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

[Akulahawk]

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.

If this is the case, I would tend to agree. If such a patient showed up in my ED and we did the same care, that patient would be coded as "respiratory failure." It very well could be the case that the "respiratory failure" patient gets coded as some sort of "respiratory insufficiency" and not failure because the patient is no longer in frank failure upon arrival at the ED.

You make an excellent point about a huge problem that may just lie at the heart of the study.

 

[Summit]

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.

THIS PATIENT WOULD BE PUT IN THE RESPIRATORY DISTRESS ALS SAMPLE

From their supplement on sample construction and diagnosis codes:

We linked ambulance claims to the nearest in time hospital claims using the beneficiary's
identification number and the date of service. This allowed us to use diagnosis codes that
described the medical emergency rather than any subsequent developments. Our algorithm
prioritized linking to the nearest in time hospital claim (up to two days after the ride) and to
inpatient over outpatient claims. In each diagnosis group, at least 91% of ambulance transports
were linked to inpatient claims. The vast majority of ambulance transports linked to outpatient
claims (at least 94% in each diagnosis group) were for patients who either had died in the
emergency department and therefore were not admitted to the hospital or were transferred to
another health facility, according to discharge status codes.
A2. Diagnosis Codes
Trauma cases were identified by ICD-9CM codes 800 to 959.9, excluding late effects of
injury, foreign bodies, complications, and burns. Falls were identified by external cause codes
E880-E888 for accidental falls (excluding E887), and were analyzed only in 2010 and 2011,
which, unlike earlier years, include separate external cause code fields. Those fields were
completed for 92% of observations.
We used primary diagnosis codes for AMI (only initial episodes, 410.x1), stroke (433,
434, or 436), and respiratory failure (518.4, 518.81, or 518.82).

The study RESPIRATORY FAILURE sample includes

1. Acute respiratory failure

Disease Synonyms

• Acute hypercapnic respiratory failure
• Acute hypercapnic respiratory failure due to obstructive sleep apnea
• Acute hypercarbic respiratory failure secondary to obstuctive sleep apnea
• Acute hypoxemic respiratory failure
• Acute respiratory failure from obstuctive sleep apnea
• Alveolar hypoventilation
• Hypercapnic respiratory failure
• Hypoxemic respiratory failure
• Postprocedural respiratory failure
• Respiratory failure

2. Other pulmonary insufficiency, not elsewhere classified

Disease Synonyms

• Acute respiratory distress
• Acute respiratory insufficiency
• Pulmonary insufficiency
• Respiratory distress, acute

3. Acute edema of lung, unspecified

Disease Synonyms

• Acute pulmonary edema
• Pulmonary edema (fluid in lungs), acute
• Pulmonary edema, acute

http://www.icd9data.com/

So yes that would be 518.81 or 518.82 even if they also get a 428.0 or 428.1 if this was a heart failure exacerbation (would exclude the 518.4). I am not a coding expert.

 

[TomB]

Thank you!

 

[NomadicMedic]

 

[SandpitMedic]

Is ACLS not the same in hospital as it is in the field? Aren't we doing the exact same things? If you collapse in the waiting room of the ER and code they are going to do the exact same thing I am going to do when I get to your house if you are in full arrest. That's the point of ALS, to provide the same level of ACLS care in the field. The drugs are the same, the algorithm is the same, and the survival rates are the same (when everything is done as it is supposed to). Sure in a ER the doc will still give bicarb, whoop dee do.

That's the point. We do all of those things and then we take the patient to the hospital... Mostly a save is a save to discharge. Many times I have seen people code in the ER and the likelyhood of them surviving does not enhance just because that's where they coded. I've seen about the same survive in hospital as out of hospital. For all of our advanced technology and understanding we still cannot save everybody. The national average is under 12% for survival with intact neurological capability to have a normal life. Most become vegetables because we sustained their physical body when they normally should have passed. That, or they do die because they're a train wreck, and their bodies have already been too taxed to live- death occurs, as it has since the beginning of time.

There are times when it is appropriate for life sustaining measures like a young person or a child... But if you are an 89 year old renal failure, diabetic patient, and your heart stops.... Your ticket is punched, no matter what we do or where you collapse.

To me that is why I find these studies flawed fundamentally. It does not take into account those patients which are/were going to expire regardless of resuscitation efforts or level of skill. There is no way to know. If they did the study on say, people under 40, or 50, or whatever arbitrary age you want to throw out there, I think it would give us more of an understanding. If my child has a cardiac arrest, I want an ALS ambulance there... I want PALS started ASAP. (Yes, I know most of the time it's hypoxia related and BLS provides that). Just in case of the worst case I want all the capabilities there and ready to go.

