Clostridium difficile infection (CDI) is a major cause of antibiotic-associated diarrhea and colitis, and it is among the most common healthcare-associated infections (HAIs). ^{1 ,} Reference Miller, Chen, Sexton and Anderson ² The CDI incidence rate doubled between 2000 and 2009, and recent National Healthcare Safety Network (NHSN) data suggest a modest 8% decrease in CDI between 2011 and 2014, which represents far less progress in prevention than has been made against other serious HAIs such as central-line–associated bloodstream infection (CLABSI) and surgical site infection (SSI). ³ The NHSN measure used to track C. difficile disease is the laboratory-identified (LabID) CDI event, which was introduced in March 2009. ⁴ LabID CDI events represent new positive C. difficile lab test results for Toxin A and/or B and are further divided into community onset (CO; sample collected within the first 3 days of admission) or hospital onset (HO; sample collected on day 4 or later). HO Lab-ID CDI events (reported as standardized infection ratios [SIRs]) are reported publicly and are incorporated into the Centers for Medicare and Medicaid Services (CMS) value-based purchasing program. ^{5 , 6}

At the same time that CDI has become the most common HAI in many US hospitals,Reference Miller, Chen, Sexton and Anderson ² advances in diagnostic technology have improved our ability to detect toxigenic C. difficile in stool samples.Reference Bartsch, Umscheid, Nachamkin, Hamilton and Lee ^{7 –} Reference Polage, Gyorke and Kennedy ⁹ Widely adopted nucleic acid amplification tests (NAATs) are far more sensitive than toxin enzyme immunoassays (EIAs), and they have increased the rate of C. difficile detection by more than 50%.Reference Moehring, Lofgren and Anderson ¹⁰ In addition, because such assays detect the toxin gene rather than the toxin itself, there is increasing recognition that some positive tests indicate the presence of a toxigenic strain in the absence of clinically significant disease.Reference Koo, Van and Zhao ^{8 ,} Reference Polage, Gyorke and Kennedy ⁹ Many centers have begun using testing algorithms that include toxin EIA testing in combination with more sensitive tests (glutamate dehydrogenase [GDH], NAAT, or both) in an effort to reduce costs, to obtain additional information about toxin production (to support clinical management), and to maintain the sensitivity of NAAT testing (to help guide infection control efforts).Reference Bartsch, Umscheid, Nachamkin, Hamilton and Lee ⁷

To improve interfacility comparison and bring fairness to pay-for-performance programs that incorporate LabID-CDI events, the NHSN utilizes a risk adjustment formula that includes several variables, including the diagnostic method used.Reference Dudeck, Weiner and Malpiedi ¹¹ The diagnostic method variable is divided into 3 categories: NAAT, EIA for toxin, and all other tests. Because our hospital utilizes an algorithm that incorporates both toxin EIA and NAAT testing, we can calculate our LabID-CDI event SIR under the assumption that we use only EIA or only NAAT to evaluate the performance of the current NHSN risk-adjustment formula.

RESULTS

From February 1, 2015, to February 29, 2016, UIHC had 213,404 patient days and 13,788 hospital admissions. During this time period, 374 patients had toxigenic C. difficile detected in stool samples by either EIA or NAAT, and 127 of these incidents were detected within the first 3 days of hospital admission (ie, were classified as community onset [CO]). Table 1 presents the differences in number of observed and expected cases of HO LabID-CDI under the EIA and NAAT assumptions, and HO SIR values.

TABLE 1 Hospital-Onset (HO) LabID-CDI Event Standardized Infection Ratio (SIR) when Using Enzyme Immunoassay (EIA) Versus Nucleic Acid Amplification Test (NAAT)

NOTE. No. of predicted LabID events=exp (β₀+β₁X₁+β₂X₂+…)× patient days. No. of predicted (expected) HO CDI LabID events=exp [(−7.8983+0.385 NAAT^a+0.0160 EIA^a)+0.3338×(CO CDI prevalence rate)+0.2164 beds^b+0.0935 beds^c+0.187 major^d+0.0918 graduate^d]×CDI patient days, where ^a indicates CDI test types and ^b indicates bed size >245 beds, ^c indicates bed size=101–245 beds, and ^d indicates medical school affiliation. EIA model: No. of predicted LabID events=exp [−7.8983+0.385×0+0.1606×1+0.3338×0.43+0.2164×1+0.0935×0+0.187×1+0.0918×1]×213,404, where bold values have been substituted into the first equation as follows: NAAT=0; EIA=1; CO CDI prevalence rate=0.43; >245 beds=1; 101–245 beds=0; medical school majors=1; medical school graduates=1; and CDI patient days=213,404. NAAT model (assumes EIA positive samples are all NAAT positive): No. of predicted LabID events=exp [−7.8983+0.385×1+0.1606×0+0.3338×0.92+0.2164×1+0.0935×0+0.187×1+0.0918×1]×213,404, where bold values have been substituted into the first equation as follows: NAAT=1; EIA=0; CO CDI prevalence rate=0.92; >245 beds=1; 101–245 beds=0; medical school majors=1; medical school graduates=1; and CDI patient days=213,404.

^a The CDI LabID SIR is calculated by dividing the no. of observed HO CDI LabID events by the no. of expected events.

The SIR for HO-LabID-CDI was almost double for NAAT (0.95) compared with EIA alone (0.50).