Key Takeaways
- Research shows automatic alerts can enhance clinical care for aortic stenosis and mitral valve regurgitation.
- Artificial intelligence and automated processes can help guide clinical decision-making to provide patients the care they need.
- Duke Heart & Vascular continues to lead research and guide clinical care in the AI space.
Although severe, symptomatic aortic stenosis (AS) is typically fatal in two years without treatment, research has shown that patients with AS are undertreated. The ALERT randomized controlled trial tested whether automatic notifications could help more patients receive treatment for severe AS or mitral valve regurgitation (MR). Artificial intelligence (AI) analyzed echocardiogram reports to generate the alerts on the patient’s health record suggesting physicians consider further care.
Across five health systems and over 30 hospitals, the electronic notifications increased the rate of valve interventions from 9.6% to 13.4% and multidisciplinary heart team evaluations from 17.9% to 22.7% within 90 days versus usual care.
“When you give providers an extra alert stating that the patient has severe AS or MR and may benefit from further treatment, this impacts care,” says Duke Health cardiologist Sreekanth Vemulapalli, MD, who served on the study steering committee and as senior investigator. “The work is ongoing, but we presume mortality will also improve.”
Clinical decision support
Generally, echocardiography reports indicate diagnoses but not whether providers should consider treatment. The alerts worked so well that Duke has already integrated the approach, according to Vemulapalli: “Now, if an echocardiogram interpreter says the patient has severe AS, the report automatically includes a statement from the American Society of Echocardiography guidelines to recommend further treatment.”
The concept of automatic notifications can be adapted to many parts of the health system, but electronic notifications are just one form of clinical decision support. “There are ways we can aid in delivering guideline-recommended, appropriate care,” Vemulapalli explains.
Although AI can be used to generate these notifications, human decision-making guides the entire process. “It’s still a human reading the echocardiogram and providing care,” Vemulapalli says.
Multidisciplinary heart care
“We selected multidisciplinary evaluation as a study outcome because physicians across specialties collaborate with the patient to decide the best course of therapy and provide holistic care,” says Vemulapalli. Often, a patient will require referral to access comprehensive care.
Vemulapalli highlights numerous benefits to Duke’s multidisciplinary heart team, including diagnostic quality, efficient transition from diagnosis to evaluation, advanced therapeutic options, world-class expertise, high quality, and access to cutting-edge treatments and clinical trials. “We want patients to go where they can be diagnosed accurately and well, get to the appropriate experts to be evaluated for treatment, and where all the treatments possible are available, with experience in all of them,” says Vemulapalli. “Duke offers all of these advantages.”