EBM, or, why do we do what we do?

[deleted109597]

Advertisement - Members don't see this ad

So, in the interest of continuing discussions that have sprung up in multiple threads, especially the pressor thread, I would like for people to discuss EBM.
As some have put it, if we simply "follow the rules", then we are in effect robots and can easily be replaced with noctors or worse. However, there are plenty of gray areas that I personally believe doctors excel at. We just don't have any studies, and since DNP researchers are doing all of the noninferiority studies to help them out, I don't think we can ever prove we are better at thinking.
But some people argue against EBM (again, specifically the pressors) in the face of ample evidence to the contrary, because of "physiology".
I'm not talking about something like tPA, where there are a myriad of studies all over the map, but we give it because of litigation risk. I'm talking about things such as dopa vs levo, NG lavage vs not doing it, and insulin bolus or bicarb in DKA. There are numerous studies that all point the same direction on these, but people still do it "because it makes sense" or some other argument. So I guess my question is, why?
How many of you out there give insulin boluses with their drips? Or bicarb?
How many NG lavage on your own, without GI telling you to?
How many choose dopa for all cause hypotension when other agents are available?
How many suture in a sterile field?
How many give IM shots to people who can tolerate PO (not counting epi or tetanus).

Of note, any other topic can be discussed, these are just the two most recent on here, and a few that have been pet peeves of mine for some time.

 

[xaelia]

- no insulin bolus. bicarbonate only in a case of hemodynamic instability attributable due to severe acidosis.
- no NG lavage.
- dopamine only temporarily when peripheral access is the only access available - although it might be mythology that it's any safer than norepinephrine through a peripheral vein.
- non-sterile suturing (and absorbables whenever and wherever possible).
- rarely anything IM than can't be given PO - usually ends up being IM dilaudid for pain unrelieved by PO narcotics (don't think we have PO dilaudid available, which actually works pretty well).

Though, I can't say these are all (some are) "evidence-based" practice patterns based on specific literature.

And, since I rarely have my own patients anymore, mostly I let the residents do whatever they want. Because they're continuously bombarded with unpredictable expectations from various attendings, unless it's truly harmful, I just try to give them a non-judgmental justification for alternative ways to do it if it's relevant. I try not to forget how painful it is to be a resident.

 

[KarlPilkington]

No insulin bolus/bicarb/NG

I do use IM narcs/antiemetics fairly often. Patients seem to feel better sooner than with PO and there seems to be a bit of perceived benefit from receiving something in a shot. The risk/expense of IM is minimal, and patients seem to respond well.

 

[Apollyon]

Well, EBM is "what have you done for me lately", as, "we have been doing 'x' forever, but, today, we're doing 'y' - but, then, tomorrow, it is 'x' again, and, "what were we thinking?""

The latest evidence is only as good as the next study being published, stating the prior evidence is wrong wrong wrong.

 

[jdh71]

But some people argue against EBM (again, specifically the pressors) in the face of ample evidence to the contrary, because of "physiology".

This didn't happen anywhere in the pressor thread, at least not the recent pressor thread. You may be referring to a different pressor thread, and in that case never mind.

EBM is only as good as it is. It's not bad, in fact it's good, but every single study is done within a specific context, asking fairly specific questions, with enough flaws that we all tear even the best studies appart in joural club. EBM is better than guessing but it's not all-know, nor all-powerful. Look at all of the early-goal directed therapy we do when it can't even be reproduced?! My position is don't try and make EBM say more than it actually does.

 

[emergiQ]

An oldie.... but absolutely a goodie.

http://www.bmj.com/content/327/7429/1459?view=long&pmid=14684649

 

[deleted109597]

Well, EBM is "what have you done for me lately", as, "we have been doing 'x' forever, but, today, we're doing 'y' - but, then, tomorrow, it is 'x' again, and, "what were we thinking?""
The latest evidence is only as good as the next study being published, stating the prior evidence is wrong wrong wrong.

I agree, to a certain extent. However, I'm not talking about changing from nicardipine to whatever new name brand HTN drug and then back every week as the competing drug companies release new sponsored trials. I'm also not talking about small studies, or case reports, or similar things. I'm talking about studies that have been repeated, time after time, sometimes over decades, and people still not following the recommendations (ie insulin boluses, or bicarb for DKA).

EBM is only as good as it is. It's not bad, in fact it's good, but every single study is done within a specific context, asking fairly specific questions, with enough flaws that we all tear even the best studies appart in joural club. EBM is better than guessing but it's not all-know, nor all-powerful. Look at all of the early-goal directed therapy we do when it can't even be reproduced?! My position is don't try and make EBM say more than it actually does.

Absolutely. That's why I wanted to talk about this. Because there are some things that I agree can be taken both ways (think tPA, or steroids for spinal cord injury, or pressors for that matter) and you probably aren't wrong as far as we know. But there is data for both sides.
But then there are other things where the multitudes of data say that one way is the right way. Things like ASA for MI. Not giving antibiotics after I and D. Sterile technique for lacerations.
I'm not saying every new issue of JAMA/JACS/AEM/Lancet/NEJM should be practice changing. And yes, every paper should be looked at with skepticism. But, sometimes the guy that recommends washing your hands between deliveries might be right. Or more recently, the guy that found H pylori. And if one study begets another, and then more, and they all point the same way, then yeah, maybe what they're saying does make sense.
A favorite site of mine is The NNT because it is able to give people simple answers for certain questions about what to do or not to do. Unfortunately, it doesn't always use multiple sources in each of their statements, but they're much better than they were even a year ago.
I agree with Jerome Hoffman in that reading the literature frequently actually makes you less likely to jump in with both feet on new data.
But we should keep doing the studies, or otherwise medicine will continue trying to balance the humors.
Besides jdh, you of all people know that if we don't either refute bad evidence or try and make our own, the government will continue to make mandates about what you need to do to satisfy the bureaucracy, regardless of how "right" or "wrong" it is.

