Insurers Are Using AI to Boost Risk Management

^{Apixio leverages artificial intelligence to help healthcare organizations minimize risk.}

As value-based care begins to spread through the healthcare system, artificial intelligence (AI) is becoming an important tool to manage risk and boost provider efficiency.

Apixio is at the forefront of these efforts. The California-based data science company has signed up around three dozen health plans and providers for its HCC Identifier tool, an AI-powered risk adjustment platform.

>> READ: People to Insurers: Don’t Use Big Data to Measure Risk

Right now, the company is focused on clients who offer Medicare Advantage products, an area that has garnered close scrutiny from regulators.

“The growth of value-based care is driving the need for deeper healthcare insights, for key activities such as risk adjustment, quality reporting, care and utilization management,” Apixio CEO Darren Schulte, M.D., told Healthcare Analytics News™.

The company’s pitch looks something like this: The Centers for Medicare & Medicaid Services is increasing audits of health plans offering Medicaid Advantage (MA), to ensure correct payments. That means health plans have to keep careful tabs on coding and payments. But unless the health plan owns the clinic, providers don’t share the same financial risks as insurers.

Apixio says about 85-95 percent of MA risk-adjustment payments are based on diagnosis codes from clinics and hospitals. The remaining payments are from chart coding. Diagnosis codes are backed up by patient notes and physician codes, but those records aren’t often reviewed prior to payment by the health plans. Thus, any errors may go undiscovered, which could cause regulatory headaches for health plans — but usually not clinics.

Schulte said AI helps address the problem because it can compile and analyze large amounts of data from disparate sources.

“The problem with healthcare data is that it is difficult to accumulate and access,” he said. “Healthcare data is spread across multiple channels: EMRs (electronic medical records), PDFs, text files and faxes (yes, these still exist).”

Another complicating factor, Schulte said, is the wide variety of electronic health record software in use, which aren’t always capable of easily transferring records from one provider to another.

“Apixio’s AI platform can access data from a variety of sources, assemble longitudinal care profiles and apply learning-based models to mine documents and predict future events,” he said.

The system also makes it easier to retrieve data when an organization is audited.

Gateway Health, a Pittsburgh-based health plan, was among the first to sign up for the service. In a press release, Gateway Health’s Gabriel L. Medley, M.H.A., MBA, said pressure from regulators drove the health plan to look into AI.

“With the frequency of audits increasing, and other government agencies like the Department of Justice and Office of the Inspector General weighing in on risk adjustment payments, the pressure to ensure every claim is supported has never been higher,” said Medley, vice president of quality and risk adjustment programs at Gateway Health.

Though Apixio began with Medicare Advantage, Schulte said the range of potential AI applications is much broader.

“We started with risk adjustment, given that it is foundational to any value-based care contract and delivery system,” Schulte said. “If an organization cannot effectively predict near-term future costs, risk-based payment models and patient management activities will be inadequate.”

Over the next 12 to 18 months, Apixio plans to focus on improving risk management for providers and Medicaid members. The company is also working to develop an enterprise data solution that can improve quality measurement and care and utilization management, all key components of value-based care.

Get the best insights in healthcare analytics directly to your inbox.

Related

EHR Data Reveal Striking Racial, Ethnic Disparities in Hypertension

What Will Humana’s New Analytics Center Tackle?

How Edge Computing Can Advance Healthcare