As the healthcare industry continues its shift into value-based care, many are looking for new ways to replace traditional care models and improve care quality. While there’s no consensus on the right strategy to make better use of the stores of data, AI may hold the key to creating a more predictive care model.

In fact, Accountable Care Organizations can only find success in the Medicare Shared Savings Program and other value-based programs with access to real-time clinical, outcomes, quality and financial data, explained Pam Hepp, shareholder of law firm Buchanan, Ingersoll and Rooney.

That access will help improve outcomes, control costs and “in effect change practice patterns,” said Hepp. “Up-to-date, comprehensive information about the providers’ own patients must be analyzed to identify care gaps and coordinate and manage care.” 

But to do that, providers need to be able to track and analyze large volumes of data to predict, for example, the patients who may be at high-risk for disease.

“AI uses data analytics but takes it much further,” said Hepp. Analytics work to summarize large stores of data in a meaningful way to help providers in their decision making.

With both retrospective and current data, providers can get feedback on a specific patient’s disease progression to develop a treatment plan, explained Hepp.

But “AI not only analyzes the data but adds predictive analytics to predict outcomes and make decisions such as by creating treatment protocols or plans that the physician may then implement,” she said.

The trouble is that many providers complain of data overload and a lack of time to focus on patients, explained Hepp. And often, those same providers are unwilling to adopt the technologies into their practices, as it would add to those issues.

So “more data - even if summarized and analyzed - in and of itself is not the solution, as such data can still be overwhelming and not suitable for practical application,” Hepp said.

“AI can be beneficial because it incorporates predictive outcomes coupled with decision-making tools, but providers need to be willing to adopt and use such technology, and that is often a difficult ‘sell’ for many practitioners,” she added.

Costs also are hindering adoption, which also includes EHRs that still need to be deployed across the entire ACO network, Hepp explained. Plus, there are regulatory constraints as to how the technology can be shared between the health system with the independent providers in the network.

This leads to concerns around HIPAA and data security: There’s a need for processes to be built into the AI software design or policies that address how and what types of data can be shared, such as mental health or HIV-related information, she said.

The industry also has struggled with how to include data from unregulated patient fitness devices, “all of which slows and adds cost to the development,” said Hepp.

She added that the broad push toward value-based care will create more alignment in the industry and, ultimately, make it more difficult for smaller providers to survive.  

“That is particularly true when considering the technology investments required for participating in such reimbursement programs,” Hepp said. “The inclusion of providers across the spectrum is vital to successfully coordinating and managing patient care.”

Twitter: @JF_Davis_
Email the writer: jessica.davis@himssmedia.com