• quality measurement
• infection control
• healthcare quality improvement

W Edwards Deming is famously quoted as having said, “If you can’t measure it, you can’t manage it”. In truth, Deming’s full quotation reads, “It is wrong to suppose that if you can’t measure it, you can’t manage it—a costly myth”.1 2 In our journey to improve our hospital’s rates of Clostridium difficile, we learned first-hand the truth of Deming’s full statement—that in fact, even without the ability to measure perfectly, imperfect measures can still help us improve quality.

US hospitals are currently required to report hospital-acquired C. difficile rates as a condition of participation in several Centers for Medicare and Medicaid Services (CMS) pay-for-performance programmes. CMS is seeking to shed light on this type of preventable patient harm and raising the stakes by putting financial penalties and a hospital’s public reputation at risk. However, there is a vigorous debate in the medical community over the C. difficile measurement technique used by CMS. CMS’s C. difficile rates are based on the National Healthcare Safety Network’s C. difficile ‘LabID’ measure. This metric identifies C. difficile infections solely by positive laboratory test results, without regard to patients’ clinical signs or symptoms. This approach makes case assignments very objective (and therefore reduces ‘gaming’), but comes at the risk of misclassifying colonised patients as infected. The C. difficile tests we have only tell us whether C. difficile is present or not. They do not differentiate between colonisation and infection. This is particularly true of PCR-based testing that looks for the presence of the C. difficile toxin gene but does not determine whether the gene is producing toxin or not. The most specific test we have is the C. difficile toxin immunoassay, but even this is imperfect.

The lack of specificity is the Achilles heel of the C. difficile LabID measure because …