Is all artificial intelligence (AI) created equal? A recent groundbreaking Partners HealthCare initiative with GE Healthcare points to no.
The initiative, three years in the making, is a platform that imports more contextual data from both imaging technologies and patients’ medical records, to inform better, smarter clinical decision-making. The platform and resulting software help to solve a longstanding problem of AI deployment—data either too narrow to apply across medical departments, or too broad to be clinically useful. With this development, specific programs can be built for different conditions, equipment, and intended uses—offering a significant opportunity for developers to identify and deploy new AI applications.
“It's kind of a landmark opportunity that we have here to be able to show that not only can we create these algorithms, but we're actually putting them in practice and actually see transformation in care,” says Keith Dreyer, DO, PhD, Chief Data Science Officer, Partners HealthCare.
An example of potential application is in stroke care. Identifying which patients would benefit from advanced clot intervention requires advanced MRI machines that community hospitals lack. If community hospitals used AI to read CT scans, however, they might better identify which patients to send for the specialized treatment. Such applications augment, rather than replace, clinicians’ expertise.
“They’re informing humans how to make better decisions,” concludes Dr. Dreyer.
Read more at The Boston Business Journal.