Predictive Tools, Preventive Practices: Strategies for Winning the Battle Against Insurance Fraud
There’s no sugarcoating it: Insurance fraud hurts everyone.
“As fraud is perpetrated, it increases claim costs, which increases overall risk, which increases premiums, which increases the cost of goods and services performed by the companies paying the premiums,” said Jon Shifflett, AVP, investigative services, CIS, at AmTrust.
“So, the cost trickles down to, really, everybody.”
However, “we’re seeing … the industry making significant progress in identifying and pursuing those who are perpetrating fraud,” Shifflett added.
That’s great news, considering the large monetary consequences that stem from intentional acts of insurance fraud.
According to Coalition Against Insurance Fraud (CAIF) data, an average $308.6 billion is stolen from American consumers each year thanks to insurance fraud. Medicare fraud costs an estimated $60 billion every year. The National Insurance Crime Bureau (NICB) reports property/casualty insurers paid $4.6 billion to $9.2 billion extra in disaster claims in 2021 — a price tag directly felt by policyholders upon renewal.
Fraud is not a new risk for the industry, but it is one that will never go away. Therefore, it’s important to remain vigilant, understand how perpetrators are trying to shirk the system and stay on top of today’s methods of prediction and prevention.
The term “insurance fraud” can mean any number of crimes.
In workers’ compensation alone, where the CAIF estimates $32 billion is stolen every year, we could see any number of frauds on a single claim: from double dipping, where an employee collects temporary disability benefits while also continuing to work and earn a paycheck, to medical provider fraud, where corrupt medical providers take advantage of the system and inflate medical bills.
Then there’s those who will claim P&C damage after natural disasters that never touched their property. Agents who will pocket premiums and business owners who withhold information on the jobs employees are doing.
Then there are those who will stage a car accident, commit arson, allow deliberate identity theft for health care coverage and so on and so on and so on.
“There’s always going to be fraud,” said Clayton Matthews, claims director at Oakbridge Insurance. “The end goal is always to deceive in order to try to gain something, be it monetary or otherwise. Sometimes, people might not even know they are committing the fraud.”
For the latter, Matthews gave the example of someone not adding a newly minted 16-year-old driver to an existing auto policy or not reporting the age of a roof to their insurance company.
“Maybe, for reasons unknown, they chose not to report the number of employees on their payroll,” he gave as a third example. Whatever the omission, these intentional or unintentional acts shirk the system, affecting costs and adding to the overall rate of fraud crime.
Again, none of this is news to the savvy insurance professional. What is new, though, are the methods fraudsters are using to cheat the system.
“Every time you think you’ve seen it all, somebody comes up with something new,” Shifflet said.
Looking at workers’ comp as an example, the latest trend he’s seen includes perpetrators, in post-termination or post-layoff claims, collecting total temporary disability while working another job.
According to Melissa Segel, partner at Swift Currie and a specialist in insurance fraud, cyber-enabled fraud is also on the rise.
“We’ve seen huge growth in cybercrime, and that really impacts all areas of the industry,” she said. The rise, she added, is due in part to the nature of how the internet works. “Anyone can be behind the scenes where they’re just working off their home computer and making things up,” Segel said.
The internet has the answers, she said. Again, anyone sitting at their home computer can Google how to commit insurance fraud and find step-by-step guides online. It opens the door for more crime from any type of fraudster from any background with any level of criminal ability.
So, how can the industry cope?
Step One: Predicting
Much like how fraudsters are turning to technology for guidance, so too can insurance professionals. (Though, a Google search alone might not yield a step-by-step guide this time around.)
Other tech tools, however, can prove meaningful in the fight to stop insurance fraud in its tracks.
“We have clients that focus on analytics. And just like the economist who can predict how the up- and downswings in the economy can impact the price of flour, analysts can predict potential insurance fraud with analytics,” said Segel.
Copious amounts of data can show patterns in fraud schemes. As the data is plugged in, analysts reviewing it, or even artificial intelligence and machine learning programmed to analyze it, can spot out-of-the-norm reporting based on previous fraudulent acts. And, hopefully, this can lead to predicting and preventing potential fraud.
“Many of us are using internal data scientists who are familiar with our claims and underwriting data and are developing our own predictive models,” Shifflett said.
“When developing a predictive model, it’s important to have strong data sets so the model can learn the unique elements of cases that have been accepted by your SIU [special investigation unit] for an investigation compared to those cases that were not.”
For smaller institutions that may not have a team of internal data scientists at their disposal, Shifflett added there are third-party vendors offering such tools. In this instance, he suggested insurers review their options to find agencies that meet their quality and standards.
“Each carrier is going to look at its lines of business that it underwrites and evaluate anomalies and scenarios that would indicate suspicious activity,” he said. “So many companies out there have developed tools in the industry to help identify these suspicious indicators using AI and algorithms. You want to review what they’re doing and match it to your needs.”
A subset of analytics tools, behavior analytics is yet another way insurers are training AI to predict potential fraud. This type of analytics provides insight into the end user, what they’re clicking, what their browser history reports and other meaningful information that shows what the person is up to and how they’re spending their time.
A tool like this gives insurers insight on a claim’s trustworthiness based on the user’s clicks, interests, location and other behaviors leading up to making their insurance claim.
Step Two: Preventing
The good news with analytics is that repeat fraud types will, more likely than not, be spotted by the predictive models. However, new methods of fraud might not be obvious to AI, and so insurers still need to employ other tactics to curb fraud.
Enter prevention. Perhaps the most important thing to remember in stopping fraud is prevention. No matter the insurance line — from comp to health care to P&C — there are several practices firms can implement to prevent fraud.
“It starts with training,” said Shifflett. “Getting in there and developing a good training program to give your adjusters a foundation to identify suspicious indicators and make referrals. The human element is going to be able to bring you the most information, the best information.”
Agents are often the first line of defense when it comes to fraud. As the people who intake the claim, they are speaking directly to the claimant in real-time, sometimes within hours of when the reported loss occurs.
Training these agents to spot misleading information or information that just doesn’t add up can go a long way in the fight to prevent fraud.
That first phone call with the claimant is also key, as several experts noted.
“It’s old-fashioned intelligence — just talking to people on the phone. You can pick up on so much of what people are saying by asking the right questions … If we’re asking the right questions up front, we’re not eliminating the fraud, but we are making it very difficult for bad actors to create the mayhem that they do,” Matthews said.
He further explained that letting the potential fraudster speak freely can lead them to reporting facts that misalign as they try to keep their story straight: “People will often tell on themselves,” he said.
This is why training is so vital: If the right questions are being asked and the initial report of a claim is being thoroughly reviewed, the potential to prevent fraud increases exponentially.
“There’s a number of things insurance companies can do to prevent fraud. Namely, be diligent and prompt. Ask all the questions possible in the first phone call. Commit to that early, deep investigation,” Segel added. &