Tuesday, October 9, 2018

Another Inadequate Paper Published on Triage ECGs, whose Conclusions Need Scrutiny.

This paper was just published:

Noll S. et al.  The utility of the triage electrocardiogram for the detection of ST-segment elevation myocardial infarction.  Am J Emerg Med 36(10):1771-1774.  October 2018.


In this paper, in a department in which they state they have only 50 STEMI per year, they looked at only 8 days worth of triage ECGs for a total of 538.  They did not find one STEMI on their triage ECGs (no surprise! They only get 1 per week!).  They did have 16 NonSTEMIs, but do not describe the ECG of these 16, nor even investigate the consequences of the ECG in these 16.  "While in the ED, one patient progressed to have ECG changes consistent with STEMI although the initial triage ECG did not meet STEMI definition guidelines."  It would be instructive to see that ECG: perhaps it showed an OMI, but not STEMI, just like the one below which was recorded just a couple days ago in triage.  They did not look to see how many of the NonSTEMI had an occluded artery at angiogram.  As far as I can tell, they did not view the ECGs!  They only looked at ED diagnosis, not at any angiogram or even at discharge diagnosis.  They do a "cost analysis" based on charges of $125 per ECG and state that it would cost $54,000 to detect 1 STEMI based on 50 STEMI per year.  

Conclusion: "Given the extremely low yield and high associated charges, current guidelines for triage ECG for identifying a possible STEMI should be reviewed."

Fair enough.  I will briefly review the guidelines here and now: 

One should never use charges to calculate cost.  Charges and cost have no relation to each other in hospital billing.  In our ED, a health care assistant (HCA) records all ECGs, in triage and elsewhere.  It takes at most 10 minutes (this is an exaggeration).  At total compensation of $50,000 per year, working 1800 hours, an HCA could record over 10,000 ECGs if that is all they did.  That is $5.00 per ECG in cost.  Let's double it to be certain, to $10.00.  Then it takes a staff physician all of 10 seconds to read it (at most!).  Is that $125.00 of cost?  No, again, charges and cost have no relation to each other in hospital billing.  One must add in the cost of paper, and the system, and the ECG machine, all of which are negligible per ECG.  

Let's say, then, that the true cost of an ECG is $20.00 (this is an exaggeration -- it would cost even less).  We at HCMC have 30 walk-in STEMIs per year; the rest come by ambulance.  We also detect at least 5 OMI patients who do not meet STEMI criteria. We record ECGs in triage on every patient with chest pain, and some other indications, and this amounts to 8000 ECGs in triage each year, costing at most $200,000 (8000 x $20.00).   That is 35 OMI detected out of 8000 triage ECGs; that is 1 OMI for every 229 ECGs, and 1 OMI for every 10.4 days.  Thus, it costs us at most $5700 to detect one OMI.  Given the dire consequences of missing a STEMI or OMI, including cardiac arrest (see cases below), $5700 is extremely cheap.  

Saving just one person from death or heart failure by early diagnosis of STEMI is worth far more than $5700, or even than $54,000, or even than $200,000.  Moreover, the costs of litigation of just one missed STEMI or delayed diagnosis, with subsequent severe heart failure or death, can run into the millions.  

By recording triage ECGs, we make the early diagnosis of OMI in not just 1 patient, but 35 patients, per year.  

That the triage ECG must be shown to the physician is demonstrated again by the following ECG, recorded just this week.

A patient with chest pain:
See the computer analysis.
Is it normal, as the computer says?
Notice there is only trace ST Elevation in several leads 
(a normal amount!).

For those of you who read this blog regularly:

You will know that these are clearly hyperacute T-waves, diagnostic of proximal LAD occlusion.

This was immediately recognized by the physician and the cath lab was immediately activated. 

V2 is actually a de Winter's T-wave.  

Cath lab needs activating.

Imagine if this patient had no triage ECG and had to wait 2 hours for placement into the ED.

Imagine if the computer read of "normal ECG" prevented it from being seen by the MD.

