Record-keeping for a patient complication used by Medicare to determine how much hospitals get reimbursed is not comprehensive or accurate, undermining the policy's value, a new study suggests.

In an effort to get more for their money, the Centers for Medicare & Medicaid Services (CMS) - as well as some individual insurance companies - don't reimburse hospitals for certain conditions listed on billing records that are costly and believed to be preventable.

One such example is urinary tract infections (UTIs) acquired in the hospital as a result of using a urinary catheter.

Such complications are "very common," according to study author Dr. Jennifer Meddings, from the University of Michigan in Ann Arbor - and at least on the surface, penalizing hospitals by not reimbursing them seems like a good idea.

The problem, she said, is that while the particular nature of a UTI - such as whether it happened in the hospital - shows up on a patient's medical record, it doesn't always make it into the billing information used by CMS to allocate payments.

"If you just looked at the billing data, you would already think that hospitals are doing a great job to prevent this complication," Meddings told Reuters Health.

She said the policy may penalize hospitals that are getting more information into the billing records.

"Their data gets translated into higher rates (of hospital-acquired UTIs) in their claims data, so they look like a bad hospital," Meddings said. "You certainly don't want to discourage hospitals from documenting well."

She and her colleagues used data from 96 hospitals in Michigan to determine how often different types of UTIs were listed in patient billing codes in 2007 and 2009. The CMS non-reimbursement policy went into effect in 2008.

Their records included close to 800,000 patients each year who stayed in the hospital at least two days.

The researchers found that UTIs were common in the billing records - about one in ten patients had one - but very few of those were listed as related to catheter use. There were only 25 cases in 2009 in which hospitals were reimbursed less than they would have been because of hospital-acquired UTIs related to catheters, Meddings and her team wrote this week in the Annals of Internal Medicine.

She said that for a hospital to lose payment for the complication, the billing record must clearly show that a patient got the UTI in the hospital and that it was caused by urinary catheter use. And that level of detail usually isn't there, she added.

One reason may be that nurses are the ones putting in and taking out catheters, so while they may note the cause of a new UTI, doctor's records - which are used for billing - might not.

Urinary catheters can help patients who have a bladder blockage, or those who are both incontinent and have bedsores. But many times they're used unnecessarily because catheters make it easier to care for a patient who is immobile or incontinent, Meddings said. And using a catheter greatly increases the risk that a patient will develop a UTI.

CMS did not provide a response to the study findings before deadline.

Other conditions that aren't reimbursed by CMS include bedsores acquired in the hospital, injuries due to falls there and some other preventable infections. The current study didn't look to see how well any of those are recorded in Medicare billing data.

Another factor used to judge quality and reimbursement that has caused controversy is the rate of patients re-hospitalized within a month of being discharged.

The new research was funded by the Blue Cross Blue Shield of Michigan Foundation.

Dr. Bernard Rosof from North Shore-LIJ Health System in Huntington, New York, who wrote a commentary published with the study, said using data from electronic health records instead of billing records may help ensure that the full picture of preventable complications is clear to CMS.

He agreed with the researchers that the current policy could end up rewarding the wrong hospitals.

"An unintended consequence of basing this strategy on inaccurate data would be to penalize those hospitals that report accurately in their coding and claims database and thus seem to have higher rates of hospital-acquired events than hospitals with rates that are low because of inaccurate data rather than because of excellent performance," Rosof wrote.

The study is published in the Annals of Internal Medicine.