Guest Column | January 23, 2020

FDA Approves AI-Based Heart Failure Screening Method

By Kayla Matthews, Productivity Bytes


Heart failure is a growing problem in the United States. In 2017, heart failure was responsible for one out of every eight deaths in the country, with the condition costing the nation upwards of $30 billion, according to the Centers for Disease Control.

In response to this growing health issue, the FDA recently approved a heart failure screening method that relies on artificial intelligence to make its diagnoses.

What will this AI-based screening method entail, and why is the FDA's approval of this new method such a big deal in the medical community moving into this new decade?

AI Heart Failure Screening

More than 5.7 million adults are living with the shade of heart failure looming over them, and roughly half will die of the condition within five years after receiving their diagnosis. It might sound like a death sentence, but there are several treatments available to help patients manage their symptoms, as long as they get diagnosed with the condition early enough. That's where this AI screening comes in.

MedTech startup Eko, in partnership with the Mayo Clinic, is working toward creating a technology that's easy to embed in something as simple as a doctor's stethoscope, something nearly every medical practitioner on the planet carries.

When paired with an AI algorithm designed to detect heart failure, these specialized stethoscopes can collect enough data during a 15-second physical exam to determine if the patient has a reduced left ventricular ejection fraction, a symptom of heart failure. This problem would typically show up on an EKG, but those tests usually aren't part of a regular physical exam.

Researchers are teaching the AI to identify this specific ejection fraction by exposing it to the data from nearly 45,000 patients who have the same condition. When tested, the system was capable of detecting this symptom of heart failure between 85 and 86 percent of the time.

Different Aspects Of Medical AI

Eko isn't the only MedTech company seeking and receiving FDA approval for AI-based diagnostics for cardiac health conditions.

eMurmur, created by an Ottawa-based company, is also using a specialized stethoscope and a machine learning program to identify and classify heart murmurs, depending on whether they're normal or dangerous. This program also has received FDA clearance and may help both cardiologists and general practitioners diagnose dangerous heart murmurs.

In time, this kind of AI-based program could also evolve, appearing in other aspects of the medical industry, as well as in medical-adjacent applications, such as during the process of performing an Impairment Rating Evaluation for workers' compensation and disability assessments. While it will likely never replace the human aspect of this kind of exam, it could complement human workers and perhaps make their job a little bit easier in the long run. It could also help make the assessment process more objective and fairer.

The First Of Its Kind

The most exciting thing about this new AI-based heart failure detector isn't just the fact that it will save lives, though that is part of it. It's the fact that it's one of the first of its kind, an AI system that has received FDA approval to predict heart failure. While it isn't the first AI in the medical industry, it is at the forefront of the application of this technology. As of mid-2019, there were only five FDA-approved AIs.

The technology is taking off and will likely become an integral part of the medical industry. One of the advantages of AI in the medical sector is that there is a massive amount of data available for this tech to use.

In 2011, researchers estimated that the healthcare system in the U.S. alone had generated more than 150 exabytes of data. For comparison, one exabyte is one billion gigabytes. By 2019, that number had grown to 8.41 petabytes. It's skyrocketed by 878 percent since 2016, and that number will only continue to increase.

AI and machine learning systems can sort through all this data in a fraction of the time it would take human researchers, comparing information both past and present to find a diagnosis, create accurate predictions or help save lives.

Medical AI Has a Promising Future

Today is a thrilling time to be part of the medical industry. AI technology is just beginning to emerge as a practical tool for diagnostics and analysis. There are only a handful of medical AI programs currently available, but it won't stay that way for long.

Medical AI will continue to grow and advance as the technology does, and it likely won't be long before artificial intelligence becomes an invaluable diagnostic and analysis tool.

The exciting thing about AI is that the more data it has, the smarter it becomes and the more useful it will be. This tech is only in the earliest stages of demonstrating its full potential to the world.

About The Author

Kayla Matthews is a MedTech writer whose work has appeared on HIT Consultant, Medical Economics and HITECH Answers, among other industry publications. To read more from Kayla, please connect with her on LinkedIn, or visit her personal tech blog at