Strengthen the business case for automated fraud detection

The amount of fraud and abuse in the U.S. healthcare system is staggering and continues to grow at an exorbitant rate. Many factors contribute to this growing incidence of insurance fraud. For example, an ever-increasing volume of transactions, antiquated procedures that were not established to assume that fraudulent claims would be made, and complex and multiple reimbursement methodologies that open insurers to the risk of inappropriate claim payments because of fraudulent and abusive billing practices.

However, healthcare payer organizations have been slow to adopt technology as a fraud-detection tool, even though manual fraud investigation techniques are labor-intensive. The number of heterogeneous IT environments also has delayed deployment. To help reduce healthcare costs and minimize losses, healthcare payer organizations should explore the capabilities of new and emerging fraud-and abuse-detection tools. The need to do so is made more urgent by the inability of insurers' manual systems to stay on top of the increasing flood of fraud and abuse.

Historically, the only available automated approach to detecting fraud was to detect fraudulent claims after they had been paid-a retrospective approach. The retrospective approach uses rule-based data-mining capabilities, searching for claims that meet well-defined conditions, such as doctors who bill for 25 hours of office visits in a single day. However, if fraudulent providers vary their schemes from case to case, this retrospective method is not effective. The post-payment data-mining technology can react to rules drawn up to detect known fraud schemes, such as bills for tests unrelated to a procedure or a provider that bills for the same patient twice. The rule engine software picks out only the most worthwhile cases to pursue. Although this widely adopted approach is necessary and valuable because it can identify and store provider patterns and discrepancies, the ability to recoup funds after claims have been paid is more difficult.

The latest approach to fraud detection is the prospective approach. This approach has the ability to stop-not just detect-potentially fraudulent or erroneous claim submissions before the claims are paid. With the prospective approach, pattern-matching technology is used. The software can sift through claims and segment them by groups. The pattern-matching system is good at looking for "aberrations" in claims by comparing a provider with a peer group. For example, if a pediatrician starts billing for allergy treatments, this system may single out the claim as departing from the norm and thus suspect. Payers should note that the prospective approach is new to the market and therefore may not fully satisfy requirements. Anecdotal information from payer references indicates that too many claims are being flagged as suspect. Because of the lack of internal resources to review these claims and to analyze the pattern-matching logic, consultants are being brought in to assist. When deploying the prospective approach, payers must take the time to review and thoroughly analyze the rules used for pattern matching to ensure that only erroneous claims are flagged as suspect.

Payers must realize that pattern-recognition systems (the prospective approach) or rule-based systems (the retrospective approach) can only see shades of gray in a mass of health claims. Some payers are augmenting their rule-based systems with pattern-recognition technologies to get the best of both solutions. However, pattern-recognition technologies and the prospective approach are still immature. Payers should determine whether they have the internal resources to support the deployment of this approach.

Visit http://gartner.com/healthcare/ or e-mail info@gartner.com
.

Joanne Galimi is research director with Gartner's Healthcare Industry Research & Advisory Services.