Common hospital-acquired infection rarely reported in the dataset used to implement hospital penalties

But a statewide analysis by the University of Michigan shows there was very little change in hospital payment due to removing pay for hospital-acquired catheter-associated UTIs. For all adult hospital stays in Michigan in 2009, eliminating payment for this infection decreased hospital pay for only 25 hospital stays (0.003 percent of all stays). This is in great contrast to the large savings anticipated, given that this condition accounts for nearly one third of all hospital-acquired infections.

The reason, U-M authors say, is that the "no-pay" policy uses billing data that is inaccurate for identifying such complications.

The findings were published in the Annals of Internal Medicine, alongside an editorial that emphasizes the national policy implications of the U-M findings. U-M's analysis, the editorial says, argues that careful dataset selection is crucial when used to measure and penalize hospital performance.

"We think the policy was well intended but its financial savings from non-payment for catheter-associated UTI are negligible because of the data used to implement the policy," says lead author Jennifer Meddings, MD, MSc, an assistant professor in the Department of Internal Medicine, Division of General Medicine at U-M Medical School.

The problem, authors say, is that the policy relies on claims data previously used only for billing -- but multiple billing codes must be listed correctly to indicate that a UTI is due to a catheter and occurred only after admission to the hospital. As a result, most cases are listed as simple UTIs and hospitals continue receiving payment as usual.

"If the cases you wish to penalize are not documented in the dataset chosen, the policy's intended impact will be limited," Meddings notes.

But the blame doesn't fall on billing coders who generate the claims data, authors say, as coders can only list the diagnoses as described by physicians in the medical record (following federal guidelines.)

Meanwhile, the well-publicized policy implemented by the Centers for Medicare and Medicaid Services has expanded to include more hospital-acquired conditions for nonpayment, such as pressure ulcers (or bed sores). But the authors warn that similar coding problems with other conditions may also stymie financial savings.

"We don't have any evidence on whether it has prompted hospitals to improve patient care to prevent these infections, but we do know that it did not lead to large financial savings," Meddings says.

The findings suggest that billing data is inaccurate for comparing hospitals by their catheter-associated UTI rates. Because billing data is anticipated to be used in 2015 to penalize hospitals with the highest infection rates, authors say, hospitals that report accurately in claims data will be unfairly penalized because their reported rates will be higher.