STUDY: BLS better than ALS for trauma, stroke, respiratory distress

BLUF: Don't hit the panic button, there are most likely methodology flaws in this study although the question raised is valid, and they are right to a certain extent.

This is the same group that wrote a paper (last year? maybe early this year?) about outcomes after cardiac arrest when comparing ALS and BLS. Although I don't have full access to the current study, given that it is the same group, using the same raw data and looking at the same question, I'm confident in saying that they did the same thing and made the same mistake. Yes, I know the first response will be to blow this off without thinking.

The problem is that they are using medicare billing codes to group people into groups for comparison. Great idea, except it goes by the belief that a cardiac arrest is always a cardiac arrest and always the same, or in this case, a stroke/trauma/MI/respiratory failure is always a stroke/trauma/MI/respiratory failure and always the same. This is patently not true; the simple act of using a medicare code to indicate what was wrong with the patient (especially when done by EMS providers) and possibly how severe the problem is does not allow the true comparison of 2 similarly coded people without actually looking at the individual patient. To put it another way; one group might say that patient X had a stroke, while another says patient Y also had a stroke and use the same codes...except patient X had a mild CVA with almost full resolution, and patient Y had a severe cerebral hemorrhage. Or, very likely given that it's EMS, patient X didn't even have a stroke but was coded that way by the ambulance company. If all that is looked at is medicare codes and who transported the patient, not the actual patient severity, this leads to a false comparison.

Before anyone says it, in the first study they did look at some of the initial hospital medicare codes as well and I bet they did the same here; unfortunately this does not validate the whatever was chosen by the ambulance companies. Medicare allows billing and coding if the problem was "present at any time" (or words to that effect), and given the problems that were looked at (cardiac arrest, MI, CVA, major trauma, respiratory failure) it would be very unsurprising if many hospitals and doctors coded these encounters in a specific way due to the initial concern and reception.

At least in the first study (and again, I don't have access to the full version of the current one) the way they analyzed the numbers to "prove" their validity also showed the flaws in their reasoning, although this was not recognized by the authors. It was recognized by several other doctors who were nice enough to write in to the Annals of Internal Medicine about it. While not meant as a disparagement towards AIM in any way, shape or form, I think there may have been a reason that both studies were published there and not in any of the emergency medicine journals.

So, is the authors conclusion wrong? Well...I don't think so, though it isn't really fully right either. I do think they reached their conclusion in the wrong way and are using a broad brush (even a spray gun one might say) to paint a picture that needs a finer touch. In reality, paramedics are very much misused in America, and even though we do tend to treat a vast number of patients (in most places) the majority of those people don't NEED any treatement; they would have the same outcome if they got a quick ride to an ER and waited for a bit longer. And when I say majority, I mean about 95% or more. The times when a paramedic will have a real impact (and in this setting comfort and symptom relief don't come into play) on a patient's outcome are very few and far between.

The sooner this is recognized the better of EMS will be.
 
Triemal, I have read the whole study and supplement. In a few moments you will also have that ability :)

The major methodology problems you predict:
1. I didn't see them. They looked at hospital codes. They looked at ISS. They performed tests aimed at identifying confounding effects.
2. If they were there, the effect should amplify ALS positive outcomes, not worsen them! SOB that turns out to be anxiety would group to the ALS sample.

Sure, no study is perfect, and I see minor flaws. They admit potential flaws as well and try to limit the conclusions in the discussion based on these potential problems, but like you said, this study is saying something that we should listen to.

We should always ask: is this really something we should listen to?
I think so.
The questions now should be: why is this happening and what should we do about it?
 
Summit said:
1. I didn't see them. They looked at hospital codes. They looked at ISS. They performed tests aimed at identifying confounding effects.
Click to expand...
I too, don't have full access to the study, but in the editorial published 10/13/15 http://annals.org/article.aspx?articleid=2456126, It does state that the study used billing codes explicitly, and the only clinical information included was their adjustment for ISS.
 
Summit said:
Triemal, I have read the whole study and supplement. In a few moments you will also have that ability :)

The major methodology problems you predict:
1. I didn't see them. They looked at hospital codes. They looked at ISS. They performed tests aimed at identifying confounding effects.
2. If they were there, the effect should amplify ALS positive outcomes, not worsen them! SOB that turns out to be anxiety would group to the ALS sample.

