Can Paramedics Recognize STEMIs?

If I show this 12-lead to a group of paramedics, I would wager close to 100% would accurately identify it as an inferior MI and call in a STEMI alert if the 50 year old male patient in front of them was clutching his chest.

Now, if I have the same patient and the 12-lead shows this;

How many paramedics are going to say this is a STEMI?

It is not a STEMI, it is a Left Ventricular Hypertrophy.

In a recent  study published in the April 2013 issue of Prehospital Emergenncy Care, 63.3% of paramedics identified this ECG as a STEMI.

The study, PARAMEDIC ABILITY TO RECOGNIZE ST-SEGMENT ELEVATION MYOCARDIAL INFARCTION ON PREHOSPITAL ELECTROCARDIOGRAMS, by Mencl, et al, goes against the findings of prior studies that have shown paramedics can identify STEMIs with a high degree of accuracy. The limitations of the other studies have been that they have used paramedics right after training classes and in many cases had them evaluate only a small number of relatively obvious ECGs. The 12-leads used in this study I believe offer a more realistic test of a paramedic’s interpretation abilities.

In this study they had 472 paramedics from 30 different EMS services in five counties with 15 different medical directors read 10 different ECGs and were asked to either identify the strip as a STEMI or not.

Here’s how they did:

The number indicates the percentage of paramedics who correctly identified the rythm as a STEMI or not a STEMI.

Inferior STEMI 96.0%, Anterior STEMI 78.0%, Lateral STEMI 51.1%

Normal 1 97.3%, Normal 2 100%

RBBB 79.2%, LBBB 39.0%, Ventricular pacing 52.8%, LVH 36.7% , SVT 65.3%

The study concludes: “Despite training and a high level of confidence, the paramedics in our study were only able to identify an inferior STEMI and two normal ECGs. Given the paramedics’ low sensitivity and specificity, we cannot rely solely on their ECG interpretation to activate the cardiac catheterization laboratory. Future research should involve the evaluation of training programs that include assessment, initial training, testing, feedback, and repeat training.”

I took the 10 ECGs from this study and gave them as a test to a large group of paramedics ranging from twenty-year veterans to new paramedics and found similar results.  These results square with my observations of paramedics bringing possible STEMI patients into our hospital.

Here’s what I take from it:

Paramedics can identify inferior STEMIs with a high degree of reliability.  Lateral and Anterior STEMIs can be more difficult.  Left bundle branch and LVH can particularly cause confusion in some paramedics.

My experience with field STEMI alerts is that paramedics tend to by shy about calling for STEMI alerts, and thus have fewer false activations than this study, if extrapolated,  would suggest.  In other words, if a paramedic is sure it is a STEMI, such as an inferior, the paramedic will call in a STEMI alert.  But while the paramedic might guess a 12-lead is a STEMI such as the LVH ECG above, the paramedic might be less likely to call in a STEMI alert with it. 

Despite this well done study, I continue to support field activation of the cath lab through paramedic interpretation alone, but I do believe we need to increase our 12-lead training efforts.

I believe that with frequent training and regular competency checking, paramedics can indeed identify STEMIs with a high degree of reliability.  But if you don’t do the training and keep it up, the results will not be as high.

In a future post – What about computer interpretation?


  • Christopher says:

    It really boils down to how we teach 12-Lead interpretation. Unfortunately it usually starts with “≥1mm ST-E in 2+ contiguous leads in the absence of frank confounders” and they work back from there.

    In reality you need to work from the other direction and only get to STEMI as a “protocol” thing, because ST-elevation myocardial infarction occurs regardless of any millimeter criteria imposed by systems.

    The class I came through stressed solid interpretation of the 12-Lead, including identification of LBBB, LVH, RVH, Early Repol (well we called it Benign Early Repol), Pericarditis, hyperkalemia, etc. When I see failings in 12-Lead interpretation of STEMI, it is usually because the impetus was on finding elevation rather than understanding the 12-Lead.

    The findings in this study are not surprising and my own service area has seen a recent uptick in “false positives”. Most of these have been what we know as frank confounders. Often times I hear bandied about “treat your patient not your monitor” as for why a “STEMI Alert” was called on an ECG that did not meet criteria or was a confounder that mimic’d the criteria necessary, and this is a problem on its own.

    This study is good, in that we know what we need to focus on: better basic education on 12-Lead ECG’s!

    • medicscribe says:

      Hi Christopher-

      Thanks for commenting. You make a great point about how we should be teaching 12-leads instead of how we do. I will keep this mind for future classes.

  • Brooks Walsh says:

    This is an intriguing study, and it suggests a corollary question: How do you construct a “fair” test of ECG reading skills?
    For example, how many “normals” or mimics do you put in a test? In the real world, a medic sees a large proportion of normal-ish tracings for every STEMI. I wonder if the medics’ reading skills appear better in “the real world,” where there’s a lot more chaff than wheat.

    A test like this, on the other hand, focuses on the medic’s ability to differentiate STEMI patterns from mimics, and would tend to magnify any mild deficit in ECG interpretation.

    Add on top of this that the example of LVH is not entirely classic; e.g. while you have strain pattern and aVL > 11 mm, the precordial leads don’t show diagnostic S or R wave amplitudes. Even the “4-step” algorithm suggested by Brady would have a tough time with that ECG! (

    I look forward to hearing your thoughts about the utility of teh computer!

    • medicscribe says:

      Thanks for the comments, Brooks. I enjoyed reading about the Brady method on your site. I think perhaps a better way to have done a 12-lead study is to add an additional option for the paramedics of “Possible STEMI.” As we know the real like ECGs are often very borderline, and the correct answer may not be “Definite STEMI” activate the cath lab or “Not a STEMI”, no need for alarm, but “Possible STEMI,” maybe not activate the cath lab, but give all a heads up that there might be a case coming in. That would best mirror what we are actually asking paramedics to do.

      A preview of my thoughts on the computer interpretation: can be helpful, but my experience is it not nearly as reliable as studies claim. I will have some stats to illustrate my point.

  • DaveB says:

    I agree with Christopher that the way 12 leads are generally taught in medic class is lacking.

    First, ST elevation=STEMI is overemphasized to such a large degree that i doubt most medics do not know STEMI is not even the leading cause of ST elevation. Recognizing which ST elevations are STEMI is far easier to teach than which ST elevations are not STEMI, such as LVH, BER, normal variant, etc.

    I believe that while limited, this study reflects the general state of prehospital 12 lead education.

Leave a Reply to medicscribe Cancel reply

Your email address will not be published. Required fields are marked *