Healthcare is in the midst of an AI gold rush, with traders, suppliers and payers pouring extra money than ever earlier than into new know-how. However in accordance with a gaggle of high healthcare CFOs and trade leaders gathered at HFMA’s annual convention, a lot of that funding is destined to disappoint.
The issue isn’t essentially the AI instruments themselves — it’s extra that the majority healthcare organizations are bolting these refined fashions onto infrastructure that’s decades-old.
Seema Verma — former CMS administrator and common supervisor of Oracle Well being & Life Sciences — put it bluntly: “You’ll be able to’t have an AI technique with out having an information technique.”
She argued the infrastructure most suppliers are working with in the present day is just too antiquated to assist what AI really requires — particularly, entry to real-time, unified knowledge throughout scientific, monetary and operational methods.
As an illustration, a doctor counting on an AI advice wants that mannequin to know whether or not a affected person’s well being plan will cowl a drugs, whether or not it’s in inventory, and whether or not the affected person has any interactions — all within the second.
With out that connective tissue, Verma mentioned even the strongest AI instruments will underperform.
Mike Marks — CFO of HCA Healthcare, one of many largest well being methods within the nation — echoed Verma’s issues and added that the sheer scale of technical debt is actively blocking progress.
“The quantity of legacy methods that we’re all coping with is basically getting in the best way of transformation,” he remarked.
It’s extremely costly to switch know-how at that scale, and the demand for AI innovation already far outpaces any well being system’s means to pay for it, which makes getting the foundational work proper much more crucial, Marks famous.
His method for prioritization is: scientific methods first, then operations and administrative capabilities.
The logic is patient-centered. Medical methods contact care supply straight, so enhancements there have probably the most instant affect on well being outcomes, Marks defined. Operations and administrative capabilities matter too, however in Marks’ view, they need to be optimized in service of a system that’s already delivering high-quality care — quite than the opposite method round.
Medical AI that underperforms has direct penalties for sufferers, so it calls for probably the most rigorous funding and oversight, Marks added.
One other well being system CFO — Scott Hawig of BJC Healthcare — agreed that the healthcare trade nonetheless has a protracted method to go earlier than AI can ship on its promise.
The present actuality, Hawig argued, is finest captured by what’s taking place between payers and suppliers. He painted it like this: two units of bots, every deployed by opposing sides, endlessly preventing over claims with nobody profitable.
“Bot versus bot — supplier income cycle bot in opposition to the denial insurance coverage bot — is the basic downside,” Hawig declared.
The AI instruments exist, the funding capital is flowing, and the urgency is actual. However as these consultants made clear, writing enormous checks with out fixing the information basis first may simply be a dearer method of staying caught.
Photograph: imaginima, Getty Photos
