Medical claims 'mined' to find fraud

As seen in USA Today. Although they need to look as the psych diagnosis a little more carefully.

Medicare investigators in Los Angeles, using sophisticated computer technology to sift through claims data, saw an unusual pattern: A single patient had apparently undergone a diagnostic rectal-probe procedure 118 times in a year — at 21 medical facilities.

"It's unlikely that could have occurred," says Kim Brandt, director of program integrity at the Centers for Medicare and Medicaid Services. "This person would not have been able to sit on a plane."

Borrowing techniques from financial and credit services, Medicare and private health insurers are increasingly "mining" claims data for suspicious patterns, comparing practitioners with their peers and larger databases of claims.

Medicare investigators found similar cases across a group of diagnostic testing companies, which used improperly obtained Medicare patient identification numbers to bill for the exams, which were likely never performed. Problems were found with other types of tests as well. As a result, 83 diagnostic centers this year lost their billing privileges, and $163 million in payments were denied.

Without the computer program, the fraud might have gone undetected.

"I call it spider-webbing: Find one common denominator (and follow the thread)," says Brandt.

Fraud is estimated to account for 3% to 10% of the nation's $2 trillion health care spending tab. Seeing a market opportunity, such companies as Fair Isaac, IBM, ViPS and Ingenix, a subsidiary of UnitedHealth Group, have launched fraud-detection software and systems in the past few years.

Using such a system, Independence Blue Cross in Pennsylvania flagged claims from a psychiatrist who appeared to be seeing up to 63 patients a day. An investigation found that a former patient had stolen the doctor's billing identity, set up a counseling service and billed Blue Cross for more than $1 million in services.

Aetna, which processes 170 million claims a year, says its fraud-detection software helped save it more than $89 million last year.

"The most widespread and damaging schemes are committed by a very small percentage of dishonest health care providers," says Mike Stergio at Aetna.

Data mining has been likened to a "witch hunt" by some doctors because it looks at factors that can be subjective: Is the doctor overtreating? Seeing patients too many times?

"We have zero tolerance for intentional strategies to defraud insurance companies or public programs," says Edward Hill, the AMA's immediate past president. He says systems need to get better at distinguishing fraud from inadvertent billing errors or legitimate medical treatments.