“AI will not replace clinicians but clinicians who use AI will replace clinicians who don’t.”

Dr Bertalan Meskó, PhD

Join us in recapping the latest discussion on enabling smarter, artificial intelligence powered cardiac screening to optimize program quality, physician workflow efficiency and identify patients at-risk for sudden cardiac arrest.

The exclusive webinar, hosted by Vizient and Mpirik, included a panel discussion with subject matter experts Dr Sanjaya Gupta (Saint Luke’s Mid-America Heart Institute), Dr Julian Booker (University of Alabama at Birmingham), and Mpirik’s CEO Logan Brigman.

Key highlights from the panel discussion:

The who: patients at risk for sudden cardiac arrest (SCA)

Dr Gupta illustrates that sudden cardiac arrest is a major national health problem. Accounting for approximately 15 to 20% of all deaths, and, within the United States alone, 500,000 patients will die every year. He notes that patients at risk for SCA primarily can be treated with a preventable Implantable Cardiac Defibrillator (ICD), but only 10% of patients who meet the clinical guidelines receive this potentially life-saving treatment. Lastly, Dr Gupta looks towards the future of how to solve this noting, “We need automated continuous AI monitoring of sudden cardiac arrest … and we need to close gaps in care between patients who are known to be at risk for sudden cardiac arrest and those who receive ICD therapy. These are things that we believe AI can help answer.”

AI in cardiac care 

Logan Brigman highlights the Cardiac Intelligence® software, discussing that it’s an artificial intelligence program which automates identification of patients with disease and lack of follow-up. The program utilizes AI and Natural Language Processing (NLP) to convert unstructured data (such as echocardiogram reports and clinic notes) to structured data and applies algorithms to detect various cardiovascular diseases such as patients at risk for SCA, severe AS, MR, LAA and more. The intention is to eliminate the undertreatment of cardiovascular disease by automatically identifying patients with disease and communicating back to the EHR to facilitate a referral to the appropriate care team.

“The ability to identify patients at risk for SCA goes far beyond risk factors and that there are many patients who require further diagnostics or medication adjustments before a therapy can become an option” says Logan Brigman. The Cardiac Intelligence® program provides surveillance of care pathway adherence and automates this complex process across disease states. Mpirik’s Cardiac Intelligence® is piloting the identification of patients at risk for SCA in collaboration with Medtronic. The collaboration also involves Vizient, a health care performance improvement organization that provides data and analytic insights.

Health equity 

During the webinar, Dr Booker explains that patients with social determinants of health tend to fall through the cracks. At UAB they see a lot of index echoes that are ordered by non-cardiovascular providers and sometimes there is a lack of appreciation for which conditions should be referred to higher levels of care.

A question was asked to Dr Booker, “Where are you seeing most patients fall through the cracks and have you determined which of these patients, not receiving follow up treatment plans, fall into categories of racial or ethnic disparities?”. Dr Booker explained that there are two categories of trend for patients falling through the crack: 1) the underinsured and uninsured, and 2) patients with an index echo that was ordered by a non-cardiovascular provider. He notes those are two “very disparate groups of people” and that AI has helped to close the gap.

Outcomes with usage of AI 

Dr Booker discusses his experience with utilizing AI at UAB. Following the implementation of Cardiac Intelligence, UAB saw several benefits including, but not limited to:

  • “106 interventions occurred following an alert”
  • Overall time from index echo identified with disease to follow up event has been reduced
  • Several surveyed alert receivers noted the message directly impacted their decision to refer
  • More sensitive detection of disease
  • Quality improvements and ability to track discrepant reports in real time

Click here to watch