Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

This is interesting, but more because it indicates that there's adequate data in a single heartbeat to do such diagnosis. In practical terms it's probably not nearly so relevant because it sounds like they were working with the raw data not tracing. By the time you have a patient hooked up to the proper equipment to do this diagnosis you're going to be getting adequate data anyway.

The main impact might be that if this holds up people could be tested with a short hook up in an office instead of with a 24-hour monitoring where they have to bring back a Holter device the next day. Of course, that 24 hour dataset may have independent value of its own for further diagnostics beyond just whether the patient has CHF.



The study is not worth paying attention to.

The datasets for positive cases and negative cases come from different databases. n=30 patients, on top of it.

All this does is recognize the patient/ECG technician who recorded the data. It's basically certain it doesnt generalize




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: