Researchers are using artificial intelligence to better predict heart failure. The process has the potential to identify at-risk patients in New York and across the country up to nine months earlier than doctors.
Researchers at Sutter Health and the Georgia Institute of Technology used a method called deep learning and combined it with graphics processing units, or GPUs, to analyze the electronic health records of hundreds of thousands of Sutter Health patients. From those, it examined 3,884 heart failure patients and 28,900 control group patients. Deep learning is a highly sophisticated type of artificial intelligence that can learn complex technical patterns. It is also very time-intensive, sometimes taking years to build a model. However, the team provided powerful GPUs that allowed the AI to analyze records at a significantly faster rate. By analyzing and understanding complex variables in health records, such as prescriptions and doctor visit summaries, the AI could predict heart failure well before doctors diagnosed it. The researchers said their model can be applied to other diseases. Their findings are being considered for publication in a medical journal.
According to the American Heart Association, approximately 6 million Americans suffer heart failure each year. Heart failure occurs when the heart muscle becomes too weak to pump enough blood and oxygen to the body. Around 50 percent of patients diagnosed with the disease die within five years.
A doctor’s misdiagnosis of heart disease and other health problems may result in a worsened condition. Patients who have been in this position may want to obtain the opinion of an attorney as to whether such mistake constituted compensable medical malpractice.
Source: Nvidia, "Change of Heart: How AI Can Predict Cardiac Failure Before It’s Diagnosed," Jamie Beckett, April 11, 2016