A recent research points to the ability of artificial intelligence in studying electrocardiogram (ECG) test results. ECG tests help in diagnosing irregular heartbeats that can be dangerous for patients. These tests also help in predicting the possibility of a patient to die within the next year. The aforementioned assertions are a part of a study by the American Heart Association. The presentation of the study is due at the Scientific Sessions of the Association from 16th-18th November.
Accuracy of Model
Researchers used over 2 million ECG test results spanning over a period of three decades. They used archived medical records stored in the Geisinger Health System of New Jersey. These records were then used to train and develop neural networks with advanced computational networks. The new study is the first to predict future health events using artificial intelligence and ECG tests. Previously, existing health problems were used to predict the occurrence of any other medical complication. A senior author of the study considers this research to be a revolution in healthcare diagnosis and testing. Professor Brandon Fornwalt believes that physicians could now work in tandem with computers to improve test results.
Use of Deep Learning Technologies
The researchers have also conceptualised a deep learning model to predict irregular heart rhythms. Hence, patients suffering from atrial fibrillation could be hopeful of a healthier future. Moreover, the model can also help in predicting the vulnerability of individuals to atrial fibrillation before it develops. The occurrence of atrial fibrillation increases the chances of heart attacks and strokes. Therefore, it is important to control and monitor the rhythmic movements of the heart. Currently, there are no significant methods to predict the occurrence of AF in individuals.