More on Operations

Artificial intelligence helps insurers identify Medicare members who also qualify for Medicaid

Medicare Advantage insurers that miss enrolling dual-eligibles are losing out on an opportunity, according to Change Healthcare.

Susan Morse, Managing Editor

Medicare Advantage insurers that miss enrolling dual-eligibles are missing an opportunity, said Change Healthcare, which has introduced a combination of artificial intelligence and behavioral science to identify candidates.


Enrolling individuals dually-eligible for Medicare and Medicaid is good business for insurers who are already in the Medicare Advantage space.

HIMSS20 Digital

Learn on-demand, earn credit, find products and solutions. Get Started >>

While this population has more chronic conditions, payers are reimbursed by state Medicaid benefits for the risk they take on.

For the provider, it adds to the panel of people who are now covered in the system.

This population is rising as baby boomers age. Unfortunately, some seniors may find they qualify for Medicaid if their retirement income is low enough. The guideline is 138 percent of the federal poverty level, an estimated $16,394 for an individual.


Of the 58.5 million Medicare beneficiaries, 34 percent live at or below the federal poverty level, indicating they might be eligible for Medicaid. But only about 10.6 million of 19 million Medicare Advantage members are dual-enrolled.

Health insurers have traditionally identified dual-eligible candidates using systems that require manual programming and continual updating. The prior method was based on a static algorithm that was limited by small data sets, according to Change Vice President of Engagement Keith Roberts.

Artificial intelligence has taken over part of the burden by rapidly analysing mass amounts of data.

Change's dual enrollment advocate system is able to isolate 15 attributes of those dually eligible, according to Roberts.

A test set of 875,000 consumer records identified 106,000 who were likely to be fully eligible, about 12.1 percent, Roberts said. It went further to identify 112,000 who were partially eligible, about 12.9 percent.

For payers in particular, the financial area of healthcare is where AI expertise lies, Roberts said. Change is making a big bet on it.

"We are investing 14 percent of revenues back into R & D," Roberts said. "It will have a big impact on reducing the cost of healthcare."

AI is expected to generate $150 billion in savings by 2026, with $18 billion of that in healthcare, according to Accenture analysis.


Change Healthcare's solution is a blend of AI and behavioral science.

The behavioral side involves member engagement teams first gaining the trust of potential beneficiaries.

"This is a population that feels preyed upon for scams," Roberts said.

The second challenge is giving consumers information relevant and personalized for them. And the third is giving them control.

The engagement teams use targeted mailings and other outreach methods to activate dual-eligible members and support them through the enrollment process.

Feedback from payers has been good, Roberts said, calling it the best quarterly reviews they've had. The method boasts a 93 percent accuracy in identifying patients who qualify for dual eligibility.


"Using AI plus behavioral science and experience design represents the modernization for a vulnerable group of consumers these days," Roberts said. "AI and machine learning can't do it alone. Behavioral science can't do it alone. And health plans and legacy systems can't do it alone. The time has come to bring these healthcare IT and scientific disciplines together to help solve a critical business challenge."

Twitter: @SusanJMorse
Email the writer: