Dive Brief:

• Abbott said Tuesday it has developed an algorithm using artificial intelligence that could help emergency room doctors determine if a patient is having a heart attack.
• The algorithm is the first created through machine learning to combine high-sensitivity troponin blood testing with other patient details such as age and sex, with the aim of providing more individualized information to inform the diagnosis, Abbott said.
• A study published in the journal Circulation found the algorithm gave doctors a more comprehensive assessment of the probability a patient was having a heart attack, especially for those who entered the hospital within the first three hours of their symptoms starting.

Dive Insight:

Patients who enter the emergency room with symptoms of a heart attack typically receive an electrocardiogram and troponin blood tests at set times. Cardiac troponin is a protein released into the blood when the heart is damaged.

Abbott said the algorithm goes beyond a one-size-fits-all approach to diagnosing heart attacks by taking more personalized factors into account. For example, the algorithm takes into consideration that women may not produce as much troponin protein as men.

In the Circulation study, researchers from the United States, Germany, the U.K., Switzerland, Australia and New Zealand analyzed data from more than 11,000 patients from nine countries to test whether the AI technology could provide a faster and more accurate heart attack diagnosis.

The research found the algorithm, called the myocardial ischemic injury index, performed better than the European Society of Cardiology rule-out pathway. The European Society of Cardiology rule-out pathway includes a rapid assessment algorithm based on high-sensitivity cardiac troponin and sampling guidelines at timed intervals to rule out heart attacks.

The Abbott-developed algorithm is more versatile than existing algorithms because it doesn't depend on fixed cardiac troponin thresholds, doesn't require testing to be performed at specific time points and recognizes different healthcare systems have different risk tolerances, the researchers said.

The study concluded the algorithm's individualized, objective assessment of the likelihood of myocardial infarction could be used as a tool for identifying low- and high-risk patients who may benefit from earlier clinical decisions.

The algorithm is used for research only and isn't now commercially available, Abbott said. The company has applied for an international patent on use of the algorithm.