Standardized STEMI Identification Assessment

This test motivated me. Thank you.

Is it okay to pass this on to my nurse wife and her nurse friends?
 
I'm leaving this test up on some of our company computers to see if anyone else will take it
 
560 folks have taken the assessment (now 561...it goes up up up), and really not much has changed!

Acc: 67%
False Positive rate: 11%
False Negative rate: 22%
Sensitivity: 67%
Specificity: 66%

That's almost identical to when it was just 40 people!

Some interesting subgroup analysis though:
  • Self identified NP/PA's as a group scored the highest overall (n=11). 72% accuracy, 10% FP rate, 19% FN rate.
  • Attendings/GPs were not far behind (n=13). 71% accuracy, 12% FP rate, 16% FN rate.
  • Residents and Paramedics are nearly indistinguishable, with the residents holding a slight edge in false negative rates (21% vs 22%).
  • The "best" scorer group (>90%) have at least one: MD, resident, NP/PA, med student, RN, paramedic, EMT, and technician (nobody non-medical). No one specialty dominates among the top performers.
  • When we ignore the license/cert/card/title, reading at least 1 ECG a day keeps the false positives and negatives away.
 
Your accuracy was: 83 %
Your sensitivity was: 92 %
Your specificity was: 67 %
You had 4 false positive(s) and 2 false negative(s).
 
So wait..explain the st-depression again. This was a big source for debate in my medic class. So how I was taught is that if there is depression in II, III or avf there will be elevation in one of the other set of leads and vice versa (that was how they explained reciprocal determinism). Explain the diffuse ischemia version of depression then to me. It was also taught that all you need is 1mm box to activate the cath. lab and that it had to be in two consecutive leads. (ie if you had elevation in V1 it had to be in V2 as well).
 
Other than reciprocal changes, it's traditionally taught that ST depression can also signify subendocardial ischemia (in other words, not a complete STEMI, just ischemia). This is true, but there's a misconception that this can be localized on the ECG in the same way as ST elevation. True subendocardial ischemia is usually widespread, or at least not localized anatomically. An anatomical pattern of depression (e.g. inferior leads or anterior leads) is more likely to indicate reciprocal changes.
 
So ischemia will show up in like....V1-V4 as opposed to just V1 and 2? If it just shows up in V1 and V2 I am looking for elevation in II, III, and aVF?
 
So ischemia will show up in like....V1-V4 as opposed to just V1 and 2? If it just shows up in V1 and V2 I am looking for elevation in II, III, and aVF?

All of the precordials are somewhat contiguous, although it's unusual for a pattern to stretch all the way across. But yes, compare for reciprocals in the inferiors and vice versa (or hidden elevation in the posterior wall).

As for mere ischemia, it's not uncommon to see it in practically every lead. This is a common finding in stress tests as well as tachyarrhythmias (demand ischemia -- heart's working harder than the blood supply can oxygenate). Even if it's not global, you won't see reciprocal elevation anywhere.
 
Good Lord, Christopher! There's not a straight forward ECG in that entire test. Each was was difficult, IMHO. I would not feel comfortable calling any of them a STEMI without clinical correlation. My accuracy: 83% with 2 false positives and 4 false negatives (I answered conservatively). You could see many of these STEMI-like ECGs with other diagnoses (e.g., sepsis, etc.). I think paramedics and physicians alike would have extreme difficulty with this test.

Tom

All cardiologists say they would rather get a false positive, than one that isnt called in. We saw one we thought was a stemi, called it in, got to the hospital, it was pericarditis. Doc said he would want us to call that in as a stemi 10/10. That being said, it's pretty embarrassing when you call in something obviously not a stemi.
 
All cardiologists say that until they are called in at 0300 for a paced rhythm or because artifact triggered the ***ACUTE MI SUSPECTED*** message on the prehospital 12-lead ECG. Pericarditis is a special case because it's rare and widespread ST-elevation could also be a massive STEMI. Every system has a false positive rate but very few hospitals will speak about it publicly. In many hospitals it's in the 30% range (I'm not talking about paramedic-activated STEMI cases in particular). Having false positives, by itself, is nothing to be ashamed of. OTOH, there are false positives (and false negatives) that simply should not happen, but do.
 
