Institutions Measure Bloodstream Infection Rates Differently

A new study looking at how hospitals identify pediatric patients who develop catheter-associated bloodstream infections (CA-BSI) found significant inconsistencies in the methods used to report the number of patients who develop them.

The study, led by Matthew Niedner, MD, assistant professor of pediatrics and communicable diseases at U-M C.S. Mott Children's Hospital, was conducted by the National Association of Children's Hospitals and Related Institutions Pediatric Intensive Care Unit Focus Group and appears in the October issue of the American Journal of Infection Control.

"There is an intense amount of attention being placed on measures of quality performance that have significant implications in pay-for-performance, and reimbursement," says Niedner, who led the study. "What you have is a desire to measure quality but a lack of perfect measures. Measures are often good enough to enable quality improvement, but can leave undesirable ambiguity when used comparatively as a metric of clinical performance."

Bloodstream infections are the most common hospital-associated infections in pediatric intensive care units (PICUs) and a significant source of in-hospital deaths, increased length of stay and added medical costs. Both adult and pediatric patients who have catheters inserted into their blood vessels face increased risk of developing an infection along the invasive plastic devices that can become deadly as it spreads into the bloodstream.

One hundred forty-six respondents from five professions in 16 PICUs completed surveys with a response rate of 40 percent. All 10 infection control departments reported inclusion or exclusion of central line types inconsistent with the Centers for Disease Control and Prevention (CDC)'s CA-BSI definition, half calculated line-days inconsistently, and only half used a strict, written policy for classifying BSIs. Infection control departments report substantial variation in methods, timing, and resources used to screen and adjudicate BSI cases. Greater than 80 percent of centers report having a formal, written policy about obtaining blood cultures, although less than 80% of these address obtaining samples from patients with central venous lines, and any such policies are reportedly followed less than half of the time. In assessing attitudes and beliefs, there was much greater confidence in the validity of CA-BSI as an internal/historical benchmark than as an external/peer benchmark.

All of the surveyed infection control practitioners in the study said they used the Centers for Disease Control and Preventions definition for CA-BSI, but none actually did, says Niedner. This has significant implications in the era of mandatory public reporting, pay-for-performance and Medicares never events." The Centers for Medicare & Medicaid Services lists CA-BSI as a never event, and no longer reimburses for such hospital-acquired infections.

The study also showed that more aggressive surveillance efforts correlate with higher catheter-associated bloodstream infections rates. This suggests "that the harder one looks for CA-BSIs, the more likely they are to find them."

"From an internal perspective, you want an aggressive surveillance system that is inclusive of all possible cases, but from a public reporting or pay-for-performance standpoint, youd like to exclude as many cases as you can," Niedner says. "There are no definitive national standards as to how to go about doing CA-BSI surveillance at the clinical practice level. It leaves wiggle room that pits hospital economics and reputation against quality improvement teams."

"If you are interested in improving quality of care, you look hard, if youre interested in reputation and reimbursement, maybe you dont look so hard," Niedner adds.

The studys findings offer a compelling opportunity for hospitals to improve their CA-BSI surveillance as a means to promote valid comparison among institutions, Niedner says. Current publicly reported data show that some hospitals report a four-fold difference in CA-BSI rates.

The current system makes it difficult to identify best performers, he adds. "You have to ask yourself, Is it because their care practices are good or is it because their surveillance is weak?"

Niedner hopes this work spurs further research into improving hospital surveillance for such infections. Improved understanding of this variability and awareness of the potential consequences provides an opportunity and rationale to define CA-BSI surveillance best practices and work toward standardizing them across institutions, he adds.

"Many problems become more manageable when we standardize procedures," he says. "Various professional bodies have put forward recommendations for CA-BSI surveillance, but not at the level that will give it real traction. Its going to take a national entity endorsing standardized surveillance practices to improve the validity of institutional comparisons."

Reference: American Journal of Infection Control, Volume 38, Issue 8 (October 2010). doi:10.1016/j.ajic.2010.04.211