For those of you who are new to this blog, see these cases:

10 Cases of Anterior Hyperacute T-waves in V2-V3

10 Cases of Anterior/Lateral Hyperacute T-waves in V4-V6

10 Cases of Inferior Hyperacute T-waves

Missed hyperacute T-waves followed by death

Hyperacute T-waves that never manifested STE despite serial ECGs with total anterior wall infarction

30 year old with hyperacute T-waves diagnosed prehospital

Missed hyperacute T-waves followed by cardiac arrest during discharge

Hyperacute T-waves called "normal" by computer

Another prehospital hyperacute T-wave case

Computer "Normal" ECGs in Triage:

A case of arrest:

Chest pain relieved by Maalox and viscous lidocaine

Another case of arrest:

Another case of arrest:

A 50-something woman with chest pain and 2 "normal" ECGs at triage

A middle-aged woman with chest pain and a "normal" ECG in triage

Chest Pain Diagnosed as Gastroesophageal Reflux

An Elderly Male with "Indigestion"

Comment by KEN GRAUER, MD (10/9/2018):
I agree with Dr. Smith — This is a faulty study for many reasons. In addition to the highly problematic issue of “cost analysis” (What is a fair and reasonable “true cost” — and what is the “value” of stemi detection in patients with chest pain?) — the methodology used indicates this was a retrospective chart review of “all patients seen in an ED over an 8-day period of who had a triage ECG performed”.
  • The intervention that should have been assessed (in my opinion) — is the potential benefit (if anyof doing a triage ECG (ie, of doing an ECG within 10 minutes of arrival when the presenting complaint is for symptoms of concern for possible stemi).
QUESTION: Isn’t there “potential benefit” — IF by doing an ECG on a patient who presents to the ED, a capable clinician decides that the patient needs to be formally assessed sooner than simply “waiting his/her turn” in a busy ED?
  • To do such a study — data must be prospectively obtained. You simply have NO idea as to how much may have been missed by retrospective analysis.
  • As per Dr. Smith — we have NO idea how many OMIs may have been missed that didn’t quite satisfy “stemi” criteria (like the triage ECG that Dr. Smith posts above).
  • In addition to looking for STEMIs and OMIs — there are other potential ECG findings that may merit expediting formal evaluation by a clinician in the ED.
  • Since the intervention being studied involves ECG interpretation — more than a single physician should interpret each tracing, with allowance for how to resolve discrepancies in interpretation.
BOTTOM LINE: I believe the wrong question was asked in this study. In addition, the wrong methodology was used. I don’t think any valid conclusions can be drawn from the results.
  • COMMENT: I find it hard to imagine how there cannot be benefit from ensuring that a triage ECG is promptly performed (ie, within 10 minutes of arrival) for patients who present to an ED with symptoms concerning for possible acute STEMI.
  • A study is not needed to determine potential benefit of parachutes for those who jump out of airplanes. Do we need to study if a triage ECG on patients who present to an ED with acute symptoms is a reasonable concept?
  • P.S.: Our very next case ( = on October 14, 2018) — shows the value of having a capable ED physician promptly review the Triage ECG!


  1. 20$? Its true cost is less than 1$. But in this situation we must not even think about costs...

    1. One must always consider cost. Health systems are unsustainable if you do not, and then everyone suffers. The true cost is probably $10.00 because of the cost of the tech doing the recording and the machine, its upkeep, the data center, etc.

  2. What is also questionable is that they calculating costs as if this was something they otherwise didn't need to do. Patients presenting with symptoms concerning for STEMI will obviously one way or another get an ECG taken some time during their hospital stay, so the cost for this intervention is of course not the cost for taking so and so many ECGs according to their billing standards, but the extra work needed to do it faster in triage.

    Their conclusion pretty much requires them also saying that patients with chest pain wont need an ECG due to low yield, so we just send them home anyway. Otherwise, we will just need to do the same ECG later which is billed the same amount, but might require more work per ECG if process is more streamlined in triage compared to us having to say that we want an ECG for every patient with concerning symptoms, requiring my nurses in the main ED having to pause other things they are doing to get the equipment to get a single ECG on one patient.

    We obviously do not repeat the ECG on every patient if taken in triage on patients with low risk for ischemia, i.e. healthy 24 year olds with atypical chest pain .


Recommended Resources