How accurate are EKG readings by computers?

[Dock1234]

Advertisement - Members don't see this ad

I'm premed and little confused about this. Some people say that they are awful and some say that they are better than senior cardiologists at reading EKGs. So how is it really?

 

[gutonc]

Neither of those two extremes.

 

[nlax30]

Generally they're fairly decent. Since they're computers they are good at calculating things such as intervals, QTc, axis and I usually trust those interpretations.

The trouble comes when you're interpreting whether those anterior ST segments are repolarization abnormalities or actual ischemic ST elevations, which the computer is not as good at.

 

[Raryn]

In my (albeit as a noncardiologist, somewhat limited) experience, they seem to tend to overcall rather than undercall. As in they have a pretty decent sensitivity but the specificity is pretty crappy.

 

[sacrament]

some say that they are better than senior cardiologists at reading EKGs

Who in the world says that? That's just absurd.
I agree with the comment that they are more sensitive than specific. If the computer interpretation is that the EKG is normal, then it is almost certainly normal. It tends to be fairly abysmal when it comes to rhythm recognition. ST changes are hit-and-miss.

 

[Packman2003]

Their major weak spot is with arrhythmias in my experience.

I really think you can't trust the interpretation much except for numbers like rate, intervals.