Important news from the CDC Tuesday: A particularly deadly kind of hospital-acquired infection, CLABSIs — central line-associated bloodstream infections — was sharply suppressed across the 2000s. There were an estimated 43,000 in intensive-care unit patients in 2001 and an estimated 18,000 in 2009, a drop of 58 percent. That’s excellent: Up to one out of four patients who contracts a CLABSI dies as a result of it. According to the CDC report, the numbers represent — just in 2009 — 6,000 lives saved and and $414 million in healthcare costs that didn’t have to be spent. Across the decade, the lives saved might go as high as 27,000, and the total savings, $1.8 billion.
But the news is not uniformly good. CLABSIs are still prevalent outside of ICUs, in general in-patient care, and especially in outpatient care such as dialysis. In 2009, there may have been 23,000 CLABSIs in hospitals but outside of ICUs, and 37,000 in dialysis-clinic patients.
Why such a discrepancy? Probably because ICUs — where you find the sickest patients undergoing the most interventions — have been the focus of the most intensive hospital-infection prevention programs. Think for instance of the Michigan Keystone Project, created by Hopkins anesthesiologist and “checklist doc” Peter Pronovost, MD, which deployed a specific, always identical, always repeated set of preventive measures across 103 ICUS in Michigan and forced CLABSI rates down to almost zero. Or think of the Pittsburgh Regional Health Initiative, a medical and community consortium that has rolled out quality improvement methods across a swath of Pennsylvania. (Coincidentally, PRHI published an editorial in the Pittsburgh Post-Gazette on Sunday, titled “Hospitals are still killing patients with needless mistakes — and it’s inexcusable because we know how to prevent them.”)
Plus, hospital infections have been the target both of justifiable patient outrage and of the mandatory-reporting laws that outrage produced; 22 states now require hospitals to report CLABSIs to the CDC as a result of state laws passed since 2000.
So prevention works — where it is applied. The bad news within those good-news numbers is that the other healthcare settings are going to pose a significant challenge to prevention efforts. ICU patients may be subject to lots of interventions that put them at risk of infection, but they’re also cared for by very experienced nurses and mid-level professionals who are extensively trained in best practices. That isn’t always true elsewhere in hospitals, and it is certainly not true of many of the United States’ 5,000 dialysis clinics. (For an excellent recap of the uneven and hazardous state of dialysis care in the United States, read ProPublica’s recent investigation “Diagnosing Dialysis.”)
And, to be fair, dialysis patients may be at unusual risk of infection. They’re not immobilized in an ICU bed; they are walking around in the everyday world, carrying some version of a port to their circulatory system — a central line or arteriovenous fistula or graft — that has the potential to introduce pathogens directly to their bloodstreams.
As extraordinary as the success in reducing ICU CLABSIs has been, it may still turn out to have been the soft target. Reducing healthcare-associated infections in settings where staff are less trained and less motivated, and patients’ experiences are less predictable or monitorable, might be the truly difficult challenge.
(The best bloggers on hospital infections are the three physician/infection-control professionals at Controversies in Hospital Infection Prevention. If they take up this report, their analysis will be worth reading.)
Update: I missed that, over the weekend, the journal Pediatrics posted a paper about using checklists and “bundles” of actions to reduce CLABSIs among tiny, fragile babies in neonatal intensive care units. Overall, CLABSIs went down by 67 percent, though the success wasn’t uniform across all hospitals. Schulman J et al., Statewide NICU Central-Line-Associated Bloodstream Infection Rates Decline After Bundles and Checklists. PEDIATRICS Vol. 127 No. 3 March 2011, pp. 436-444 (doi:10.1542/peds.2010-2873)