Sure, no study is perfect, and I see minor flaws. They admit potential flaws as well and try to limit the conclusions in the discussion based on these potential problems, but like you said, this study is saying something that we should listen to.

We should always ask: is this really something we should listen to?
I think so.
The questions now should be: why is this happening and what should we do about it?
Click to expand...
Thanks for that, I'll take a look when I have time today or tomorrow and give some more feedback.

1. If they did take the ISS into account for trauma patients then that does alleviate some of my concern for that specific subset, and, to be honest, that's one that I don't have much doubt about being correct anyway. For real trauma patients, the things that can be done prehospitally at the paramedic level are often a) done poorly, which makes the situation worse, or b) the wrong thing to do in the first place. While I think there are exceptions to that, taken as a whole I'd say it's a pretty accurate blanket statement.

1.5 The tests that they performed to see if there were confounders are what gives me pause; if it's similar to what was done in the cardiac arrest study it actually lends credence to their interpretation being off; I'll look.

2. I'm not sure what you mean here: "SOB that turns out to be anxiety would group to the ALS sample" No, it would group to whichever group brought them in and how it was billed; if billed as ALS it would go to the ALS group no matter the true illness/severity, and the same for if it was billed BLS. The issue would be if enough of the BLS calls were coded as something that was more severe than they actually were; if that's the case the outcome of the patient would be better because the problem was less.

I don't know that I would go so far as to say that this is something that we should LISTEN to...but it absolutely is, hands down, without a doubt something that we shoulding be TALKING about, and something that does need to be answered. Like I said, I think they are likely more correct than not, just that the reached that conclusion in an erroneous way that could lead to the wrong things being done.
 
triemal04 said:
2. I'm not sure what you mean here: "SOB that turns out to be anxiety would group to the ALS sample" No, it would group to whichever group brought them in and how it was billed; if billed as ALS it would go to the ALS group no matter the true illness/severity, and the same for if it was billed BLS. The issue would be if enough of the BLS calls were coded as something that was more severe than they actually were; if that's the case the outcome of the patient would be better because the problem was less.
Click to expand...

Per the Study:
If dispatched as ALS then downgraded to BLS, the call can still be billed to Medicare as ALS. Since doing so is perfectly acceptable, it is standard practice and these downgraded patietns are still grouped in the ALS sample, which should work in the ALS sample's favor, yet the ALS results down benefit enough from this bias in their favor to flip the results to what we would expect, which is that ALS interventions save more lives than when not present.

Also:
Moreover, BLS patients were older and had more comorbidities than ALS patients, which suggests that any unobserved differences in severity may actually have biased our results against BLS.
Click to expand...
Yet, the BLS sample fared better.

The authors make a compelling argument that their data, if biased, is biased against BLS and in favor of ALS, yet, despite this, the data do not show the expected benefit from ALS. The data show the opposite.
 
Last edited:
chaz90 said:
E
The percentages are small, but in a sample this large even a small change is compelling.
Click to expand...
They were percentage points, not percentages. If we compare 3% mortality to 7% mortality (or survival), that is 4 percentage points difference but a 133% increase or 57% decrease. 4 percentage points difference in absolute risk looks like a small difference but can actually be a large relative risk (or odds ratio).

This next example is based on the study lead author's own estimates about 4 minutes into the video.

Consider 100 severe trauma patients
63 survive with BLS or ALS
15 die with ALS care but survive with BLS care.
22 die whether they get ALS or BLS care.
Thus, this 15 percentage point improvement is a 24% higher survival rate going from ALS to BLS or a 19% lower survival rate going from BLS to ALS.
And this 15 percentage point improvement is actually a 40% lower mortality rate with BLS (vs ALS) or a 68% higher mortality rate with ALS (vs BLS).
This is not a small percentage, this is a mass casualty.

I would also point out that the study lead author points out that ALS took twice as long (27 vs 13 minutes) from arrival on scene to arrival at ER, though this was from a different study, and considers this the most likely cause of the difference in outcomes.

Also, scoop and run by police instead of waiting for EMS did not hurt patient outcomes (after adjusting for patient accuity) for GSW/stab wounds; it appears speed and skill/equipment were running neck and neck there with speed having a slight edge. In an urban environment with multiple trauma centers and short transport times, where the difference in transport times could be a pretty substantial portion of the total.
http://onlinelibrary.wiley.com/doi/10.1111/j.1553-2712.2010.00948.x/pdf
 
SandpitMedic said:
http://www.ems1.com/research-review...not-determine-whether-ALS-is-better-than-BLS/

A rebuttal to the OP's cited study.
Click to expand...
How is that a rebuttal? He doesn't refute any claim nor claim the study is flawed. He points out that no study can control for everything, that the study is worthy of consideration.

His point is well taken. If you cannot control for everything, you can either decide:

1. One cannot know anything with enough certainty to change practice based on evidence.
2. One can use the best evidence available to inform practice and further inquiry.

What one absolutely cannot do is choose option 1 when a study says something undesirable and option 2 when it matches a preconceived desirable position. There are some around here who are willing to do exactly this.

EMS has too many problems to use such mental gymnastics to road block potential progress.

Similarly, anyone who thinks this study supports eliminating ALS is also missing the point.
 
Last edited:
Well that's a rebuttal. He is rebutting the presentation that BLS is better for patients than ALS. That is what a rebuttal is, his counter point is that the study is inconclusive with too many variables to decide a "winner."
 
I really think the results are due to time. Since BLS has options less for clinical management, their scene times are probably going to be shorter than ALS. I would think that this quicker delivery to definitive care would result in improved patient outcomes. Due to the variety of diagnostic tools, increased education and treatments, i would think ALS tends to remain on scene and do ALS things, even if its stuff like ET over a King, multiple attempts at venipuncture ect. Granted this is my conjecture but i think it seems logical that with more tools, the providers will just spend more time doing even a basic (to their level) assessment.
 
I mean... I'm just sharing the article as it relates to the topic. I'm not arguing for or against. I don't really care- nothing will change regardless of what this study finds or doesn't find. To me, we should be placing our intelligence and efforts elsewhere, but that's just my opinion.
 
SandpitMedic said:
Well that's a rebuttal. He is rebutting the presentation that BLS is better for patients than ALS. That is what a rebuttal is, his counter point is that the study is inconclusive with too many variables to decide a "winner."
Click to expand...
That is not what your link said!

And the study I posted didn't say that BLS is outright better than ALS.

It says that ALS has worse outcomes in non-urban environments for some types of patients where we think that ALS should have had better outcomes.

There is no "winner."

ALS is not worthless or bad.

The study suggests that something about the way we are doing ALS is doing more harm than good for these patient types in non-urban environments, and we should take a close look for the causes and change practice to correct the issue.

There is a "loser": THE STATUS QUO!
 
This study probably isn't going to change any practice anywhere, and it shouldn't. What are we doing to do based on it, just stop sending ALS? Of course not. "ALS" is not a procedure or a drug that we can stop using. In the context of this discussion, ALS should not even be viewed as a "level" that is distinct from BLS. ALS is simply an arbitrary point on the continuum of clinical care that is provided to sick and injured people. It is an artificially delineated set of skills, and it isn't even the same from place to place. What even makes a skill "ALS" vs. "BLS"?

So, because this study didn't provide any information on what exactly it was that ALS personnel were doing differently from BLS personnel that appears to negatively affect outcomes, there's no way that we can use this study to change our practice.

That doesn't mean it isn't important information. The value in this study is that it will hopefully get folks thinking about the things that we do and whether or not they are actually helpful to our patients. History is full of all sorts of things that seem so glaringly obvious that no one bothers to question it for years and years. The earth being flat and the center of the universe, MAST pants, intracardiac epi, spinal immobilization - the legitimacy of all these things once appeared so self evident that to suggest they were wrong was heresy. So what's next? What will paramedics in 20 or 30 years sit around laughing at the thought of us doing way back in 2015? Is a normal SP02 perhaps bad in the long-term for a COPD or CHF patient? Is hypotension actually protective in trauma patients? Which co-morbidities require a different approach to management, which we don't do now? Which chemical mediators are actually helpful in stroke, but we are wiping them out with our treatments?

