Fraud prevention: Dragons at the gate

Governments and businesses use data to combat fraud.

by Patric Helmaan

Fraud is the menacing, monstrous dragon that eats a substantial percentage of the world's public and private sector economies every year. As businesses look for ways to keep the dragon outside, many find data warehousing makes their task exponentially easier.

In both government and enterprise deployments, data warehouses are playing increasingly important roles in identifying fraud—ferreting out non-compliant taxpayers, for example, or red-flagging credit card usage patterns indicative of card or identity theft—and in preventing fraud in the first place.

Factor in abuse and waste, whether deliberate or inadvertent, and you have a global, multi-trillion dollar problem that robs businesses of profits, steals legitimate benefits from deserving people or organizations, undermines public confidence in government institutions and private businesses, and adds to the cost of products, services and public programs.

"There's no doubt that the problem is huge," says Thom Rubel, former vice president for government services at META Group, a Stamford, Conn.-based consulting firm. "But there's also no doubt that things are getting better quickly."

The source of the improvement? Rubel believes it's data warehousing. "Technologies that allow information sharing and analysis of comparative data are precisely the tools governments need to track fraud and, more importantly, to catch potential fraud up front and prevent it," he says.

Rubel notes that the trend toward using data warehousing in the government agencies and departments that fight fraud is accelerating. "Naturally, many of the earliest implementations were in the private sector," he says. "That's understandable: Every enterprise dollar lost to fraud is a dollar of lost profit. Government, on the other hand doesn't go out of business if it doesn't make money."

But as public sector IT budgets have improved in recent years—and as data warehouse-driven anti-fraud initiatives have delivered striking successes—the technology has established a large and growing presence at both state and national levels.

"There are two large drivers for government IT investment: Reducing the cost of government itself, and improving the quality of services and their delivery," Rubel says. "In both areas, the ability to identify and prevent fraud is undeniable."

Private sector initiatives have brought data warehouse resources to bear upon scams and schemes ranging from insurance fraud to embezzlement, from accounting irregularities to governance violations and more. Similarly, public sector programs are attacking fraud on a number of fronts.

Among the programs generating the greatest success:

• identifying non-compliant taxpayers at both state and national levels
• spotlighting multiple billings for services
• locating deadbeat parents
• tracking illicit prescriptions for pharmaceuticals
• weeding out fraudulent benefits applications
• reconciling tax and wage figures with benefits applications
• eliminating bureaucratic and communications bottlenecks.

In these areas and hundreds more, the flexibility provided by a robust data warehouse has empowered users to uncover avenues of fraud and misuse that were previously invisible.

"Before the data warehouse, it could take several weeks just to set up the preliminary parameters for a fraud investigation," notes Steve Puleo, director of communications for Bull Services, a Massachusetts-based public sector information solutions company. "Even a simple investigation of transportation billing patterns, for example, required notifying IT specialists of the parameters, waiting for the data to be assembled and run, then reviewing the results."

Those results, according to Puleo, could amount to thousands of pages of reports. Analysis and comparison of the data contained in those reports was often done by hand, further slowing the process.

Data warehousing technology has changed that. For one thing, Puleo points out, the sheer flexibility of the data warehouse platform makes it possible to aggregate and compare information in ways that were simply impossible with previous systems.

"Departments and agencies locked into rigid data architectures will never be able to benefit from the fraud identification techniques and strategies that data warehouses enable," he says.

Equally important, in his view, is the liberating effect the data warehouse provides department personnel. Word gets around to other agencies and departments, encouraging the addition of more data from more areas, which in turn empowers more approaches to eliminating fraud, abuse and waste.

That empowerment comes directly from the user focus at the heart of the data warehouse deployment. "In agencies and departments with data warehouse capabilities," Puleo continues, "the investigative power now rests with the actual end users. And because of the speed and agility the warehouse provides, it's absolutely possible for a user to wake up in the morning with an idea for a new area of inquiry and begin that investigation almost immediately."

The ad hoc query capability, in fact, may be the most effective anti-fraud tool in investigators' arsenals.