An oldie.... but absolutely a goodie.
http://www.bmj.com/content/327/7429/1459?view=long&pmid=14684649

I think this paper is good evidence for why you need to be able to understand what you're reading and make sure the study is powered to give the answer they're looking for, but far too often people use this paper as a reason to disbelieve everything published.

 

[deleted109597]

A few more recent things popped up. I was wondering if anyone here wants to weigh in on the recently evaluated but old data on using oxygen with MI, or the more recent therapeutic hypothermia data.

How many people had their hospitals jump two feet in with the hypothermia stuff, buying big expensive units and special catheters?

 

[WilcoWorld]

A few more recent things popped up. I was wondering if anyone here wants to weigh in on the recently evaluated but old data on using oxygen with MI, or the more recent therapeutic hypothermia data.

How many people had their hospitals jump two feet in with the hypothermia stuff, buying big expensive units and special catheters?

Couldn't access the link, but as an advocate of TH I'm very interested - could you please post a citation or abstract?

 

[gutonc]

Couldn't access the link, but as an advocate of TH I'm very interested - could you please post a citation or abstract?

The TL;DR version is:

TH is good, but hemodynamic stability is better. Stabilize them first, then chill them. They need to be cold, but probably not until they're settled into the ICU.

The money quotes are here:

The researchers arbitrarily chose a cutoff of two hours to distinguish early (<2 h) and late (>2 h) initiation of hypothermia. The target temperature of 34 degrees C was reached at a median of 180 minutes in the early group vs 410 minutes in the late group.

Neurologic outcomes were similar for both groups - but ICU mortality was twice as high with early vs late hypothermia: 47.4% vs 23.8%. Six-month mortality was also significantly higher after early hypothermia: 60.8% vs 40.5%.

As expected, cerebral performance category at ICU discharge and at six months was significantly better with hypothermia than without, although ICU mortality was similar for treated and untreated patients (38.8% vs 45.3%, respectively), as were mortality rates at six months.

It reminds me of the "they're not dead until they're warm and dead" trope. If you can't push the blood around at 37C, you're not going to have better luck at 34C and neurologic outcomes (the real reason for doing TH) don't matter that much in dead people.

 

[WilcoWorld]

The TL;DR version is:

TH is good, but hemodynamic stability is better. Stabilize them first, then chill them. They need to be cold, but probably not until they're settled into the ICU.

The money quotes are here:


It reminds me of the "they're not dead until they're warm and dead" trope. If you can't push the blood around at 37C, you're not going to have better luck at 34C and neurologic outcomes (the real reason for doing TH) don't matter that much in dead people.

Thanks gutonc. I should probably defer further comment until I've been able to read the actual article, but hemodynamic instability (defined as SBP<90) has been a contraindication to TH all along. If that's what this is getting at, then it's not news to me. If there's more to it, such as saying delayed TH is better than early TH... well that goes against all the pathophysiology I based my understanding on, and as such I'd better revisit the issue.

 

[deleted109597]

Couldn't access the link, but as an advocate of TH I'm very interested - could you please post a citation or abstract?

Sure thing.
http://www.sciencedirect.com/science/article/pii/S0300957211006885
As I've mentioned before, I'm not saying this is or isn't proof to do or not do what we are doing, but evidence that we need to keep studying everything, not just looking for new things. It also should serve to help stop the many instances we have had in our community of EMS or rural hospitals cooling patients prior to resuscitating them properly.

 

[ORL10]

I agree with always pushing the envelope about what we know and understand. When Manny Rivers published EGDT everyone jumped on the wagon, now we are all jumping off the wagon, and the reality is the truth likely lies somewhere in between. I think the PROcess data is going to show that the EGDT arm of the trial is better than usual care and maybe slightly better than the PSC arm, but which variable is making the patients outcomes better (maybe because we have a nurse/MD in the room titrating drips, hanging abx at the exact time they are suppose to be given, and are titrating fluids aggressively??).

As for TH, I am a big proponent. The study you gave posted is observational, which as such lacks the ability to make anything but casual relationships. Who knows the real reason why the difference in outcomes was identified. first in baseline characteristics the patients who got delayed cooling had: lower baseline SAPS II scores, and lower lactate levels (shown to be prognostic) the earlier group needed more vasopressors and more patients got catheter based cooling.

Maybe the MAP was lower in the early group because of some other factor unrelated to TH (they had a larger infarct, they had less reserve etc), and because of this they would be expected to have worse outcomes (bigger infarct, lower BP, requirements for vasopressors higher SAPS scores = poorer outcomes).


Basically this study tells me nothing other than that we need more studies. We can always tear apart even the most well designed studies, but I'm hard pressed to stop initiating cooling early in patients until someone does a RCT telling me otherwise.

-Thanks for the article