McCabe JM et al. Physician Accuracy in Interpreting Potential ST‐Segment Elevation Myocardial Infarction Electrocardiograms J Am Heart Assoc. Oct 2013; 2(5): e000268. doi: 10.1161/JAHA.113.000268
The full text of this article is available here: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3835230/
A little bit of background about the article first, and then the fun part is down a bit further.
About the Article
Using a set of 36 ECGs that had previously resulted in putative STEMI diagnoses, a group of 124 emergency physicians, cardiologists, and interventional cardiologists were asked: “based on the ECG above, is there a blocked coronary artery present causing a STEMI?”. The reference standard for ascertaining the STEMI diagnosis was subsequent emergent coronary angiography.
The sensitivity to identify “true” STEMIs was 65% (95% CI: 63 to 67) and the specificity was 79% (95% CI: 77 to 81), with a poor overall inter-observer agreement (kappa = 0.33). Notably the computerised diagnosis had a sensitivity of only 48%, but a specificity of 83% (so if it says it's there, look carefully!).
Emergency physicians had the highest sensitivity (74%), but lagged behind the cardiologists in terms of specificity. This likely represents a preference to "overcall" a STEMI than miss one. The comparison of sensitivity, specificity and positive and negative predictive values are shown in this table:
The Fun Bit - Try it Yourself!
@AmboFOAM has provided a link to an online version of all the ECGs used in the study, allowing people to test their accuracy in identifying STEMIs. You can have a go at it here. This page will then provide you with your accuracy, sensitivity and specificity so you can compare it with the results from the study.
I did... and to add extra realism, I made it a little more like working in the Emergency Department by doing so in the presence of loud distractions and constant interruptions (just a standard day with the family). I was a bit disappointed with my initial accuracy of 67% - I had good sensitivity (72%) but poor specificity (58%), so at least that was being reasonably safe. Given that, I sat down again tonight in quieter conditions to do them, and was quite happy with my overall accuracy of 83%.
