(See the article by Tedja et al⁷ on pages 1330–1334.)

Increasingly, payers, legislators, regulators, consumer groups, and the general public have become interested in the prevention of hospital-acquired infections (HAIs). Most states have mandated public reporting of HAIs, and some financial reimbursements are now linked to how effectively hospitals prevent infections.

For example, HAI prevention is a central part of the Centers for Medicare and Medicaid (CMS) Hospital-Acquired Condition (HAC) Reduction Program. In this program, 65% of a hospital’s total HAC score is based on the observed-to-predicted number of central line-associated bloodstream infections (CLABSIs) and catheter-associated urinary tract infections (CAUTIs). When a hospital falls in the highest quartile of HAC scores, a stiff financial penalty is imposed: CMS reduces the hospital’s reimbursements by 1%.

Evaluating hospital quality, however, is fraught with difficulties. A good HAI metric should be objective, reflect a clear episode of patient harm, and be preventable through the implementation of evidence-based practices. The question at hand is whether these standards are met by the National Healthcare Safety Network (NHSN) surveillance definition for CAUTI as part of the HAC Reduction Program.

Whether a NHSN-defined CAUTI represents an episode of patient harm is debatable. Because the surveillance definition does not coincide with physician practices, each CAUTI does not necessarily represent a preventable infection or antibiotic prescription.Reference Al-Qas Hanna, Sambirska, Iyer, Szpunar and Fakih ¹ Furthermore, because the criteria used to define a CAUTI are nonspecific, it is unclear what exactly the metric is measuring. To meet the current CAUTI definition, a patient must meet 3 simple criteria: (1) the presence of a Foley catheter for >2 days; (2) a urine culture growing ≥10⁵ colony-forming units per milliliter (CFU/mL) of bacteria; and (3) signs or symptoms suggestive of a urinary tract infection (UTI). This final criterion has been met by a temperature >38°C in 79.7% of cases reported to the NHSN.Reference Gould ²

The nonspecificity of the NHSN criteria is highlighted in the intensive care unit (ICU), where the simultaneous occurrence of these 3 findings is common. Foley utilization in the ICU ranges from 48% to 76% depending on the type of unit, while fever, including those secondary to noninfectious causes, occurs in 26%–70% of patients.Reference Dudeck, Horan and Peterson ^{3 ,} Reference Rehman and Deboisblanc ⁴ In addition, many patients have a positive urine culture even before they reach the hospital. Asymptomatic bacteriuria is reported in up to 19% of community-dwelling elderly, 27% of diabetics, and 50% of elderly within long-term care.Reference Nicolle, Bradley and Colgan ⁵ The risk of acquiring bacteriuria increases by 3%–7% for every day a Foley remains in place, but the constellation of these 3 findings—Foley, fever, and bacteriuria—does not necessarily indicate an infection of the urinary tract.

As a clinician, a CAUTI is often a diagnosis of exclusion. The Infectious Diseases Society of America CAUTI guidelines emphasize the importance of ruling out an alternate cause of a catheterized patient’s fever before attributing the fever to bacteriuria.Reference Hooton, Bradley and Cardenas ⁶ In contrast, the NHSN definition assigns all fevers to the urinary tract as long as the other CAUTI criteria are met.

Although surveillance definitions will never entirely coincide with clinical infections, the wide gap between the 2 classifications for CAUTI warrants investigation. To quantify this gap, Tedja et alReference Tedja, Wentink, O’Horo, Thompson and Sampathkumar ⁷ reviewed all NHSN-defined CAUTIs that occurred in their medical center’s ICU during 2012–2013. In 97% of patients with a NHSN-CAUTI, the urine culture had been ordered because the patient had a fever. Upon further chart review, the investigators noted that 68% of these febrile cases had a potential alternate explanation for the fever. In fact, 68% may be an underestimate because a standardized, exhaustive search for the cause of fever was not routinely performed. Thus, the investigators were unable to state the definitive cause of each case of fever(s). Nevertheless, since experts agree that catheter-associated bacteriuria and candiduria are rarely the cause of fever in ICU patients,Reference O’Grady, Barie and Bartlett ⁸ it is probable that many, if not most, of these febrile “CAUTI” patients did not have a true urinary tract infection.

