Artificial intelligence (AI) models can be trained to identify warning signs of sudden cardiac death (SCD) in electronic health records (EHRs), according to new research to be presented at Resuscitation Science Symposium 2023, a two-day international meeting presented by the American Heart Association (AHA).
The advanced AI algorithm was designed to detail a patient’s risk of SCD based on a long list of factors, providing a snapshot of their overall heart health. If the patient’s EHR said they had been diagnosed with hypertension, for instance, the AI model would highlight this detail and explain just how much this increased their risk of SCD within a specific timeframe.
The AI model was built using data from more than 12,000 SCD cases and a matched control group of patients who did not experience SCD. The model was then validated with data from more than 11,000 additional SCD cases. Overall, the model was able to build personalized SCD prediction models that achieved an area under the curve (AUC) of 0.80, a positive predictive value of 77% and a sensitivity of 68%.
“We have been working for almost 30 years in the field of SCD prediction, however, we did not expect to reach such a high level of accuracy,” lead author Xavier Jouven, MD, PhD, a professor of cardiology and epidemiology at the Paris Cardiovascular Research Center, said in a statement. “We also discovered that the personalized risk factors are very different between the participants and are often issued from different medical fields (a mix of neurological, psychiatric, metabolic and cardiovascular data) – a picture difficult to catch for the medical eyes and brain of a specialist in one given field. While doctors have efficient treatments such as correction of risk factors, specific medications and implantable defibrillators, the use of AI is necessary to detect in a given subject a succession of medical information registered over the years that will form a trajectory associated with an increased risk of sudden cardiac death.”
For AI-powered risk assessments to work, the researchers noted, consistency would be required from one EHR to the next; an algorithm trained to extract data from one resource may not be reliable if using input from a completely difference resource.
Read the full abstract here.
Resuscitation Science Symposium 2023 is scheduled for Nov. 11-12, 2023, in Philadelphia. Click here for more details.
Michael has more than 16 years of experience as a professional writer and editor. He has written at length about cardiology, radiology, artificial intelligence and other key healthcare topics.