Back to adults:
14 minutes of ACLS in the field is no different than 14 minutes of ACLS in the hospital. That's why we stay on scene and do those measures; because bumping down the highway reduces the efficiency of compressions. Stay and play- in the ER that is essentially what they are doing. The problem is mainly getting folks to not interrupt compressions, which is the latest push from the AHA for the last several years, and that has indeed brought up the stats of survival anywhere from about 3-6% depending on what study you look at.

ALS vs BLS?

If I code, I want ALS.

 

[SandpitMedic]

I realize some of you may strongly disagree with me and feel as though we should want to provide life saving efforts for every single living organism on the planet, so I apologize if my slightly morbid opinion offends some of you.

There is a line between common sense, giving someone the dignity of an honorable and natural death, and being in a field in which you are bound to provide efforts even if you morally object. I get that. I have no problem with palliative care, and giving someone a morphine pump with an all access VIP button to do with as they wish. I just don't think we ought to be doing full ACLS on 90 year old people who have lived a full life and haven't contemplated a DNR.

And before someone says, what about your family- my grandma is 96... And if she died in front of me I would call my family to come say goodbye, not beat up her chest just so she could be taken and connected to machines, in a coma, for an extra 7 days of what some people consider "life."

Sorry for the derail.

 

[SeeNoMore]

But the question is do you need ALS on that code or just CPR and defib. I'd like to imagine that ALS would provide some benefits post ROSC , but I don't think any studies show that. If we can't show any benefit to ALS it's fair to consider getting rid of it , as much as that would mess up my prospects. Although who knows maybe helicopters would be called more if there was no ALS at all for certain high acuity patients.

 

[Carlos Danger]

To me that is why I find these studies flawed fundamentally. It does not take into account those patients which are/were going to expire regardless of resuscitation efforts or level of skill. There is no way to know.

Sure it does.

Those propensities (the statistical likelihood of a given outcome) are taken into account and controlled for, if the statistical analysis is done correctly. It's not perfect in a retrospective study such as this one, but generally the larger the sample size the more reliable the findings, and this one was huge.

A good clinical statistician is a wizard.

 

[SandpitMedic]

Yeah, I think you can say that all day long...

But if you aren't Jesus, from the future, a wizard, or whatever diety you prefer - there is no way to say who would live and who dies anyway in a matter of fact way.

Math me all day long, but this point is another thing we will permanently and desicively disagree on. Stats class be damned.

 

[cruiseforever]

But the question is do you need ALS on that code or just CPR and defib. I'd like to imagine that ALS would provide some benefits post ROSC , but I don't think any studies show that. If we can't show any benefit to ALS it's fair to consider getting rid of it , as much as that would mess up my prospects. Although who knows maybe helicopters would be called more if there was no ALS at all for certain high acuity patients.

What would the benefit of a helicopter be? Unless if ground transport to a Trauma Center is greater than a hour.

 

[Carlos Danger]

Yeah, I think you can say that all day long...

But if you aren't Jesus, from the future, a wizard, or whatever diety you prefer - there is no way to say who would live and who dies anyway in a matter of fact way.

Math me all day long, but this point is another thing we will permanently and desicively disagree on. Stats class be damned.

"Screw you and your big words and fancy science. We don't believe in nun uh dat that book stuff 'round here. Dem numbers iz for heathens!"

Demonize what you don't understand. Why am I not the least bit surprised? Maybe this is why paramedicine is advancing so rapidly as a highly respected clinical specialty.

 

[jrm818]

"Screw you and your big words and fancy science. We don't believe in nun uh dat that book stuff 'round here. Dem numbers iz for heathens!"

Demonize what you don't understand. Why am I not the least bit surprised? Maybe this is why paramedicine is advancing so rapidly as a highly respected clinical specialty.

Except that the claim that pure statistical methods can tell us anything about causality is NOT all that well accepted. As far as I know, conventional scientific wisdom still suggests that it cannot.

This topic actually seems fairly ill-suited to study using non-randomized study design, given the high number of variables that need to be accounted for with a purely-statistical method (including not only patient factors, but overall system design, stupid things like how aggressive the billing department is (eg often they harass physicians to include higher acuity billing codes), the care provided at the hospital...which may well depend on care provided in the field, etc.).

Here's an interesting little excerpt about these statistical methods:

http://bayes.cs.ucla.edu/BOOK-09/ch11-3-5-revised3.pdf


And personally I still can't get over the fact that someone who appears to be a non-clinician published this in an internal medicine journal. It's not just that it's weird....there is a lot of perspective that is lacking when someone is reading or writing about a topic they have limited background information on.

Edit: and I will say that sometimes it is appropriate to be skeptical of things that are not easily explained. I have no shame in saying that this propensity score nonsense is a bit over my head.

