HealthcareITNews
Experts on AI in healthcare: 'We need to be more realistic'
By: Tom Sullivan
CHICAGO — Artificial intelligence in healthcare has become more buzzword than practicable application thanks, in part, to bold claims that AI will be able to transform healthcare by combining billions of data points and diagnosing diseases, notably cancers, faster than any doctor ever could.
“We do ourselves a disservice when we refer to AI as something that we just take a ton of data, push a button and cure cancer,” Leonard D’Avolio, CEO and co-founder of machine learning startup Cyft said at the HIMSS and Healthcare IT News Pop Health Forum in Chicago Monday. “We need to be more realistic about what these tools can and can’t do.”
John Supra, vice president of solutions and services at the Care Coordination Institute, agreed that just having machines tell doctors exactly what to do is not necessarily what healthcare really needs.
That said, both see big opportunities in the artificial intelligence, cognitive computing and machine learning realms for hospitals.
D’Avolio pointed to Amazon, Facebook, Google, Pandora, Spotify and Target as companies making use of machine learning to improve the customer experience.
“Amazon does not consider everyone to be the same consumer, it understand you based on the data. That’s the foundation for transformation,” he added. “The opportunity now is to blend this customization and understanding into healthcare.”
Supra said that machine learning will help to drive a shift from people spending a lot of time on reporting in siloed EHRs and, instead, transforming those resources toward analysis and predictive modeling.
“Value-based reimbursement needs to bring [disparate data sources] together,” Supra said. “Those interactions are complex.”
D’Avolio added that since 95 percent of care today is fee-for-service, “that glass is 5 percent full,” meaning that particular 5 percent is actually incentivized to keep people healthy.
He urged hospitals to work toward expanding that to 10 percent or 15 percent of care they deliver being under a value-based model — and to know that AI is one of the tools for making that happen but it really is just a tool.
“We’re moving from one-size-fits-all to more of a precision care management approach,” D’Avolio said. “That is the opportunity, not to cure cancer in two years necessarily but to do a better job of things we do every day.”
