Breaking up interagency data silos helps the state of Michigan catch cheaters.

Each year, the Michigan Department of Human Services (DHS) grants approximately $5 billion in public assistance to some 1.2 million Michigan residents — including some people who are lying to get it. Although the information that investigators need to help expose fraud is available, until recently they couldn’t easily match up data because it existed in different agencies’ information systems.

The Michigan Department of Information Technology (DIT) has been busy remedying that problem by integrating data from multiple repositories into a single data warehouse at the DHS. It has also put business intelligence query tools and reports into the hands of hundreds of investigators and staffers for the DHS Office of Inspector General (DHS-OIG), who identified $8.7 million in fraud last year. Most of that was associated with the state’s DHS Child Development and Care assistance program.

“We [at the DIT] help agencies share data across the enterprise to produce these kinds of results,” says Ken Theis, chief deputy director at the DIT.

The day care program, which provides assistance to working parents, had been expanding rapidly for the past five years. “We had a feeling this was a problem,” says Dave Russell, department specialist at the DHS-OIG, adding that some applicants were creating fake pay stubs using home computers. “We’ve tossed around different ways to figure this out, but we’ve never had the tools to help.”

The DHS uncovers cheating by comparing quarterly wage data from the Michigan Department of Labor & Economic Growth (DLEG) to see if recipients of day care assistance are indeed working. Before, the agency relied on calls to a hot line or worker complaints. Now, the DHS receives quarterly electronic updates from the DLEG and uploads those into its own data warehouse. Using query tools that DHS-OIG staffers built using Business Objects software, investigators drill down to identify suspicious activity, such as recipients who earned less than $500 in wages but received a high dollar amount of day care reimbursements.

“We’re able to target the higher-dollar areas,” Theis says, and investigators are now more efficient at finding fraud. Detection rates rose in the first two years of the initiative — identifying $3.3 million in fraud in 2004 and $9.2 million in 2005 — but the numbers dropped back a bit in 2006, to $8.7 million. Publicity about the program and subsequent prosecutions may have contributed to the drop. “Fraud is starting to go down because the word is getting out,” Theis says.

While several hundred staffers and investigators can create their own queries, the DIT also puts reports into the hands of more than 3,000 DHS staffers as downloadable PDF or spreadsheet documents.

“The fact that critical data, such as wage data, is locked in complex legacy applications makes progress very difficult,” says Gene Leganza, an analyst at Forrester Research Inc. “Michigan’s ability to build a data warehouse across programs and departments definitely puts them in the front of the pack” compared with other states.

The DHS had a Teradata data warehouse and Business Objects business intelligence tools in place before it began the project. “It wasn’t a matter of bringing in new technology but bringing the pieces together,” says data warehouse contractor Jim Davis.

The DHS has used the technology in other areas as well. Theis says the DHS saved $1.6 million last year by identifying and removing from welfare rolls people who had left the state but were still receiving assistance. The DIT integrated data from an electronic benefits transfer vendor (benefits are provided on a debit card), and investigators could then create queries to discover which clients were using their benefits in out-of-state locations.

“If they’re not a resident, it’s a resource drain,” Theis says, citing the state’s role in paying out Medicaid premiums to HMOs, administering the issuance of federal food stamps and providing some direct cash assistance.

Now the DHS is working with the federal Public Assistance Reporting Information System, an information exchange established by the U.S. Department of Health and Human Services. The service matches up Michigan DHS data with that from other states, as well as with federal wage and Veterans Benefits Administration data. “We’re developing our query tools and will be using the data warehouse,” says Theis.

To Theis, the success of the initiative isn’t so much about catching fraud as it is making sure resources get to those who need them. “It is enabling DHS to direct more and more resources toward their clients during a time when their need is the greatest and funding is stretched,” he says.

See more of the Best in Class special report.

See the list of 2007 Premier 100 winners.

Copyright © 2007 IDG Communications, Inc.

Bing’s AI chatbot came to work for me. I had to fire it.
Shop Tech Products at Amazon