By Monica E. Oss, Chief Executive Officer, OPEN MINDS
Using artificial intelligence, referred to as “AI” (and sometimes called “augmented intelligence”) always seems like a “tomorrow” issue. Interesting, but not of much use right now in the day-to-day delivery of services.
It’s not that investors aren’t enthusiastic about AI health care technologies. Just last week Woebot closed their Series B funding with $90 million. Wysa, an AI digital platform for mental health, announced that it has secured $20 million in financing, which it will use to expand into additional global markets. Cleerly, a platform that uses AI to identify the features of plaques that may cause heart attacks, announced a $192 million Series C round of funding, bringing its total funding to $248 million. And Merck has begun taking applications for a new biotech accelerator program to support startups offering AI and machine learning.
The reason for investor enthusiasm is the broad range of current and future uses for AI in health care. That was the focus of The 2022 OPEN MINDS Strategy & Innovation Institute session, Using Advanced Analytics, Workflows & AI To Decrease Provider Burden, Improve Quality Care & Maximize Revenue, by Michael Lardieri, senior vice president for strategy at Core Solutions. He provided an overview of health care uses of AI—from more precise analytics to smartphone-assisted consumer and condition monitoring to precision diagnostics and treatment.
The question that busy executives have is what can AI do for my organization right now? Mr. Lardieri was quick to point out that AI is being used now by a number of health and human service organizations, giving them a leg up in improved performance and reduced cost. In particular, he called out the current use of AI in improved schedule management, in better consumer engagement, in predictive analytics for consumer care coordination using electronic prescribing data, and in risk stratification for population health management.
Improved schedule management—With ongoing staffing challenges, AI can assist with two scheduling challenges. The first is enabling provider organizations to better predict consumer no shows and cancellations. Current software can automatically overbook specific consumers based on the probability that they will show up at the scheduled time. The second is to use the software to better manage clinical staff’s work schedule to reduce burnout and improve staff engagement.
Better consumer engagement—AI-infused platforms currently offer the ability to track consumers use of an organization’s services and their health status. If an organization can automatically track which consumers it hasn’t seen in a while, by both clinical profession and service line, this can help rank how engaged your entire population is with those staff and service lines. If management sees that a consumer isn’t engaged, and the health outcomes are suffering, they can make adjustments to outreach and service delivery.
“You can look at consumers coming in by treatment type,” said Mr. Lardieri. “And you can look at your data by clinicianand what the drop off rate is. I don’t like to look at ‘just engagement’ because I want to also look at outcomes. So, based on the level of engagement are they still getting better? I think so many times we just do sessions every week or every two weeks. But can this be done every month? Modifying treatment intensity depending on the needs of the specific patient opens up the use of so many other services.”
Predictive analytics for care coordination—AI provides a broader and deeper range of predictive analytics for care coordination. By leveraging this ability, provider organizations can “know what’s coming,” and make better estimates on how many resources or staff they will need and make adjustments that can diminish unnecessary cost or invest more heavily in certain areas to provide services when they are needed.
Risk stratification for population health management—In population health management, risk stratification, which is the ongoing process of assigning all consumers a risk status based on health measures, is key. AI assists with early identification of consumers who are at greater risk of moving to a condition where they need expensive acute care and interventions. This provides the opportunity for early and specific interventions that improve consumer health status and decrease the cost of their care.
To make any of this work, Mr. Lardieri pointed out that provider organization executive teams can’t simply “adopt AI.” They need to understand the processes and functions that allow their organizations to create value for consumers and payers—and then understand how AI can make those processes and functions better, faster, and less expensive. “The improved use of AI is going to be required for organizations to be efficient and effective, and to maximize revenues and improved outcomes,” he said. “This is what we need to look toward and for.”