56 A Quality Improvement Program to Reduce Blood Culture Contamination in the Emergency Department

Saturday, April 2, 2011: 12:00 PM
Cortez Ballroom (Hilton Anatole)
Thomas R. Talbot, MD, MPH , Vanderbilt University School of Medicine, Nashville, TN
Jacki Ashburn, RN , Vanderbilt University Medical Center, Nashville, TN
Alan B. Storrow, MD , Vanderbilt University School of Medicine, Nashville, TN
Theodore Speroff, PhD , Vanderbilt University School of Medicine, Nashville, TN
Robert S. Dittus, MD, MPH , Vanderbilt University School of Medicine, Nashville, TN
Wesley H. Self, MD , Vanderbilt University School of Medicine, Nashville, TN

Background: Blood culture contamination is a common, costly and preventable problem that may lead to unnecessary hospitalization and unneeded treatment.  The Clinical and Laboratory Standards Institute benchmark for the maximum acceptable percentage of contaminated blood cultures is 3%.  During 2009, 7.18% of the 8,254 blood cultures collected in the Vanderbilt University Hospital (VUH) Emergency Department (ED) grew “potential contaminants,” defined as bacteria that commonly inhabit human skin and when grown in culture represent contamination > 50% of the time.

Objective: To decrease the percentage of potentially contaminated blood cultures collected in the VUH ED to less than 3%.  

Methods: We formed a quality improvement team consisting of ED nurses, physicians and administrators, the chief hospital epidemiologist, microbiology laboratory personnel and quality improvement leaders. Using surveys, interviews, chart reviews and direct observation, we investigated the causes of elevated blood culture contamination in the VUH ED. A quality improvement (QI) program to transform blood culture collection into a fully sterile procedure was then developed. This consisted of: (1) a new materials kit that contained all the equipment necessary to collect a culture using sterile technique, including a chlorhexidine skin prep device and sterile supplies; (2) a checklist within each kit outlining how to perform a sterile blood draw; (3) an internet-based training video; (4) weekly feedback of contamination rates. The primary outcome was the monthly percentage of potentially contaminated blood cultures, which was plotted on a Statistical Process Control (SPC) p chart to assess for changes over time. We implemented the QI program on January 31, 2010.  The pre-intervention baseline period included March 1, 2009 – January 31, 2010; the post-intervention period included February 1, 2010 – October 23, 2010.   

Results: During the pre-intervention period, 491/7,389 (6.65%, 95% CI 6.08 – 7.21%) blood cultures grew potential contaminants; during the post-intervention period, 133/4,964 (2.68%, 95% CI 2.23 – 3.13%) blood cultures grew potential contaminants (p < 0.001). On an SPC p chart of monthly percentages of cultures growing potential contaminants, the pre-intervention period was in statistical control (Fig).  During the post-intervention period, nine consecutive months had percentages less than the lower control limit defined by the pre-intervention period. 

Conclusions: Implementation of this program led to a statistically significant decrease in blood culture contamination in the VUH ED.  Collecting blood cultures using full sterile technique and routine feedback of contamination rates to frontline staff is feasible in the emergency department setting and effective at reducing contamination.