The same technology that informs you when someone makes unusually large purchases on your credit card is now being tested in healthcare billing transactions. Its aim: to flag fraudulent or wrongful claims in a flash, benefiting both the providers worried about how long the adjudication takes and the payers worried about getting taken.
Despite their best efforts to prevent fraud, health insurers lose billions of dollars a year from schemes that go undetected or are discovered after the trail went cold, industry experts say. But those best efforts still constitute a drag on claims resolution, contributing to a payer-provider tension that has resulted in passage of prompt-payment laws in 48 states.
In this acrimonious climate, stoked by aggressive insurance commissioners and impatient provider associations, payers are finding ways to unearth fraudulent claims sooner.
For honest providers, this could mean fewer claims flagged and faster reimbursement. For outliers--those providers whose claims history ranks them outside the norm--it could either finger dishonest providers sooner or explain unusual billing patterns faster through better data analysis.
Crooked, sloppy and error-prone healthcare providers who rely on systemic inefficiency and antiquated fraud-detection systems to overlook their faulty claims may be surprised by advances in fraud-detection technology being adopted by insurers.
For example, a large insurer of federal employees is testing a new antifraud software system designed to flag fraudulent and wrongful claims before they are paid rather than months afterward, potentially saving the company millions of dollars.
"There's no question about it: A good prospective antifraud software tool will be the wave of the future," says Lawrence McEnroe, vice president of corporate services and general counsel at the Government Employees Hospital Association. "We all want to find fraud before it gets out the door. But right now there is no standard for this in the industry."
The case for early intervention
Most antifraud software products now on the market operate retrospectively, sifting through data on claims already paid to discover claims that should have been rejected, says William Mahon, president of the Washington-based National Health Care Anti-Fraud Association.
One drawback to the retrospective method, experts say, is that three to six months might elapse by the time the paid claims are flagged and forwarded to company special investigative units, allowing crooked providers plenty of time to leave town without forwarding addresses.
"Retrospective technology has, by its nature, some significant limitations," says Bob Greene, manager of data analysis for the GEHA, the third-largest fee-for-service health insurer for federal employees, covering 450,000 people. Greene also evaluates new technology for the organization.
"There's enormous potential in technology that can catch potentially improper claims before they go out the door," says John Bentivoglio, who served as special counsel for healthcare fraud at the U.S. Justice Department in the Clinton administration. "The value of a prospective program has been demonstrated in the credit card industry, and healthcare is no more complicated.
"The technology itself won't solve fraud problems for private payers or government programs," says Bentivoglio, now with the Washington office of law firm Arnold & Porter. "But it can make the process smarter and make it more difficult for fraudulent providers. And it can certainly benefit honest providers by reducing delays in payments that have been a source of frustration for providers for a long time."
HHS' inspector general's office last month estimated that Medicare improperly paid nearly $12 billion to providers in 2001. The private health insurance industry does not track how much it is annually defrauded, although Mahon estimates that the private insurers that his association represents lose billions nationally to fraudulent providers.
A product being tested at the GEHA, developed by San Diego-based HNC Software, is one in a new software generation aimed at prospectively analyzing claims data to detect fraud in real time.
Greene says a prospective tool can more swiftly and accurately pinpoint patterns of fraud and allow the insurer wiggle room to delay payments while investigators run further analysis on providers whose claims are flagged.
That capability becomes more crucial as healthcare makes the transition from paper to all electronic claims. On one hand, the move to computerized adjudication can greatly reduce the turnaround on claims payment. But without a means of sifting through electronic claims as they come in, increased automation works against the current methods of spotting questionable billings.
"Those transactions happen at the speed of light," Mahon says. "So the idea of a prospective system stopping a questionable claim before it goes out the door is attractive."
Greene says the number of eyes watching transactions is gradually decreasing.
"There is less actual human handling of claims now, and we're moving toward more automatic adjudication. Over half of the claims we pay now electronically," Greene says. "So we have less time and fewer people to make sure our claims are being paid correctly. That's why I'm involved. I'm trying to find better tools that will screen claims automatically."
The Health Insurance Portability and Accountability Act of 1996 requires the nation's health insurers to implement electronic data capability by October 2003. "The payers do not have the option of not paying claims (electronically)," Greene says. "So they must adopt systems that pay quickly but accurately."
Besides the mandates of HIPAA and prompt-payment laws, a growing number of states--most recently New Jersey--are requiring insurance companies to deploy antifraud programs approved by state insurance departments. Since the mid-1990s, the Centers for Medicare and Medicaid Services has required Medicare contractors to install antifraud technology products.
Sifting through millions of claims
The GEHA's special investigative unit comprises a staff of 10 and must peruse an estimated 6 million claims annually, a gargantuan task that requires a focused effort, Greene says.
The effort is complicated by the company's broad reach, with members in every state. Although there are advantages to having a national presence, it presents problems as well, particularly in the world of fraud.
"Because we're spread far and wide, we don't represent a lot of any single provider's business. Most of the existing technology is based on statistical models and rules that require you to collect a lot of data on a single provider to be effective," Greene says. "That's something we don't have. We've looked at retrospective technology and found for the price point it was difficult to make a solid case for return on investment."
He says the company was seeking something more proactive, not reactive, a cutting-edge system that could prepare provider profiles and assist the company's special investigative unit. He heard about HNC's new product more than a year ago and began to pilot the program in December 2001 to see what the prospective product had to offer.