"Often you don't always know what sorts of questions you're going to ask, or even what sorts of questions need to be asked, until you're well into an investigation," Puleo says. "The ability to follow a line of inquiry down a particular road and then very quickly shift to another path if necessary is a huge advantage that the data warehouse provides, one that can pay enormous dividends."

The variety of inquiries and investigation paths are themselves exponentially increased as more and more information from a broader range of departments, agencies and pre-existing databases are added to the data warehouse's view.

"That 'single view of data' really means that for the first time a government's left hand can know what its right hand is doing," says Rubel.

Closing the gate
Counter-intuitively—especially for those whose intuition was honed on old-style rigid data architectures—the more information that's loaded into a fully capable data warehouse, the more that can be done with it quickly—with more results achieved in less time.

"Wider varieties of data offer larger advantage," notes Brian Minnix, president of Woodbury, Minn.-based data warehouse tax solutions software company Minnix Software, Inc. "External, internal and departmental data can all be combined to find more non-compliant taxpayers." The same is true for other types of fraud.

For example, the State of Michigan, a Bull Services customer, was able to track parents who owed back child support. The increased capability came as a direct result of the State of Michigan's consolidation of departmental data into a Teradata Warehouse. In addition to seeking those parents through tax and drivers' license records, investigators correlated names with hunting and fishing licenses, producing immediate results and enabling the pursuit of payments from parents who had previously been hard to find.

"Michigan is an excellent example of the efficiencies achieved through inter-departmental data," Puleo says. "Five of the largest departments in the state are now included in the data warehouse, with a total of 8,000 users, each able to apply their skills to exceptional data resources. And of course, you couldn't add that data in the first place if you didn't have the data warehouse's flexible architecture to receive it."

Some estimates place Michigan's cross-departmental ROI as high as $250 million per year, or approximately $1 million per business day.

While some ROI is achieved through revenue re-capture—back taxes, prosecution fees or fines for illegal claims against healthcare and other services—the largest returns are earned in money not lost to fraud and abuse.

"Simply in terms of identifying and preventing false claims and ineligible applicants up front, the data warehouse provides a valuable service," Rubel notes. "By catching fraud before it occurs, the data warehouse makes it possible to catch potential fraud up front, to mark it and stop it before it takes place. Obviously, that eliminates fraud faster than discovering it after the fact, when it is more difficult to recover overpayments and false claims."

Puleo offers another preventative advantage. "I don't think we fully appreciate the 'sentinel effect' of the data warehouse yet," he says. "But as agencies (and businesses) become more and more effective at using the data warehouse to ferret out, eliminate and prevent fraud, word gets around. People become wary of trying to commit the fraud in the first place."

While enterprise and business anti-fraud efforts are often unsung (for obvious reasons, most institutions are more than reticent about revealing how much money they lose to fraud each year), government success stories are another matter.

"Getting the word about data warehouse success out to the public in a good, coherent way helps to instill confidence," Puleo says. "It tells the public that government is performing its functions more effectively and efficiently, which is one of the large justifications for investing in the data warehouse."

By eliminating fraud and, increasingly, preventing it from being committed in the first place, governments are better able to focus on the provision of services and better equipped to see that government dollars are used properly, efficiently, and effectively.

The money and services, in other words, reach those who need them most, rather than being lost to the dragons of fraud, abuse and waste. T

