Regarding the use of artificial intelligence (AI) to provide efficiency gains in various healthcare processes, there have been a number of articles suggesting how to launch this new technology. I have discussed AI analytics in multiple blog notes (see, for example: Healthcare AI Market to Surge; Already Commonplace in Radiology; Cerner Offers AI Tool to Combat Physician EHR Burnout; Which Healthcare Jobs Are Safe from Replacement by Automation or Robots?). Hospitals will certainly be pursuing this goal on their own in order to reduce risk and costs. A recent article indicated that CMS was accelerating this process by launching a contest with a $1M reward for AI developers who come up with a way to predict unplanned hospitalizations (see: The Centers for Medicare and Medicaid Services is launching a challenge for AI developers to predict unplanned hospitalizations). Below is an excerpt from the article:
The Centers for Medicare and Medicaid Services (CMS) launched a new challenge that pits artificial intelligence (AI) developers against each other to create algorithms that can best predict unplanned hospitalizations. To attract the cream of the AI-developing crop, the agency plans to dole out a $1 million reward to the winner. The CMS' AI Health Outcomes Challenge comes amid speculation that a lack of healthcare AI regulation could spawn misuse, but recent federal government involvement in AI could help put some concerns at bay. Here's what it means: The US government is making moves to push AI deeper into healthcare....Dubbed The American AI Initiative, the order calls for the HHS, among other agencies, to prioritize funding the tech. But they'll have to use existing funds, as no new funds were allocated to the cause. With a newly restricted budget, it's hard to say if an uptick in healthcare AI investment is in the cards for the HHS. The US Food and Drug Administration (FDA) has shifted its regulatory coverage to encourage the development of AI healthcare products. In an effort to catch up with rapid digital innovation in the healthcare industry, the FDA announced last April that it was working to facilitate the inclusion of AI in digital health tools by looking into how it can apply its pre-certification program for tools based on AI.
I think that this AI contest is an appropriate step for CMS to take. After all, a million dollar reward will look trivial compared to the savings that CMS may gain on the basis of successful predictive algorithms. I do have several concerns about this idea, however. The first is that, obviously, any algorithm used to predict unplanned hospitalizations needs to be extensively tested to endure its validity and safety. Secondly, the current hospital EHR environment is very heterogeneous, sometimes even for hospitals running the same EHR software. Such an environment poses a number of challenges not the least of which is that EHR vendor support for AI predictive analytics tools will slow their emergence. Algorithms that reduce costs and increase patient safety will also be extremely valuable. The intellectual property of the developers of predictive analytics needs to be protected in these early hospital testing stages.
Comments