Understanding randomized controlled trials is easy; as such that is a very useful trial design because the majority of readers can evaluate the strengths and potential weaknesses of such a trial, and make conclusions about the applicability of the trial to their own practice

Even if this fancy magic-math can indeed prove causality, I simply cannot wrap my head around methods that use such crazy math (and I took stats, calculus x2, etc. at one point). As such, when I approach such a paper I have no clue if the methods were used appropriately, what the potential pitfalls of the methods are, and how to apply the research to my own practice. Even if it's right, there is no way for the majority of even well-informed clinicians to be confident about the information. Not helpful....

 

[SandpitMedic]

I didn't say demonize science or things one does not understand- I said the math is not going to give you accurate enough information. There are to many random unknowns... And even unknown unknowns (thanks Donald Rumsfeld).

You are making this a matter-of-fact issue, and the reality is it is anything but that. I am not completely discounting the universal language, but I am saying it may not be as applicable here as you'd like it to be. There is no way to tell who would have benefitted from BLS over ALS or ALS over BLS after the fact of death in too many situations. Some, sure they are generally easy to hypothesize; most though, is just grief and emotion getting in the way of logic. No one could know for certain.

If anything, I'd like someone to present a study of survival to discharge after cardiac arrest from say, the 1950's, to today. I'll bet my bottom dollar it is better now then it was then, right?

NOPE! http://www.healio.com/cardiology/vascular-medicine/news/print/cardiology-today/{9e0f9dd0-f2b5-4a6f-9505-bc34043540c1}/out-of-hospital-cardiac-arrest-survival-rates-stable-but-not-improved-after-30-years

They had BLS efforts then, and ALS efforts came... But look, the numbers don't lie according to you...

Statistical analysis can be skewed and variables tweaked to come up with just about any kind of finding one would hope to see. That is my point.

We can not save every one. That is it... BLS, ALS, meat wagon, flight team, in-hospital arrest. It doesn't matter. We just do what we are trained to do and adjust and adapt to the latest science based evidence - not math based probabilities. Maybe that is too simple for a wizard like you.

 

[SandpitMedic]

The real issue is who we let be paramedics, and the speed at which one can obtain the patch. There needs to be more barriers to entry, more experience required, more education, more clinical time, and more maturity before we turn anyone who wants loose to "practice medicine." There are many medics out there who don't even want to be medics.... It's just a stepping stone, one which they may not take as seriously as folks on this forum. That also has an impact on the way in which they perform their duties.

That is a derail, but that is evident in most systems. Most can point to one or more of "those guys" that they work with.

 

[Carlos Danger]

Except that the claim that pure statistical methods can tell us anything about causality is NOT all that well accepted. As far as I know, conventional scientific wisdom still suggests that it cannot.

I've never suggested - or even seen anyone else suggest - that pure statistical methods can show causality.

This topic actually seems fairly ill-suited to study using non-randomized study design, given the high number of variables that need to be accounted for with a purely-statistical method

There is no way to do an RCT on a sample of tens of thousands of patients on BLS vs ALS. So retrospective studies like this are done to identify new research questions. Again, they are not intended to show causality or to change practice.

I have no shame in saying that this propensity score nonsense is a bit over my head.

The nuts and bolts of it are well above my head, too. It's high level stuff for statisticians, not for most clinicians. But the basic concept is simple: if X condition is known to have Y probability of a given outcome, then the outcomes of patients who have that condition can be predicted mathematically and compared to the actual outcomes of the sample to ensure that they are representative.

Skepticism is healthy and necessary and I support it 100% in every case. But dismissing a study - and any research that doesn't support your biases, really - out of hand because you've never taken the time to learn even basic research methods and terminology and because "only Jesus knows who is going to die" isn't skepticism, it's ignorance.

 

[SandpitMedic]

That was sarcasm... I'm not on the bible jock.
I was illustrating my point in a satirical way.
No one can see the future, in this case who will live or who will die.

Color me skeptical.

Don't try to paint me in the loony column because I made a joke and don't fall in line with your line of thinking.

Nice misdirect.

 

[jrm818]

It's funny...I just got an e-mail that there's going to be a 40 minute round-table with the author at this years ACEP conference next Monday. I'm sorry I can't go....

I've never suggested - or even seen anyone else suggest - that pure statistical methods can show causality.

On a careful re-read, the paper does indeed dance around implying causality. The author does really heavily suggest she is describing a causal relationship in her presentation posted on youtube, however....she even goes as far as to imply that her primary question is "how does ambulance type affect survival.

She also says
"if all of these patients instead got basic life support than I estimate that an additional 15 people would live to at least 90 days."
"This is a big problem, because currently if we call 911 we get the advanced ambulance."
"so how is it that advanced is worse than basic ambulance."
"Using data and statistics we can study causality in real world settings that are otherwise difficult to replicate in experiments"

Very heavy on the implying that ALS is the cause of the bad outcomes....

