1159. Validation of an Electronic Screening Algorithm for Coronary Artery Bypass Graft Surgical Site Infection Surveillance
Session: Poster Abstract Session: Surveillance HAIs: Advancing the Science
Friday, October 19, 2012
Room: SDCC Poster Hall F-H
Background: Infection preventionists perform 100% chart review to identify surgical site infections (SSI) at The Johns Hopkins Hospital.  Electronic surveillance screening algorithms may reduce the number of charts reviewed and the time spent performing SSI surveillance.  In 2009, an electronic algorithm was developed from a test dataset that was designed to be sensitive enough to identify all patients with SSI, but specific enough to significantly reduce the number of post-surgical charts reviewed.  

Methods: Validation of an electronic algorithm to identify SSI following coronary artery bypass graft (CABG) surgery that had a sensitivity and specificity of 100% and 58% in the development dataset. The screening algorithm utilizes electronic infection surveillance software (TheraDoc) to identify positive bone, tissue or wound cultures, and orders for acetylcysteine, piperacillin/tazobactam, or vancomycin >72 hours after surgery.  Patients undergoing CABG January 2010 through December 2011 were included in the validation. Patients with SSI identified through the electronic screening method and manual chart review were compared.  

Results: 886 patients underwent CABG in the study period and full chart review identified 20 patients who met the National Healthcare Safety Network (NHSN) criteria for SSI.  The electronic surveillance algorithm selected 216 (24.4%) CABG patients for review.  Of these, 14 had SSI.  Six patients with SSI were not selected by the algorithm for review.  Only 1 of these 6 had a culture collected; most met the NHSN criteria based upon symptoms which were identified from follow-up appointment notes.  Compared to chart review, the electronic screening algorithm had a sensitivity of 70%, specificity of 77%, positive predictive value of 6.5%, and negative predictive value of 99.1%. 

 

SSI by Chart Review

No SSI by Chart Review

Positive SSI by TheraDoc Alert

14

202

No SSI by TheraDoc Alert

6

664

Conclusion: This electronic screening algorithm significantly reduced the time and number of chart reviews for SSI surveillance; but it was not sufficiently sensitive for SSI detection.  Careful development and validation of these methods are critical to evaluate their reliability in different populations and over time.

Melanie A. Gavin, M(ASCP), CIC1, Jacqueline A. Galluzzo, RN, BSN, CIC1, Polly Trexler, MS, CIC2, Aaron M. Milstone, MD, MHS1, Sara Cosgrove, MD, MS, FIDSA, FSHEA1 and Lisa L. Maragakis, MD, MPH1, (1)The Johns Hopkins Medical Institutions, Baltimore, MD, (2)Johns Hopkins Medical Institutions, Baltimore, MD

Disclosures:

M. A. Gavin, None

J. A. Galluzzo, None

P. Trexler, None

A. M. Milstone, None

S. Cosgrove, None

L. L. Maragakis, None

Findings in the abstracts are embargoed until 12:01 a.m. PST, Oct. 17th with the exception of research findings presented at the IDWeek press conferences.