One of the problem with research in EMS is that paramedics tend to identify on a personal level with their scope of practice, and this makes it hard for them to accept that maybe some of the skills that make them feel worthwhile as a professional are actually not good for patients. So even when the methodology of a study is impeccable, people still find an excuse for why that finding isn't applicable where they practice. That's no more obvious anywhere that in the airway management debate. "But things are different here. We have a 99.99% first pass success rate with our intubations. Those medics where that study was done must just suck". On some level there is some truth to that - and this is another one of the problems with EMS research - from place to place, paramedics have different competency levels, different drugs and protocols, and different logistical needs. So that add a level of complexity to the generalizing of research findings.
 
More randomized trials are the best outcome of a study like this. This kind of study is like John Snow and the pump handle -- we have no idea (playing fast and loose with phrasing here, I know) why ALS isn't showing better outcomes when we have plausible reason to expect that it would. What we need now is to run more OPALS-type studies...
 
I think the truth is that the vast majority of people don't know or care what type of ambulance they get. They want to go to the hospital. If enough press is generated about how ALS is delaying this without benefits , it could be towns and cities will decide to abandon Paramedics in favor of cheaper BLS. And maybe that's a wise decision. If not we need some focus on hown ALS can be made more effective and we need to stop lashing out as those who question our usefulness.

I don't know if it would make a difference but I work part time in a system with a small number of ALS units and often arrive on scene as the pt is being loaded into the rig. The delay in transport is often a handful of minutes with interventions being performed en route (average transport times of 15 minutes.) I don't have any proof but I feel like fewer Paramedics with more experience might be part of the answer.
 
Interesting study, it raises good questions. But I'm skeptical of its validity, though it appears they really worked hard to control for confounders. This is based on billing data, and I'm not too confident that billing is going to capture patient acuity or level of care provided as well as we'd like. (Anecdotally, emergency physicians often under-bill for services. I had met with a semi-retired EM physician that did consulting work for practices and health system for physician documentation. She would compare the billing codes to the actual charts and often found that patients were usually much sicker than billing codes indicated. EMRs have changed this somewhat, which has raised alarms with insurance companies, so now they want to decrease payment since they are now having to reimburse for actual services provided, which is more than it has been historically.) They do make the point that if ALS is dispatched because of perceived need for ALS that ALS can be billed. But, I'm not so sure, unless they are referring to billing for "ALS assessment". I actually question whether all EMS' actually do this (I worked for a service that did not bill for ALS assessment). Considering that the majority of EMS systems are "all-ALS", it's really kind hard to know if BLS is really just BLS. Also, we don't know if if all the BLS was really part of the "true" EMS system, as in are many of these BLS transport done by private ambulances that are not actually part of the local EMS system. One thing that makes me think this is so is that the SNF transfers were far and away more frequently performed by BLS. Its tough to know how this could skew things since SNF patient are older with more comorbities and have the potential to be very sick. To some degree (I know we all have our stories about the staff of SNFs), patients are eval'd by someone with medical training and often with consultation with a physician, so it would be hopeful that when the EMS is activated it is because the patient is sicker.

Will this change practice? Maybe (big maybe). If this research is used by CMS to justify decreasing billing for ALS or for further scrutinization of ALS billing, I could see some systems struggling to continue ALS at the present level.
 
I would be curious to know whether folks here believe that the difference in pt outcomes is primarily related to decreased on scene to ED time for BLS. This makes sense in the case of serious Trauma and perhaps Strokes, though I would hope that no matter who is responding every effort would be made to not stay on scene for these patients.

I would have expected that ALS would provide some benefit for the sickest respiratory patients, and that early intervention with meds, CPAP and possibly well done intubation would improve outcomes vs BLS.
 
Here's my question. EMS responds to a patient with difficulty breathing with trouble speaking between breaths. The patient has a cardiac history. S/he is hypertensive with adventitious breath sounds. Initial SpO2 in the high 70s. The patient gets a 12-lead ECG, nitroglycerin, and CPAP. By arrival in the ED the patient is doing much better. Respiratory rate is now 24 (down from 40) and SpO2 is 98. Does it get "coded" at the hospital as "respiratory failure"? If not this entire study needs to be thrown in the garbage.
 
You must log in or register to reply here.
Back
Top