"They pioneered this technology in the financial services market and thought a prospective system would represent a step forward," Greene says.
"The financial industry relies on prospective fraud technology and have validated and quantified its worth," Bentivoglio says. "Those same benefits can translate to healthcare where the dollar amounts are similarly large and the potential for reducing fraud similarly great."
GEHA Vice President McEnroe says crooked providers have hit his company with the same schemes afflicting other payers.
Although he declines to speculate on the amount of revenue lost annually to fraud, he acknowledges: "We know it is large and we know it's in the millions. We know there are fraud rings, and organized crime has gotten involved because it's just so lucrative."
He says the schemes range from upcoding and unbundling claims to abusive laboratory charges and falsifying records to qualify for reimbursement. For example, he says, some cosmetic surgical procedures like a rhinoplasty (nose job) are disguised or fabricated to appear medically necessary to obtain payment for services not covered under a plan.
Betty Drake, who manages the GEHA's special investigative unit, says her office encounters falsified infertility treatments, charges submitted for more extensive services than delivered and bills for medically unnecessary services.
"We've had people break into doctors' offices and steal provider and patient identification numbers and flood the plan with bogus billings," she says. "We think the prospective review will see much of that upfront before payment is made."
Keeping a running score
An investigation begins with the premise that the provider is honest but the claim is unusual, says GEHA investigator Katie Pritchett. There isn't a presumption of fraud.
"We start out giving them the benefit of the doubt," she says. "We want to prove their billing is correct rather than fraudulent."
So the department runs a search on the number of tax identification numbers a provider has, how much the provider has billed the company, how many offices the provider operates and how many patients the provider treats. The GEHA contacts members to validate claims information.
"Sometimes we never even contact the provider," she says. "We want hard-core facts before we seek (providers) out."
Pritchett says physicians and physician groups, rather than hospitals, account for most of the flagged claims. Because the GEHA represents government employees, the company benefits from access to government investigative resources such as the FBI and U.S. postal inspectors.
"Right now we get referrals from our claims office, from our members and from other government agencies or plans investigating providers," she says. "Sometimes we get requests for information from the Office of Personnel Management inspector general."
The system from HNC, called Payment Optimizer, compiles a profile on each network provider, for whom it develops a cumulative composite score. Based on the insurer's history with the provider, the system will track the number and types of claims and seek out patterns of clinical and billing behavior.
For example, it will ask if a doctor regularly bills on Sundays or charges for services that are obviously inappropriate--billing prostate exams for female patients or Pap smears for males. The system detects patterns of billing, flagging especially unusual ones. For example, has every patient undergone the same test, procedure or diagnosis? Each transaction becomes a part of the profile fabric, a constantly evolving dynamic. Based on that experience, the provider is ranked and scored every time a new claim is submitted.
Greene says once a provider achieves a certain score, an investigator is assigned. "Right now we're looking at a 50/50 false positive," he says. "That's really great compared to the way we've done it in the past, claims we've been reviewing internally. All of these are unusual enough to merit attention, but half are worthy and good cases deserving attention. Our early returns are very positive."
Drake says the HNC pilot program is being tested on previously paid claims so that current GEHA providers are not adversely affected by a new system that still might have a few flaws.
"There's not enough sampling yet to make a determination," Drake says. "We're not on live claims yet. But we think based on the kind of information this gives us that we'll be able to evaluate it on paid claims data. We like what we're seeing so far."
Pritchett says she's excited about the brief sampling done retrospectively.
"They've been able to dissect, slice, dice and review data on these providers. We've been able to look at frequently repeated codes and providers with high-dollar exposure with us, and it's working very well so far."
Catching up to other industries
Healthcare is only beginning to apply the power assist of technology that's available to the financial-transactions industry with the swipe of a credit card, says Andrea Allmon, product director for HNC's Payment Optimizer.
She says more than 85% of U.S. credit card transactions employ HNC technology, which allows a business transaction to be scored for fraud risk in real time.
As healthcare moves toward real-time transactions to deliver electronic payments to providers, it opens a huge risk to payers, Allmon says.
"Someone can come in and take out large amounts of money in what are called `bust-out schemes,' large numbers of claims submitted quickly from crooked people, not even necessarily providers, who just walk away once they've been paid," she says. "Those kinds of patterns will evidence themselves more quickly with a prospective system."
Allmon says the Payment Optimizer analyzes patient interactions, not just provider profiling, to give information about whether the care is delivered in an appropriate context.
"We can look at everything that has gone on between providers and their patients in the past, and even though this is transpiring in real time, it offers a richness of history," she says. "In the second stage, artificial intelligence technology is brought to bear on the transactions."
The combination tells where the fraud is occurring and picks up new and emerging schemes.
"It's the ones you don't know that will scare you," Allmon says. "This system can find programmatic weaknesses sooner, as well as providers doing things that might not amount to fraud but may require further education. This system accomplishes in less than an hour an investigation process that used to take four or five weeks."
"This is a sharper instrument to identify unusual behavior," Greene says. "It's not just a blunt ax. It's a refinement toward identifying aberrant provider or pa"tient behavior."
Most GEHA providers are honest and likely will never be flagged, McEnroe reiterates. Less than 1% of providers fall into the outlier category.
"Those who are within the norm will have their claims paid routinely, and we will focus our reviews on those who exceed the norms," he says. "Their claims will be (held) while we conduct the reviews. The GEHA is a not-for-profit membership association. We are owned by our members and we pay their claims as quickly as we can."