17 November 2010—On 13 October 2010, the Federal Bureau of Investigation arrested 55 people involved in a scheme to defraud Medicare, the U.S. government’s health insurance program for the elderly and disabled, of US $163 million. It was hailed as the largest Medicare fraud bust in U.S. history until a week later, when federal agents took down a separate group, in Florida, that had tried to defraud the Medicare system of $200 million. In both cases, the government had already paid out millions of dollars for medical services that were never provided. Most of the money will never be recovered.
These kinds of schemes cost the U.S. government billions of dollars each year—there was an estimated $47 billion in improper Medicare payments in 2009 alone—so it should come as no surprise that cracking down on fraud has become a key part of the White House’s health-care reform efforts. The goal is to reduce improper payments by half by 2012. The biggest challenge facing Medicare fraud investigators is the sheer size of the system. Sorting through all the data can take months. Phony clinics, such as the ones the FBI shut down in October, often have collected their money and moved on to new locations by the time investigators discover the fraud.
The Obama administration has put considerable energy toward fixing these problems. For example, the government is working with the database-software company Teradata to merge all the government health-care databases. Fraud indicators, such as multiple wheelchairs for a single patient, should be easier to find automatically in a merged database, but investigators need better ways of weeding out the false positives, says Rasim Musal, an assistant professor at Texas State University, in San Marcos, who has worked in the health-care fraud detection field.
One of the most promising solutions to this problem is the use of geographic information systems, known as GIS. Medical insurance claims are typically organized by patient information, but the geographic data surrounding a patient’s medical care may actually tell a more complete story, says Bill Davenhall, a health-care expert at Esri, a company that specializes in GIS. "Analysts don’t always see geography as a predictive tool, but it can be very useful," he says.
Fraud investigators already use postal codes to identify suspicious claims, but the geography of Medicare is complex. Medicare covers medical treatments received out of state, and physicians can practice in more than one state at a time, which means both physician and patient could be located in a different state from the clinic filing the claim. "It’s not unusual to have six different addresses by the end of the billing process," says Davenhall. With GIS, investigators could narrow the focus down from postal codes to street addresses, which would help investigators catch fly-by-night operations faster.
Incorporating health statistics can make geographic-based models even more powerful, says Davenhall. For example, he says, "a patient flying from New York to Florida for an elective procedure might not raise any red flags, but studies have shown that compliance goes down the farther a patient is from the clinic. So if that patient is going back for follow-up visits every week, that’s cause for suspicion."
Geographic data are informing the government’s attempts to prevent Medicare fraud, too. The Department of Justice has established regional Medicare task forces in the geographic areas most associated with fraudulent activity, and the government is in the process of using one of Esri’s GIS templates to map calls to the national Medicare fraud hotline in real time. Davenhall suggests that providing this information to the public could speed up fraud investigations even more. "The more eyeballs that see this data, the better," he says.
About the Author
Erica Westly is a freelance science writer based in Brooklyn, N.Y. In the July 2010 issue she reported on how engineers use technology to demystify the black art of barbecue.