informa
/
2 MIN READ
Feature

Tech Tools: Insurers Tap Two Strategies To Fight Fraud

In contrast to rules-based approaches, which require people who know a pattern of fraud to construct precise rules, pattern-matching software looks at masses of data from many angles.
Software that tries to counter health-care fraud might never be able to eliminate it entirely. But insurers are getting some more powerful weapons in the fight.

The most-commonly used tool is a rules engine, such as IBM's Fraud and Abuse Management System. IBM constantly adds to the system, so the engine includes some 3,000 rules that look for suspect conditions in the claims data. If it finds a match, it can flag human investigators for follow-up. IBM is adding to its system a module that can look at claims before they're paid. To avoid delaying payment, the system wouldn't test each claim with all 3,000 rules. It would apply small sets of rules deemed applicable as needed, says Tom Schamber, health-care consultant for IBM's Business Consulting Services. Ingenix Inc. and VIPS Inc. also offer rules-based systems that mine claims for signs of fraud.

A second approach to fraud detection involves pattern-matching technology, which pores over masses of data and generates norms for groups of health-care providers. Cases that look out of sync get flagged. In contrast to rules-based approaches, which require people who know a pattern of fraud to construct precise rules, pattern-matching software looks at masses of data from many angles. So an oncologist who also claims to work as an allergist would quickly come to light.

Fair Isaac Corp.'s Payment Optimizer uses pattern matching and might have found the "rent-a-patient" scheme employed by several Southern California clinics. The clinics flew in patients from out of state and paid them $800 to undergo minor surgery, then billed their insurers for the operations and a variety of unrelated treatments. One red flag: A patient who had a sweat gland removed was also treated for sleep deprivation.

A rules-based system might need to at least start with patients who had traveled outside their normal network of providers to receive treatment. That can be an unwieldy group amid the mass of claims.

A pattern-matching system, however, could find clinics that combine sweat-gland surgeries with unrelated treatments and rank them according to how far from the norm they were operating. "We would highlight the example that hasn't happened before," says Joel Portice, VP for health care at Fair Isaac. "The system learns the new aberrant practice."

The goal of both technologies is to get ahead of creative crooks who come up with new ways to steal from the health-care system.

Return to main story, Filter Out The Frauds

Editor's Choice
Brian T. Horowitz, Contributing Reporter
Samuel Greengard, Contributing Reporter
Nathan Eddy, Freelance Writer
Brandon Taylor, Digital Editorial Program Manager
Jessica Davis, Senior Editor
Cynthia Harvey, Freelance Journalist, InformationWeek
Sara Peters, Editor-in-Chief, InformationWeek / Network Computing