the modest survival improvement is due to the helicopter or the fact that helicopters are usually staffed with MDs, nurses and advanced paramedics as opposed to paramedics and EMTs on ground ambulances.

The literature is replete with papers stating that helicopters do not increase survival rates. In fact one that was presented at the American Association for the Surgery of Trauma just appeared last month in the Journal of Trauma. Other than prehospital time which is significant longer for helicopter patients, there were no differences in any parameters including survival.

In the discussion of this paper, a well-known trauma thought leader, Dr. Ken Mattox, said, “So my question is, is it time for the ROC [Resuscitation Outcome Consortium] or this organization [AAST] or organized medicine to take the power of our convictions and our evidence and make a policy statement on this very expensive advertising mode that does not really alter outcomes?

Bottom line. The study left out almost as many patients [mode of transport unknown and those not taken to a level I or II trauma center] as it included. Many of the patient records include in the study had missing data. The cost per life saved was calculated to be as favorable as possible and did not include the substantial costs of unnecessary flights. The abysmal safety record of medical helicopters was not mentioned. The helicopter may not be the factor leading to the modest [1.5%] increase in survival rate, but rather it may be the way they are staffed.

How do the headlines look now?

14 comments:

RobertL39 said...

Please post a link for the "NTSB Report, page 32"
Thanks.

Skeptical Scalpel said...

Link is posted in the blog. Sorry, I omitted it originally.

Joey Tranchina said...

One, your calculation assuming current costs, therefore, the cost would be less with greater utilization of choppers... possibly much less.

Two, given traffic blockage — like in LA anytime or summer in the South of France — it seems that the time savings to trauma center, therefore the advantage, would be greater,

Three, if 1/65 life savings is demonstrated, there must be ways to both triage and leverage the service, to offer grater savings at lower cost, while targeting its use to the most appropriate patients...

Skeptical Scalpel said...

Joey, I'm afraid I don't understand your comment about costs. The costs are fixed. They don't decrease with use.

Traffic can be an issue in urban areas but sometimes there's no place to land at the scene either. Even the most ardent helicopter advocates admit that ground ambulances are faster in cities.

There are ways to triage but it's not being done. That's why I said there aren't any accepted guidelines. If some EMT wants to see the helicopter, he just calls and asks for it.

peter sherren said...

All very valid points and there are major concerns regarding the paper.

However, I have a few points.

Helicopter EMS safety in America is for some reason lagging a long way behind Europe and Australasia. Nice video summary on transport by the rfds http://www.youtube.com/watch?v=UWfIc4cYSWY . Worth noting not only inter country variability but also risk of death with ground ambulance transfer.

Cost issues are variable. Greater number of helicopters, flight hours and competition for ems contracts could drive down individual flight costs.

As for the fact a physician/highly trained paramedic is carried on a helicopter ems, while I agree that may well affect mortality rate I feel that is the whole point of a helicopter ems. Its about putting the right resources in the right place at the right time. Sometimes in severe trauma that will involve a helicopter and physician. I presume you wouldn't let a junior resident operate alone in a small hospital miles away from a trauma centre and attending trauma surgeon.

While I understand a lot if this is subjective and difficult to prove, I would rather have a well trained physician and paramedic team looking after me following a severe trauma which can transport me safely and rapidly to a trauma centre than standard road ambulance.
While a physician is not much use in the pre hospital setting in uncontrolled haemorrhage (although can admister blood) they can perform a number of time critical interventions (rsi, thoracostomies, blood etc etc) at the scene that can benefit patient. Its the time to meaningful/critical intervention that is key, not time to arrival at hospital ED. An extreme example is pre hospital thoracotomies undertaken in arrested patients with penetrsting injuries in the UK. Small but significant number percentage of survivors (18%). Difficult to argue which you would prefer for your own family if there was a twenty minute transfer time to the ED. http://www.ncbi.nlm.nih.gov/pubmed/21131854/?

The Erstwhile Medic said...

Joey, then you start factoring in the costs of more funerals and lawsuits associated with the increased number of crashes inherent with the increased number of flights, etc. Also, it would consume more fuel, demand more crew members, etc. Your argument falls apart with even the most shallow of a review.

Skeptical Scalpel said...

