How prepared are insurance payers to detect fraud and abuse? A surprising number of claim departments still rely on manual detection processes, which simply are not sophisticated enough to identify the many patterns and types of claimant fraud. Adding to that is the volume of cases on any adjuster's desk. With an industry average of 250 claims at any time, adjusters are often too overwhelmed to perform the detailed analysis needed to find complex patterns that indicate fraud on top of their growing caseload.

According to the Coalition Against Insurance Fraud, 20 percent of total fraud at most is detected, and much of this fraud is detected late in the life of the claim. This late discovery dramatically increases the total cost on payers. In addition, a high percentage of claims that are referred to Special Investigative Units (SIUs) often are not fraudulent leads. The time and cost to investigators to pursue suspicious claims that don't turn out to be fraudulent or abusive contribute to the high cost of workers' comp.

Compounding this, a Coalition Against Insurance Fraud study noted that nationwide, fraud bureau budgets and head-counts have declined significantly in the past three years, while fraud referrals to SIUs are growing. With a growing volume of fraud referrals per investigator, properly identifying the patterns typical of truly fraudulent or abusive cases (or exceptional but not fraudulent claims needing aggressive case management) has become a huge challenge.

Greater Predictive Power

That's why insurance payers are turning to highly sophisticated analytic software called predictive models. Adjusters can spend significantly less time reviewing and processing claims, while spotting truly suspicious claims faster with greater accuracy.

Often based on a highly advanced software technology known as neural networks, predictive models can enable insurers to identify fraudulent, abusive and high-risk claims much earlier and with a higher degree of accuracy than any other known method. It can do this while swiftly and accurately processing the majority of claims without adjuster intervention.

Predictive models are largely responsible for the dramatic turnabout in fraud detection in the credit card industry. Today, predictive models are used to screen 85 percent of U.S. credit card transactions for fraud, resulting in a 50 percent reduction in industry losses. This same technology, used so successfully by 65 percent of the world's credit card issuers, is now increasingly being deployed to contain the spiraling costs of workers' comp.

In seconds, predictive models can scan thousands of data elements simultaneously to find subtle, complex, and hidden pat-terns of suspicious behavior. Their analytical and processing strength enables high-volume claim departments to perform a rigorous, objective review of every claim. While human experts are capable of identifying some red flags and simple fraud patterns, sophisticated modeling techniques are required to find more complex patterns of fraud. Based on historical examples already deter-mined to be fraud and abuse claims, neural network predictive models can learn which subtle patterns are associated with a high likelihood of fraud and which are not.

As claims evolve, they leave a trail of data. This data often shows up as first reports, followed by payment transactions that indicate workers' compensation payments, medical services rendered, vocational rehabilitation, and other important claim events. These patterns of activity provide the raw material that is used by the predictive model to score each claim from 1-1,000. Normal claims will typically have scores of 300 or less. But a high score is a signal that the claim is out of profile when compared to its peers. Each time the predictive model scores a claim, reasons are produced to explain the model score. This can provide useful information to assist in the analysis of a claim being suspected of being fraudulent, abusive, or exceptional by looking at the reason codes returned with the score.

Predictive models are accurate because the technology recognizes patterns from the data itself, not from pre-existing assumptions on what the data means. As a result, the system provides high-quality referrals to investigative units, further reducing losses. In addition, neural networks also are the only systems sophisticated enough to detect fraud types that have not been seen before. Claims are scored frequently to detect any change in the status of the claimant.

Finding Claims that Matter

Interestingly, only a small percentage of claims account for the majority of claims costs. These more costly cases include fraudulent, abusive, as well as legitimate claims that for various reasons require special handling. These exceptional situations could include a claimant who is not acting fraudulently or showing abusive behavior, but instead is someone in need of aggressive case management. Predictive analytics can help insurers identify these high-risk claims rapidly, allowing these exceptions to be routed to experienced case managers or investigative units, freeing adjusters to process the remaining claims with no outside re-sources.

The trick is separating these high-risk claims from the rest. Many fraudulent or abusive claims don't look all that remarkable at first, even to the eye of a well-trained adjuster. Clues may be subtle and submerged in an ocean of data. In the case of a bodily in-jury claim, for example, medical factors indicating a need for special handling may become evident only after some amount of treatment. In the case of fraud, opportunists who initially file legitimate claims may eventually fall prey to the temptation to exaggerate or misrepresent their cases. In fact, the majority of insurance fraud starts out this way.

Another factor contributing to high claim costs is that present methods are skewed toward apprehending fraud rather than identifying other types of high-risk claims. Predictive analytics are highly effective at reducing claims costs because they are extremely effective at identifying these high-risk exception claims, both legitimate and fraudulent. By accurately culling out high-risk claims, predictive analytics can make it practical and safe for insurers to process and close a vast majority of claims faster. Insurers save money by identifying suspicious and high-risk claims at the earliest possible moment, enabling preemptive action to be taken to prevent losses from occurring.

Advantages of Predictive Models

Predictive models don't replace the skills and experience of an adjuster. Instead, the power of predictive analytics augments the work of adjusters to work more efficiently and effectively. In a fraction of a second, predictive models can consider thousands of variables simultaneously, looking at complex relationships between data and deciphering subtle clues that might be missed by even the most seasoned adjuster.

These advanced methods are applied not only to the initial claim documents, but also to every transaction associated with the claim over its entire lifecycle. Each new piece of data coming in is analyzed thoroughly against not only the claim history, but a vast database captured from industry claims. In addition, predictive analytics are accurate because they are objective. Predictive models can recognize patterns from the data itself, not assumptions about it. This pattern recognition capability is dynamic; when the data indicates something new, the software updates its detection criteria.

Early detection is just one of the total benefits of deploying predictive models. Armed with evidence of fraudulent or abusive activity, claim adjusters can be proactive in managing claims more effectively. Often, fraud or abuse can be stopped quickly by a call or letter to the claimant, helping them understand that their activity is being monitored. Once contacted politely, malingering injured workers often stage miraculous recoveries from their injuries. This preemptive effect also prevents claim costs from growing. In the case of an injured worker whose activities are not fraudulent or abusive, claim adjusters can proactively partner with case managers to provide the care needed for better outcomes and a faster return to work.

Better Bottom-Line Results

Overall, insurers using predictive analytics software technology to detect fraudulent and abusive claims have experienced a re-turn on investment of 20-to-1, or up to $300 per claim in savings. In comparison tests, 55 percent of claims were identified by models weeks or months before they were discovered manually, and the software often discovered suspicious claims or claims needing case management that would have been missed by insurance claims analysts.

With the advent of predictive software models, insurance payers have a powerful tool in the fight against fraud. Losses and administrative expenses that were previously accepted as a normal cost of doing business can now be substantially reduced. Predictive models allow claim managers to make better decisions earlier, be more productive, and proactively prevent further fraud and abuse before it happens. The proper case plan can be established for each claim, including referral to SIU or special case handling.

It's an exciting advance in the fight against one of the largest challenges facing claim managers today. With predictive analytics, insurance adjusters have a powerful tool that can dramatically improve the climate of workers' compensation.

Kevin Lisle is product manager of property and casualty analytics at Fair Isaac Corporation. Contact In-formation: 949-655-3300, www.fairisaac.com, email [email protected].