Heart Health

Are you ready for a computerized cardiologist?

Published 4.19.2017
Machines with artificial intelligence (AI, or in other words, a complicated software program) can "teach" themselves how to diagnose illnesses including heart disease. However, humans do not understand how they do this. That’s great so long as the outcome is positive, not so great when (not if) mistakes or misdiagnoses result.

Basically, they feed a ton of data into the machine, then tell it to look for associations. The historical data includes whatever disease or affliction the patient developed in the end. Then using those associations, newer data with no end point is fed in and the software is asked to predict who is going to have a heart attack or stroke (or other dread affliction).

Because the machines can consider many more associations than human doctors can, they can be more accurate at predicting outcomes. The machines are not limited to the associations that human doctors use, which is both an advantage and a risk factor.

All four AI methods performed significantly better than the ACC/AHA guidelines. Using a statistic called AUC (in which a score of 1.0 signifies 100% accuracy), the ACC/AHA guidelines hit 0.728. The four new methods ranged from 0.745 to 0.764, Weng’s team reports this month in PLOS ONE. The best one—neural networks—correctly predicted 7.6% more events than the ACC/AHA method, and it raised 1.6% fewer false alarms. In the test sample of about 83,000 records, that amounts to 355 additional patients whose lives could have been saved. That’s because prediction often leads to prevention, Weng says, through cholesterol-lowering medication or changes in diet.

Again, humans do not know how the machine is making the determination. This is promising technology, computers should be used in medicine. However, caution should be urged before surrendering entirely to our robot overlords.

In other heart disease related news, a new guideline to decide who needs to be cut open and have their aortic valve replaced was adopted. The primary change seems to be recognizing that different sized people will have different sized aortas, so taking into account the height of the patient is a good idea.

Not a medical professional and I don’t play one on the internet, but really? How is it that patient size is always part of the calculus? Sex was also not included in the existing guideline. I can’t pretend not to be both mystified and horrified by this knowledge. Even post menopause female hormonal systems are not the same as those of males. Yikes. More reasons to stay out of the cardiologist's (computerized or not) office.