As the authors acknowledge, their findings may be limited now that the NHSN-CAUTI definition has again been revised. Based on these 2015 revisions, urine cultures with <10⁵ CFU/mL will no longer meet CAUTI criteria. It is unclear how many cases in the Tedja et al study qualified based on these lower colony counts. In addition, the 2015 changes state that patients with only candiduria—50% of the Tedja et al cohort—no longer qualify as a CAUTI. Furthermore, the 2015 definition widens the infection window period, which means that a patient who had a fever as many as 3 days before or after a positive urine culture may now qualify as a CAUTI. While eliminating candiduria helped improve the specificity of the surveillance definition, the latter change will have the opposite effect.

Ironically, while hospitals like the one described by Tedja et al face intense pressure to prevent CAUTIs, CAUTIs themselves have the lowest financial and clinical impact of all HAIs. There is no high-quality data showing that CAUTIs prolong a patient’s length-of-stay, and a meta-analysis found that CAUTIs accounted for only 0.3% of the total annual costs attributable to the 5 major HAIs. In comparison, surgical site infections account for 34% of all costs, ventilator-associated pneumonias account for 32%, CLABSIs account for 19%, and Clostridium difficile accounts for 15%.Reference Zimlichman, Henderson and Tamir ⁹

A heightened focus on CAUTIs is certainly promoting more judicious use of Foley catheters, but paradoxically, this has not consistently led to lower CAUTI rates. Reducing unnecessary Foley use can actually increase a hospital’s CAUTI rate by decreasing the rate’s denominator: catheter days.Reference Wright, Kharasch, Beaumont, Peterson and Robicsek ¹⁰ As a result, implementing the primary strategy for CAUTI prevention might fail to improve the “quality” metric by which hospitals are graded.

Hospitals with high CAUTI rates are looking beyond routine prevention measures to meet the metric. Not surprisingly, the frequency of obtaining urine cultures affects a hospital’s CAUTI rate.Reference Al-Qas Hanna, Sambirska, Iyer, Szpunar and Fakih ¹ Some hospitals are therefore encouraging physicians to reduce their use of urine cultures when there is a low suspicion of UTI. Unfortunately, defining “low suspicion” is challenging, and it is easy to imagine situations in which the urinary tract truly is the source of infection but, due to institutional pressure, a urine culture would not be sent before empiric antibiotics are started. A reflex urinalysis has been proposed to prevent culturing of the urine when there is no pyuria,Reference Stovall, Haenal and Jenkins ¹¹ but this approach is also problematic. While the absence of pyuria helps rule out a CAUTI, the finding of pyuria is nonspecific for infection.Reference Hooton, Bradley and Cardenas ⁶

In the end, the current CAUTI metric is forcing hospitals to invest limited resources in surveillance and prevention efforts without substantial benefit to patient care. The time and effort spent on CAUTI, as currently defined, carry significant opportunity costs that might be better spent targeting other HAIs or quality-improvement metrics.

Given the limitations of the current NHSN-CAUTI metric, a few alternate metrics deserve consideration. One would be Foley utilization—a simple process measure of a how frequently a hospital uses this potentially harmful medical device. Alternatively, hospitals could be evaluated on the incidence of bacteremia secondary to CAUTIs—a more direct indication of patient harm. According to NHSN data from 2009–2013, 5.2–6.1% of reported CAUTIs were complicated by a secondary bloodstream infection.Reference Gould ² Similarly, Tedja et al found that 6% of all CAUTI cases grew the same organism in both blood and urine cultures. A third consideration would be to measure the incidence of NHSN-defined CAUTIs in which the patient received antibiotics targeting the urine culture results.Reference Flanders and Saint ¹² Such a revised metric would have the added benefit of discouraging treatment of asymptomatic bacteriuria, a common reason for antibiotic misuse.

The underlying goal of all quality metrics is to promote better patient care. Since urinary catheters can cause harm, we need objective and clinically relevant methods for measuring this harm. To that end, revising the CAUTI metric could turn an opportunity cost into a greater opportunity to improve patient care.