Objective: Blood culture contamination rates (CRs) in emergency departments (EDs) vary from 1% to 9% in previous studies. High CRs cause unnecessary admissions, antibiotics, and costs. Different measures have been tried to improve CR. This study sought to determine the ED CR at Arkansas Children's Hospital (ACH) and evaluate an educational intervention.
Design: A slide presentation of the ACH's blood culture collection procedure was made. The ED nursing staff viewed this individually and demonstrated the procedure on a mock extremity. Data before (from June 1 to November 30, 2007; PRE) and after (from February 1 to July 31, 2008; POST) intervention regarding blood cultures positive for organism submitted to the ED and any descriptive data were collected retrospectively. Descriptive statistics described the CR, positivity rate (PSR), and pathogen rate (PR). Frequent collectors were identified as those submitting more than 72 blood cultures during PRE and during POST. Contamination rate in frequent collectors was independently assessed for improvement. This study was approved by the institutional review board of the University of Arkansas for Medical Sciences.
Results: Positivity rate during PRE was 6.8% (n = 15; 95% confidence interval [CI], 2.9-10.7); 114 were contaminants (CR, 5.0%; 95% CI, 1.04-8.98) and 40 were pathogens (PR, 1.8%; 95% CI, ±4.28). Positivity rate during POST was 6.3% (n = 157; 95% CI, 2.6-10.0); 124 were contaminants (CR, 4.9%; 95% CI, 1.14-8.66) and 33 were pathogens (PR, 1.4%; 95% CI, ±4.22). For both PRE and POST, contaminants represented 74% or greater of all the positives and more than 75% of contaminants occurred in patients 0 to 36 months old.
Contamination rates of individual practitioners ranged from 0% to 17% (PRE) and from 0% to 21% (POST). The CR was lower in POST for samples submitted from frequent collectors (practitioners submitting >72 blood cultures per study period, n = 6), was evaluated separately, and found to improve in POST (from 4.1% to 2.7%). A t test for matched samples for these samplers revealed a significant decrease in the POST period, P = 0.03.
Conclusions: The CR in the ACH ED for peripherally drawn blood cultures is approximately 5% monthly with less than 2% PR. For each culture positive for organism, the chance of contamination is 75% or higher. An intervention emphasizing the appropriate technique, monitored training, and improved CR of a group of frequent collectors is warranted.