Background: Central-line–associated blood stream infections (CLABSIs) are linked with significant morbidity and mortality. A NHSN laboratory-confirmed bloodstream infection (LCBSI) has specific criteria to ascribe an infection to the central line or not. The criteria used to associate the pathogen to another site are restrictive. This objective to better classify CLABSIs using enhanced criteria to gain a comprehensive understanding of the error so that appropriate reduction efforts are utilized. Methods: We conducted a retrospective review of medical records with NHSN-identified CLABSI from July 2017 to December 2018 at 2 geographically proximate hospitals. Trained infectious diseases personnel from tertiary-care academic medical centers, the University of Virginia Health System, a 600-bed medical center in Charlottesville, Virginia, and Virginia Commonwealth University Health System with 865 beds in Richmond, Virginia, reviewed charts. We defined “overcaptured” or O-CLABSI into different categories: O-CLABSI-1 is bacteremia attributable to a primary infectious source; O-CLABSI-2 is bacteremia attributable to neutropenia with gastrointestinal translocation not meeting mucosal barrier injury criteria; O-CLABSI-3 is a positive blood culture attributable to a contaminant; and O-CLABSI-4 is a patient injecting line, though not officially documented. Descriptive analyses were performed using the χ2 and the Fisher exact tests. Results: We found a large number of O-CLABSIs on chart review (79 of 192, 41%). Overall, 56 of 192 (29%) LCBSIs were attributable to a primary infectious source not meeting NHSN definition. O-CLABSI proportions between the 2 hospitals were statistically different; hospital A identified 34 of 59 (58%) of their NHSN-identified CLABSIs as O-CLABSIs, and hospital B identified a 45 of 133 (34%) as O-CLABSIs (P = .0020) (Table 1). When comparing O-CLABSI types, hospital B had a higher percentage of O-CLABSI-1 compared to hospital B: 76% versus 64%. Hospital A had a higher proportion of O-CLABSI-2: 21 versus 7%. Hospitals A and B had similar proportion of O-CLABSI-3: 15% versus 18%. These values were all statistically significant (P < .0001). Discussions: The results of these 2 geographically proximate systems indicate that O-CLABSIs are common. Attribution can vary significantly between institutions, likely depending on differences in incidence of true CLABSI, patient populations, protocols, and protocol compliance. These findings have implications for interfacility comparisons of publicly reported data. Most importantly, erroneous attribution can result in missed opportunity to direct patient safety efforts to the root cause of the bacteremia and could lead to inappropriate treatment.

Funding: None

Disclosures: Michelle Doll, Research Grant from Molnlycke Healthcare