There is no way to do an RCT on a sample of tens of thousands of patients on BLS vs ALS. So retrospective studies like this are done to identify new research questions. Again, they are not intended to show causality or to change practice.

Fair, if this were a question generating piece of research. I'm sorry, but I really see the author's presenation of the work as implying causality, and implying that the work implies that we need to start hacking away at ALS level care. The author's responses to critical letters after their first paper suggest the same.

The more I think about this piece of research, the more I dislike it. There is no good research question that comes out of this. It isn't an attempt to establish equipoise to justify an RCT of ALS vs BLS. It has too many topics to really evaluate any patient population or intervention, so we end up with a poorly elucidated skim job.

My other problem is that I get heavy overtones of financial savings in the papers and her presentations (and her CV); and the quality of the science here is nowhere good enough to make money-saving (aka resource slashing) decisions.

I get a bit too much of a feeling that the author is happy she found the results that she did. It seems likely to me that a career would get more mileage out of demonstrating to medicare how they can save money, so I really start to wonder about the underlying biases of the author. I suspect there is more incentive to generate research which finds expensive interventions not useful, which makes me a bit more skeptical from the get-go.

 

[Carlos Danger]

On a careful re-read, the paper does indeed dance around implying causality. The author does really heavily suggest she is describing a causal relationship in her presentation posted on youtube, however....she even goes as far as to imply that her primary question is "how does ambulance type affect survival.

She also says
"if all of these patients instead got basic life support than I estimate that an additional 15 people would live to at least 90 days."
"This is a big problem, because currently if we call 911 we get the advanced ambulance."
"so how is it that advanced is worse than basic ambulance."
"Using data and statistics we can study causality in real world settings that are otherwise difficult to replicate in experiments"

Very heavy on the implying that ALS is the cause of the bad outcomes....

OK. So what? This is where skepticism comes in. If you think the authors conclusions are wrong, then fine. But if you want to convince others that this study is BS, then you need to define specifically why it's BS.

Something other than "I don't understand it, so it must be wrong."

 

[Summit]

more experience required, more
education, more clinical time, and more maturity before we turn anyone who wants loose to "practice medicine." There are many medics out there who don't even want to be medics.... It's just a stepping stone, one which they may not take as seriously as folks on this forum. That also has an impact on the way in which they perform their duties.

I agree with what you said in the above quote. But, that is just my opinion.

But are you proposing this as an explanation and potential solution that addresses the results of the study that you say we cannot trust? Upon what evidence do you base your claims?

We just do what we are trained to do and adjust and adapt to the latest science based evidence - not math based probabilities.

This statement, however, makes it clear you fundamentally do not understand how evidence based medicine works... but you think you do...

Is ACLS not the same in hospital as it is in the field? Aren't we doing the exact same things? If you collapse in the waiting room of the ER and code they are going to do the exact same thing I am going to do when I get to your house if you are in full arrest. That's the point of ALS, to provide the same level of ACLS care in the field. The drugs are the same, the algorithm is the same, and the survival rates are the same (when everything is done as it is supposed to).

Many if not most hospitals do not stick to the ACLS algorithm. ACLS is a starting framework. Also, hospital inpatients are a different population than out of hospital arrests.

If anything, I'd like someone to present a study of survival to discharge after cardiac arrest from say, the 1950's, to today. I'll bet my bottom dollar it is better now then it was then, right?

NOPE! http://www.healio.com/cardiology/vascular-medicine/news/print/cardiology-today/{9e0f9dd0-f2b5-4a6f-9505-bc34043540c1}/out-of-hospital-cardiac-arrest-survival-rates-stable-but-not-improved-after-30-years

They had BLS efforts then, and ALS efforts came... But look, the numbers don't lie according to you...

Look... you don't even have an abstract interest in evidence based medicine. I say that in an overly literal way. As in you lacked the interest to read even the abstract of the study that was poorly summarized in the link you provided. If you had an interest, you'd find that it says something a little different than you imply:
http://circoutcomes.ahajournals.org/content/early/2009/11/10/CIRCOUTCOMES.109.889576.full.pdf+html
If you wanted to read the actual study, you'd realize that they only included studies from 1980-2008 and their generalizations were that there wasn't a worldwide improvement over ~30 years as of 2008. Their biggest conclusion was that bystander CPR is the biggest way to improve outcomes and they speculated that TH could make a difference (and they would have speculated that the 2010 AHA guidelines might as well).

 

[SeeNoMore]

I dont have access to the full text , but I found this in a story on ems1.com "The study found that ALS medics take longer than clinicians in hospitals to perform procedures, such as the administration of intravenous fluids and drugs and intubation." and goes on to imply this is part of the reason for better BLS outcomes. Does the study address this? If so it almost would seem to suggest that earlier ALS intervention would be beneficial. I'm not claiming that though. Do you folks think this will actually generate additional studies or just fade into the background?