Jacob Kupietzky is chairman of Healthcare transformationa company dedicated to providing experienced interim executives to hospitals.
Not so long ago, if you wanted to explore the intersection of healthcare and artificial intelligence (AI), you would confine yourself to the pages of science fiction. Not anymore: In recent years, AI has evolved from what is possible to what is practical, and consumers and practitioners alike are increasingly drawn to the possibilities of how AI can revolutionize healthcare.
While the promise of AI in this area is just beginning to be realized, we are already seeing it having a real impact on patients’ lives. Here are three ways AI is disrupting healthcare practice today.
Automation of labour-intensive tasks
Delivering the quality care patients are looking for takes time – something that is already in short supply in most healthcare facilities. The more clinicians can shift labor-intensive or administrative tasks to AI solutions, the more time they can spend with patients. AI can perform those labour-intensive tasks with less (or no) human assistance and even reveal inefficiencies in current practices.
How AI can help
While some tasks may seem tailor-made for AI-based solutions, such as chatbots and schedule reminders, they are just the tip of the iceberg. For example, automation software company Olive AI recently launched its Autonomous Revenue Cycle solutionwhich can automatically check insurance eligibility, identify benefits and ensure claims are processed correctly. The Qventus perioperative solution uses AI and machine learning to address inefficiencies in manual operating room scheduling, helping health systems across the country add more OR cases per month, schedule procedures further into the future, and increase revenue by more than $10 million per facility can increase.
AI can also be used to analyze and improve current practices. A 2022 UPenn study used AI to analyze the inpatient and outpatient notes of nearly 2 million patients and found that about half were duplicates of previous notes—a practice that “questions the accuracy of any information in the medical record.”
Virtual patient care
It’s easy to see how AI could revolutionize healthcare, even through virtual care alone. It is time-consuming and prohibitively expensive to provide 24-hour care to each patient. However, emerging virtual care solutions use AI to alert healthcare providers when care is needed and to monitor and even influence patient behavior to create better health outcomes.
How AI can help
VirtuSense wearable technology solutions use AI to prevent injury and even improve patient quality of life: VSTAlert reduces falls in skilled nursing by 75% and hospitalization cases by 78%, and VSTBalance can detect gait and function disorders, enabling the mobility of residents in residential care centers by up to 85%. Biofourmis’ virtual care management service uses analytics, AI and wearables to remotely monitor patients’ complex chronic conditions. And facility automation platform care.ai uses AI to monitor patients in the facility for falls, pressure ulcers, elopement, and compliance with protocols.
Diagnose medical problems
The promise of AI in healthcare isn’t just because it’s faster; in many cases it is better than its human counterpart. This is perhaps the most exciting aspect of AI development: the potential that it can accomplish tasks and achieve results that humans alone cannot.
How AI can help
Although it is still in its infancy in healthcare, AI is already making incredible strides in medical diagnosis. Viz.ai’s artificial intelligence software can compare CT images of a patient’s brain to its database of scans to find early signs of major blood vessel occlusion strokes — then alert doctors, who can see the images on their phones . The Mayo Clinic’s AI-assisted screening tool “Identified people who were at risk for left ventricular dysfunction 93% of the time. To put that in perspective, a mammogram is accurate 85% of the time.”
Digital diagnostics uses AI to detect diabetic retinopathy and even some types of skin cancer. And a recent study reported in Nature found that a trained neural network (a complex form of machine learning) can classify hip fractures 19% more accurately than even experienced human observers in a clinical setting. Because classification is a major determinant of treatment, greater classification accuracy can lead to better treatment, better patient outcomes and lower costs.
The need for practitioners to keep up
However, for all its promises, AI still faces significant obstacles to widespread development, including cost and evidence of effectiveness. But perhaps the biggest obstacle lies with the healthcare managers themselves: According to a 2022 study by IT services and consulting firm Capgemini: “The acculturation level of C-level executives is lagging, especially for organizations that need it most: pharma, medtechs and hospitals.” As healthcare executives, we know that AI is not the future of healthcare, but the present, and we need to lead the way.