Thanks for commenting. The paper citing successful on scene thoracotomies by anesthesiologists and ED docs is indeed remarkable and unlikely to be duplicated anywhere else in the world. As for time on scene, even the JAMA paper showed it was longer for helicopter transport. And people conveniently leave out the time it takes to launch the helicopter and for it to get to the scene.

There is an impending shortage of doctors in the US. Is deployment on helicopters the best use of MDs?

I agree that unlike other commodes or services, more helicopters will not drive down costs.

peter sherren said...

A little extra scene time to undertake a meaningful intervention for example a rsi, start neuroprotective measures and address any amenable chest injury saves time in hospital. Its the time from injury to CT scan/intervential radiology/theatre that is key, not scene time. A well trained and equiped pre-hospital team with good ongoing clinical governance can perform ~90% of the immediate resuscitation that would occur in an ED. Not only can they perform vital procedures they can do it safely and usually quicker than in the ED due to strong teamwork and drilling of the procedures. Instead of spending 30 mins stabilising a combative and hypercarbic tbi in the ED, you can receive a packaged patient and be in ct in 5-10 mins. Might of spent an extra 10-20 mins on scene but this saves time in hospital. Clearly this does not apply to all cases and there are times when a 'scoop and run' approach is appropriate such as exanguinating chest/abdo/pelvis injuries.

I think the above issues simply highlight a difference in thinking and practice of pre hospital care in America vs Europe / Australasia. However, as always its not often possible to prove beyond doubt that one practice is superior to another.

Thanks for all the posts.

Doctor Bob said...

All of these studies have significant flaws, but everyone I know who works at a Level I trauma center shares my experiential observations: each and every shift I work carrying the trauma beeper, I see patients brought in by helicopter from scene runs who could have been taken by ground to the local community hospital, appropriately evaluated (by the competent docs I know who work there), and discharged. You could write an entire blog essay on why that happens and why we're not doing anything about it.

Skeptical Scalpel said...

Peter, thanks for your interest. You make some good points but Doctor Bob brings up an issue that many people don't want to talk about. There is a lot of wasted money in helicopter transports that turn out to be nothing.

Skeptical Scalpel said...

Peter, I accidentally deleted your most recent comment. Please resubmit it. Thanks.

Skeptical Scalpel said...

I realized that the comment by Peter was still present in the email notifying me of it, so here it is.

From peter sherren:

"By having a senior doctor such as doctor bob making such triage decisions on scene would avoid such frustrations of inappropriate trauma centre transportation.

May never agree on the cost issue. Simply depends on your trauma centre , catchment population and geography will dictate what is utilised and whether helicopters are a 'waste' of money. There is so much practice that occurs on a dailey basis which has no evidence base but we know is the clinically right thing to do. I think for all physicians working in pre hospital care this is the case.

93490578-c6dd-11e1-830b-000bcdcb471e said...

I operate an air medical program. I agree with most of your perspective I offer some procedural aspects of my program that address some of the issues. I operate a team that consists of a critcal care RN and a paramedic. Both must have a minimum of five years documented experience and none have less than ten. We provide RSI with a greater than 99% success rate. Each team member must maintain a competency in advanced airway by rotating through surgery with anesthesia in which they perform 5 - 10 intubations every rotation. This is a quarterly rotation. 100% of flights are reviewed for appropriateness and we are at 100% BUT this is only because we have a policy that the crew ride in the ambulance to the closest hospital if the patient does not meet criteria for flight. 100% of our flights are both peer reviewed and reviewed by our medical control. Our aircraft has all the FAA recommended equipment for safe HEMS operation. We work closely with all our competitors and even meet on a regular basis so that we ensure that the closest aircraft is dispatched in the safest manner possible. We are not for profit and I am not pressured to meet a minimum number of flights. We strive for 10 minute scene times for trauma but only meet this about 75% of the time because a majority of these patients require RSI and most EMS agencies in the area do not have RSI capability. Ground transport to the closest trauma center is close to an hour for the majority of our response area. Partnerships between private, state and public entities could decrease the operating costs while increased regulation on the State level would assist in some aspects of both cost and safety.

Skeptical Scalpel said...

Dear 93490578, it looks like you are doing it correctly. Yours may be a situation where the helicopter works best due to the long ground transport times.

How does your crew get back to your hospital if they ride in the ambulance to the nearest hospital? What does that do to the availability of your helicopter while the team is off on that side trip?

Post a Comment