New York residents may eventually have their heart disease diagnosed long before any dangerous events such as heart attacks take place. Researchers have been using electronic medical records, GPUs and artificial intelligence to predict heart failure up to nine months earlier than doctors are able to do so.
Approximately 6 million Americans have heart failure annually. Earlier detection would mean that doctors could prescribe medicine or suggest lifestyle changes that could delay or prevent the disease.
The research team consists of scientists at the Northern California-based Sutter Health and at the Georgia Institute of Technology. The project required computers to do complex analysis of records such as prescriptions or summaries of doctor visits that were in varied formats as well as tracking patient records over time. This required teaching the machines to learn rather than having a human define all the factors the computer should consider.
Other organizations and companies are working toward similar results. The technology company NVDIA has partnered with Massachusetts General Hospital to use AI in improving disease detection. The researchers on the heart disease project say that the same technology can be used to detect other ailments and that their next focus may be sepsis. This infection may occur in a hospital setting and is often not detected in time.
An infection like sepsis might happen as the result of surgical errors. Furthermore, there may be cases in which failing to identify it in time is because of negligence. A patient who believes that an infection has occurred as a result of an error or that a diagnosis has not been made on a timely basis may want to meet with an attorney to learn if any recourse is available.