Medical Errors: The problem with getting to zero

by Dr. Anish Koka

Published Dec. 30, 2015, 12:37 p.m. ET

In a time not so long ago, medical error was a matter to be discussed only among physicians in hospital conference rooms. Boy, has a lot changed. The first real chink in the armor came with the now-famous Institute of Medicine report in 1999, that up to 98,000 people may die every year of medical errors. An avalanche of studies, some based on a mixture of faith and evidence, followed, pushing this number ever higher and firmly ensconcing medical errors and patient harm in the public consciousness.
Many fronts have been opened in the war on patient harm, but a particular focus has been on infections related to central lines — catheters in large, close-to-the-heart veins used in critically ill patients.
Central line-associated bloodstream infections, or CLABSIs, are ground zero in the war on patient harm. The physician widely regarded as the conqueror of these infections is Peter Pronovost. He published a landmark study that used a variety of evidence-based "intervention bundles" including education, culture change, and the use of checklists to dramatically reduce infection rates. Some form of his bundle soon was in hospitals everywhere. Dramatic reductions in CLABSI rates followed.
A hard lesson I learned after a peanut-butter-and-jelly-sandwich orgy applies here: There can be too much of a good thing. Pronovost started on a mission to reduce infections, but now the mission is to get to zero. CLABSIs are deemed "preventable harm," suggesting that any such infection can be prevented. Medicare has pushed forward with policies to withhold reimbursement for any CLABSIs, because they are all deemed preventable.
But is this realistic? Even Pronovost's study reduced infection rates 66 percent. A 2012 study provocatively titled "Zero risk for central line-associated bloodstream infection: Are we there yet?" found that we were not. The study, based in 37 ICUs, sought to identify the longest time a central line could stay free of infection with another "insertion intervention bundle." The authors found no intervention that was associated with zero risk. In today's non-Star Trek world, central-line catheters are sometimes necessary to save a life. These catheters are used, not in the vacuum of space into sterile objects, but within the milieu of a hospital teeming with pathogenic organisms, and into patients who are far from sterile.
Tying reimbursement to an impossible goal creates pressures that have consequences that are unintended but predictable. A study comparing hospital infection-control practitioners to a standardized computer algorithm demonstrated significant discordance. The medical center that had the lowest rate of central-line infections as judged by infection-control practitioners (2.4 per 1,000 central-line days) had the highest rate by the standardized computer algorithm (12.6 per 1,000 central-line days). The implication: Infection-control practitioners have an incentive to subjectively apply CLABSI definitions to get a desirable result.
We must respect the power of words. Using the terms medical error and patient harm too liberally and loosely strips away their meaning. How do we continue to make progress if we don't know what the real numbers are? Remarkably, yet predictably, the current iteration of transparency combined with unrealistic expectations has the potential for more harm than good.

Anish Koka, M.D., a cardiologist in private practice in Philadelphia, blogs on health-care policy and delivery at anishkoka.blogspot.com. For a longer version of this post, see The Health Care Blog.