Case study: New York State's Department of Health

As the largest Medicaid program in the United States, New York State's Department of Health (DOH) processes more than 300 million claims from 3.7 million participants annually. The price tag? A staggering $41 billion per year. Even a single percentage point of fraud would remove more than $400 million in critical-care funding from those least able to afford the loss. The solution? A DOH-commissioned Bull Services enterprise data warehouse—eMedNY—designed to consolidate hundreds of millions of disparate Medicaid records into a single repository, accessible and usable by more than 650 staff members across 10 state offices and agencies, one federal agency and 17 counties. More counties are being brought into the eMedNY program every year, and the number of users is expected to exceed 1,100 in the near future. Currently, tools available to the DOH data warehouse's users include analytical software for spotting trends and forecasting their outcomes, mapping software that provides geographic portraits of the data, and business intelligence and query tools that enable users to query the data directly. Historical perspective is provided by more than five years' worth of data—a collection of more than 1.4 billion claims—that helps the DOH understand present needs and operations and forecast future needs. Data mining tools enable fraud detection and elimination as well as decision support and utilization forecasting. Additionally, a library of 180 executive-level and subject-specific reports helps to simplify the creation of the most critical and frequently requested reports. Reports are refreshed as new data is added. The results? In its first year of operation, eMedNY processed more than 150,000 queries, resulting in: More than $16 million in savings from the identification of 3,800 Medicaid claimants actually covered by managed-care plans At least $5 million recouped in less than six months through the analysis of duplicate payments More than $1 million recovered from inappropriate clinic billings $63 million in savings through decision-support analysis related to policy decisions and implementations $30 million saved by moving the growth hormone Serostim to a "prior approval" list The program also: Increased efficiency of resource allocation: New York's Office of Alcohol and Substance Abuse Services (OASAS) found that 12% of its clients accounted for 52% of the office's costs; further analytics are being aimed at increasing the efficiency of treating that 12% and reducing their overall cost to the system Identified overlap between home care and hospital claims, eliminating double-dipping Built a predictive model to identify those most likely to commit fraud or abuse Enhanced auditability and control, enabling audit queries to be posed and answered in less than an hour. Additionally, eMedNY collects acute-illness and pharmaceutical purchase information, as well as emergency room data, as a means of identifying and/or tracking potential bio-terror attacks or incidents.

Case study: Michigan's Department of Community Health

The Department of Community Health (DCH) is the largest of the State of Michigan's departments, serving more than 1.2 million clients with an annual budget of $9.5 billion. DCH, which is responsible for many of Michigan's most critical health programs, including Medicaid, WIC (Women, Infants and Children) and children's immunizations, took an enterprise-wide approach to its Bull Services data warehouse, which has generated tremendous efficiencies and savings. Its development process has served as a model for other states considering an enterprise-wide approach to consolidating information and increasing data usefulness. Michigan's data warehouse was implemented in carefully planned phases over the past decade. The first phase of the project brought together the state's Medicaid and Public Health data in the Department of Community Health, using the data warehouse's resources to generate reports, detect fraud and abuse, and model programs. Phase two broadened the scope of data with the addition of state welfare eligibility information for reporting and budget/policy analyses. During this phase, Michigan's Department of Treasury added its taxation and revenue audit information and solutions to the data warehouse as well. The third—and current—phase includes increasing the number of data sets loaded into the data warehouse. The information enables Michigan's "unique client identifier" process, an effort to dramatically increase the effectiveness of programs, the granularity of information available and the savings generated by government agencies by linking and tracking individuals across programs as well as within specific programs. For example, one initiative will further enhance the ability of the data warehouse to support troubled schools—and individual students—by linking Medicaid and other health data with child law enforcement and education data. Thanks to its robust data environment, DCH has: Reduced Medicaid administrative costs by 25% Saved $75-100 million annually through forecasting, program overlap and service quality assessments Doubled the rate of fraud identification Developed the ability to track expenditures directly back to a specific initial claim, and to do so in minutes or hours rather than days Coordinated communication and distribution of death certificate information, overcoming an existing bureaucratic problem Eliminated $4 million in disallowed Medicaid claims to prisoners Assigned unique client identifier (UCI) numbers to each client, permitting tracking of clients through and across overlapping programs Developed predictive analysis of Michigan's healthcare needs Aggregated more than 800GB of data that is now accessible by more than 200 users Implemented access tracking to ensure accurate records are kept of all data accesses—a crucial achievement, given heightened concern for privacy and security of medical and other records

*teradata.com, "Data Warehousing Technology May be the Solution to $600 Billion Fraud Problem"

Patric Helmaan is a writer and consultant whose work deals with the acceleration of technology and the challenges of technological innovation.