North Carolina Fights Medicaid Fraud With Analytics

In two years, IBM data mining software has identified $191 million in potentially false Medicaid claims.

Marianne Kolbasuk McGee, Senior Writer, InformationWeek

May 23, 2012

4 Min Read

The Great ICD-10 Debate: Healthcare Coding Transforms

The Great ICD-10 Debate: Healthcare Coding Transforms

The Great ICD-10 Debate: Healthcare Coding Transforms (click image for larger view and for slideshow)

States budgets are under pressure, and Medicaid funding is often among the hardest hit. But in some states, including North Carolina, data analytics software is helping to uncover millions of dollars in potential savings through the detection of fraudulent Medicaid billing.

North Carolina's department of health and human services began collaborating with IBM in 2010 to develop analytics to help identify suspicious billing patterns by healthcare providers.

To date, North Carolina has identified $191 million in potentially false Medicaid claims by 206 outpatient behavioral health providers in the state. North Carolina processes about 88 million claims totaling about $12 billion annually from 66,000 providers who treat the state's two million Medicaid patients.

Using the IBM software, the state has identified suspicious claims that include submissions from providers with patterns of billing that don't fit "the norm," said Al Delia, secretary of North Carolina's department of health and human services in a recent blog posting.

[ Too many organizations approach IT projects from the wrong angle. Learn why in Kaiser Permanente CEO: Health IT Must Focus On Quality. ]

Examples of suspicious activity uncovered include: providers billing for more than 24 hours of service in a single day; overbilling in the treatment of individual patients; and double billing for the same services--such as billing for group therapy sessions, and then billing for individual sessions with the same patients.

Using three years' worth of North Carolina Medicaid claims data, IBM data mining software, which features special algorithms and modeling capabilities, was applied to detect common fraud and abuse schemes, said Shaun Barry, IBM's fraud and abuse management leader heading the North Carolina project.

Also applied in the project was IBM InfoSphere Identity Insight, which is software "that helps the agency understand 'who is who,' and resolve any identity conflicts that arise," said Barry in an email interview with InformationWeek Healthcare.

"For example, it resolves many different distinct provider numbers into single entities based on shared attributes, characteristics, and numbers," he said. "This helps North Carolina determine if providers are using different names to provide services in an effort to expand their bill capabilities," he said.

To date, North Carolina has completed 10 investigations into cases of suspicious outpatient behavioral health billing, yielding a total of $6.2 million in potentially fraudulent payments. These cases have already been referred to the Medicaid investigations unit of the North Carolina attorney general's office for further investigation.

Meanwhile, investigations into many other suspicious cases are underway.

In addition to the North Carolina project, IBM is working with a variety of different agencies across a half-dozen states "to root out fraud" using analytics software, Barry said. However, "North Carolina is the only one using this specific combination of tools," he said.

In the future, the U.S. healthcare industry's transition from ICD-9 to ICD-10 coding for medical billing also should make it easier to detect fraud, said Barry.

ICD-10, short for the International Classification of Diseases, 10th edition, expands by thousands the number of medical diagnosis and in-patient procedure codes used for clinical, billing, and financial systems of healthcare providers, payers, and other covered entities, compared to ICD-9. The U.S. Department of Health and Human Services recently announced a proposed rule to postpone the compliance date for ICD-10 by one year to Oct. 1, 2014.

The transition to ICD-10 will provide health insurers with "better, more detailed information on how providers treat their patients," Barry said. "They will be able to spot suspicious behaviors faster and with greater precision" before claim are paid.

However, in the short term, Barry suspects that the implementation of the ICD-10 coding system "will give fraudsters a greater opportunity to cheat the system."

That's because for many types of medical procedures, ICD-10 adds considerable complexity through the increased number and granularity of codes. "Fraudsters will take advantage of this complexity--during the transition to ICD-10, even the most well-intentioned provider may find it difficult to comply in full with the design of the new codes."

Get the new, all-digital Healthcare CIO 25 issue of InformationWeek Healthcare. It's our second annual honor roll of the health IT leaders driving healthcare's transformation. (Free registration required.)

About the Author(s)

Marianne Kolbasuk McGee

Senior Writer, InformationWeek

Marianne Kolbasuk McGee is a former editor for InformationWeek.

Never Miss a Beat: Get a snapshot of the issues affecting the IT industry straight to your inbox.

You May Also Like

More Insights