162. Walking the Fine Line between Causation and Association – New Clinico-Epidemiologic (CE) Criteria for Categorizing Laboratory-Confirmed Bloodstream Infections (LCBSI) among Patients with Central Lines
Session: Oral Abstract Session: Infectious Diseases Practice Challenges
Friday, October 21, 2011: 11:00 AM
Room: 151AB

Background: NHSN defines LCBSI as primary or secondary using criteria that lack specificity. As a result, many BSIs are inappropriately categorized as central line-associated BSIs (CLABSI) because such BSIs did not meet strict requirements for classification as secondary BSI.  Criticisms of the current NHSN definition include: low specificity for CLABSI, poor inter-rater reliability, and inappropriate inference of causal association when cause is clinically uncertain. We aimed to create and validate new CE criteria for classification of LCBSI.


We created new CE criteria by incorporating opinions of experts and feedback from focus groups with infection preventionists and hospital epidemiologists.


The CE criteria were derived and validated in a cohort of patients at Duke University Hospital. Two teams of blinded reviewers used a draft of the CE criteria to categorize BSIs from a “derivation cohort” (DC) consisting of 24 NHSN-defined CLABSIs detected between 6/2010-11/2010. The DC was used to determine content- and construct-validity and reliability of the CE criteria. We then integrated comments from reviewers to create a final version of the CE criteria, which were tested with a “validation cohort” of 18 patients with NHSN-defined CLABSIs detected between 12/2010-3/2011. We calculated overall agreement and kappa to determine inter-rater reliability of the final CE criteria.


The final version of our CE criteria is shown in Figure. Of the 18 NHSN-defined CLABSIs in the validation cohort, six (33.3%) were categorized as Empirical CLABSI by both reviewers; 3 (17%) CLABSIs had single enterococcal isolates; 5 (28%) were considered as “probable secondary”; the remaining case (6%) was categorized as “possible primary”. Reviewers disagreed on 3 cases.  Overall inter-rater agreement for BSI categorization using the CE criteria was 83.3% and the Kappa was 0.76.



We created and validated new CE criteria to categorize LCBSIs. The CE criteria are easy to use, include additional, clinical-based categories, and achieved high inter-rater reliability in our study. Our CE criteria will need to be validated in multiple and different community and tertiary medical centers prior to widespread implementation. 

Subject Category: N. Hospital-acquired and surgical infections, infection control, and health outcomes including general public health and health services research

Luke F. Chen, MBBS, MPH, CIC, FRACP, Daniel J. Sexton, MD, FIDSA, Pamela Isaacs, BSN MHA CIC, Charlene Carriker, RN BSN CIC, Becky A. Miller, MD, Lisa M. Cooper, RN BSN, Sheila Vereen, RN BSN CIC, Rebekah W. Moehring, MD, Nancy Strittholt, RN BSN CIC, Russell Staheli, MPH, Judie Bringhurst, RN MSN CIC and Deverick J. Anderson, MD, MPH, Duke University Medical Center, Durham, NC


L. F. Chen, None

D. J. Sexton, None

P. Isaacs, None

C. Carriker, None

B. A. Miller, None

L. M. Cooper, None

S. Vereen, None

R. W. Moehring, None

N. Strittholt, None

R. Staheli, None

J. Bringhurst, None

D. J. Anderson, None

Findings in the abstracts are embargoed until 12:01 a.m. EST Thursday, Oct. 20 with the exception of research findings presented at IDSA press conferences.