The digital transformation of healthcare is underway, but it will advance further and faster if key stakeholders work together. That is the idea behind the Applied Health Innovation Consortium (AHIC) led by the American College of Cardiology (ACC). Together with researchers, clinicians, patient advocates and technology developers, the ACC seeks to develop a roadmap for Artificial Intelligence (AI) that will enable clinicians to predict and address an individual patient’s needs with greater precision, ideally before the patient gets sick.
For example, one complex problem that cardiologists face regularly is the cardiac arrhythmia known as atrial fibrillation (AF). According to the Centers for Disease Control, an estimated 12.1 million people in the United States will have AF in 2030.1 The American Heart Association says patients with AF are five times more likely to have a stroke, three times more likely to suffer heart failure and are hospitalized twice as often as patients without AF. 2 In addition to its clinical implications, AF costs the United States health care system an estimated $26 billion each year.2
While strides have been made to combat AF, the magnitude and complexity of the problem demands a new approach to maximize value from the mountains of clinical data we now possess to diagnose and treat patients with speed, accuracy, and efficiency. Could AI be the answer? Dr. Jonathan Lindner, the M. Lowell Edwards Professor of Cardiology at Oregon Health & Science University, believes so.
“Having an AI algorithm that is extremely well vetted and that has data behind it in terms of being able to not only help clinicians predict these diseases, but also predict who is likely to respond to certain therapies, would allow us to essentially go into a patient’s room with much more confidence and tell them that they are highly likely to either benefit from a certain therapy or not,” said Dr. Lindner.
While the prospect of being able to better predict and possibly prevent disease before it takes hold is both exciting and transformational, there are barriers to integrating AI with clinical medicine.
Successfully Applying AI to Digital Healthcare
According to Dr. John Rumsfeld, the Chief Innovation Officer and Chief Science and Quality Officer at the ACC, “The promise of AI for healthcare is that we could be more efficient, improve clinician well-being and patient engagement, and improve health and health outcomes if it’s done correctly.”
The ACC has long anticipated the digital transformation in healthcare – or shifting the focus of care to virtual care and remote monitoring – and is committed to helping facilitate and execute an ongoing innovation strategy. Dr. Rumsfeld says the keys to AI being successfully applied to digital healthcare are high-quality data, clinical use cases, effective integration into workflow and training future generations that this is the way we deliver care—and that we can move further faster to deliver trust in the form of evidence and standards if we collaborate cross-functionally.
Dr. Rumsfeld lauds strong industry participation, such as GE Healthcare’s involvement with the AHIC.
“If we’re going to have tools like AI make a fundamental difference in the way we deliver care, no organization can do that alone. We need collaborators who share the same vision and bring different strengths to the table. GE has been right at the forefront of helping the ACC reimagine healthcare delivery,” he said.
Currently GE Healthcare’s Edison platform forms the technological basis for the company’s many AI offerings that are poised to become an integral part of the advanced cardiac technology used by clinicians in the diagnosis and treatment of more than 145 million hearts each year.
Dr. Rumsfeld envisions a future in which clinicians are using AI predictive algorithms to highlight people who may be at such a high risk for developing complications for which preemptive action is recommended.
“Today, we’ll look at a patient and say they have moderate aortic stenosis, but one such patient may stay in that state forever and never need a valve intervention while another may have a digital phenotype that suggests they’re going to have rapid progression and need earlier intervention. The individual human eye isn’t going to be able to interpret that based on images alone. This is where AI can make us better at what we do,” Dr. Rumsfeld said. “AI won’t just interpret the image, but it actually can see patterns in the imaging that will make us better at diagnosis and better at predicting future course of care and outcomes for patients.”
- https://www.cdc.gov/heartdisease/atrial_fibrillation.htm Last accessed 06/2/2021
- https://www.heart.org/-/media/files/professional/quality-improvement/get-with-the-guidelines/get-with-the-guidelines-afib/afib-infographic-pdf-ucm_473263.pdf?la=en Last accessed 05/30/2021