Due to its insane popularity, I've begun work on vNext of the assessment (sneak peek pics are here). The images have already been moved to a content delivery service, and I'm moving off Google Apps to an actual website with an actual database*.

If you included your email in the first go-around, I'm going to work on migrating your results to the new version. I haven't figured out exactly how to handle the account sign-up bit...but likely you'll get an invitation email with a magic link and stuff. Tentatively the site will allow Guests...but who knows!

The closed beta will include just the McCabe assessment, so you won't get a chance to bust your chops on any new ECG's.

Again, thank you for all of the support!

* for those interested: Amazon Web Services is hosting the server, an EC2 instance running Debian Wheezy. Site served via nginx, proxying dynamic requests to node.js+express.js 4, dynamic client side scripting with Jade templates and Knockout.js, CSS is via Twitter's Bootstrap. Data layer housed in a combination of REDIS and MySQL (Amazon ElastiCache and RDS respectively). Static content served via Amazon's CloudFront + S3.</nerdalert>
 
That is awesome dude
seriously cool stuff

Due to its insane popularity, I've begun work on vNext of the assessment (sneak peek pics are here). The images have already been moved to a content delivery service, and I'm moving off Google Apps to an actual website with an actual database*.

If you included your email in the first go-around, I'm going to work on migrating your results to the new version. I haven't figured out exactly how to handle the account sign-up bit...but likely you'll get an invitation email with a magic link and stuff. Tentatively the site will allow Guests...but who knows!

The closed beta will include just the McCabe assessment, so you won't get a chance to bust your chops on any new ECG's.

Again, thank you for all of the support!

* for those interested: Amazon Web Services is hosting the server, an EC2 instance running Debian Wheezy. Site served via nginx, proxying dynamic requests to node.js+express.js 4, dynamic client side scripting with Jade templates and Knockout.js, CSS is via Twitter's Bootstrap. Data layer housed in a combination of REDIS and MySQL (Amazon ElastiCache and RDS respectively). Static content served via Amazon's CloudFront + S3.</nerdalert>
 
That is awesome dude
seriously cool stuff

Thank you!

For those technically inclined here is a picture of how the sausage is being made.
overengineered.png
 
And for those who are still following this thread, I'm trialing an ECG viewing feature and would like user input on its "look and feel"...and whether my hackjob CSS/JS actually works :)

http://s3.amazonaws.com/ecgs/sp/lightbox.html

I'm begging, borrowing, and stealing the look/feel that Wikipedia has moved to for images.

Also of interest is how this looks on tablets / phones of various types/etc.

Any bugs you may find, I'd love to know about:
- I already know that image zooming on phones gets a bit odd...I'll probably provide a fullscreen mode for that.

Thank you!
 
Took this in class. There were definitely some challenging ones in there. Thanks
 
Took this in class. There were definitely some challenging ones in there. Thanks

Thank you for the support. vNext is approaching rapidly. If you used your email address you will have an account waiting for you in the new system.
 
this is an outstanding little quiz. I remember attempting the very same assessment about a year ago, when a fellow student recommended it in our cardiology class. I don't recall how I fared the first time—dreadfully, I'm sure. Today I scored with 89% accuracy [96% sensitivity/75% specificity] from the 36 mentally-straining ECGs, and I'm cool with that.
 
  • Your accuracy was: 78 %
  • Your sensitivity was: 67 %
  • Your specificity was: 100 %
  • You had 0 false positive(s) and 8 false negative(s).
================================

Not very happy about the 8 false-negatives :(
 
So I've had a chance to go back through my results, and now I'm a little happier. The ECG I thought was Wellen's syndrome is a good beat. I missed at least one very blatant STEMI, I'm still not sure how. My biggest downfall was failing to consider absolutely small but proportionately large ST changes in the inferior leads, and I obviously have a need to be a little more systematic, and less reliance on strict reference values.

This is an excellent teaching / learning resource. Thanks for producing this!
 
Back
Top