In the Kernan case, the government claims the number of patients who had kwashiorkor as a secondary disease increased a hundredfold — from three in 2005 to 358 in 2008 — making up more than a third of all Maryland cases, according to statistics from the state Health Services Cost Review Commission.
In court documents arguing against UMMS attempts to dismiss the case, federal officials said Kernan systematically upcoded and that the patients were not malnourished. One patient, they said, was 5 feet 2 inches tall and weighed between 170 and 200 pounds.
The documents said that after Maryland's cost review commission implemented new rules to make secondary diagnoses more financially rewarding, Kernan trained its staff about malnutrition as an important and common secondary diagnosis. The government said physicians were sensitized to Kernan's goal to increase compensation by using the diagnosis. The hospital also did no auditing or monitoring of the secondary diagnosis even as it increased significantly, court documents say.
While not claiming their patients had kwashiorkor, UMMS officials said some patients do have severe nutrition deficiencies that share the same coding. Such patients may have been in car accidents and then been treated in the hospital for a long period of time, attached to a feeding tube.
The officials say there was no intent to defraud the system and that is why they are fighting the claims.
"If you wanted to make it appear as though there was over-coding, we said we'll dispute that," Chrencik said. "We're not going to roll over."
The Justice Department would not discuss the case because it is still pending.
Some say the government shouldn't be so quick to allege fraud.
A trade group that represents medical colleges, including the University of Maryland, said that coding can be a complicated process. Ivy Baer, director and regulatory counsel for the Association of American Medical Colleges, said hospitals aren't trying to deceive the system. They frequently audit their systems and are on constant lookout for mistakes, she said.
Faulty data input can be attributed to human error or poorly trained staff, Baer said.
"When the government does an audit, a lot of times they are running data and just looking at what is being coded," Baer said. "Even if they see an unusual pattern and it raises a red flag that doesn't mean there is fraud. When you look at it, it may appear as if there is fraud until you gather all the facts."