Automated phenotypic method versus manual method in the identification of microorganisms isolated from blood cultures: clinical and microbiological outcomes

Abstract

Introduction: Automation is increasingly used in microbiology laboratory, however, few studies assessed clinical outcomes compared to traditional methods. In Brazil, no studies with this objective were detected. Objective: To analyze the clinical and microbiological impacts after implantation of an automated phenotypic method in a microbiology service. Methods: Observational and retrospective study carried out on the microbiology laboratory involving blood culture test from intensive care unit (ICU) patients. Data were collected from hospitalized patients between January 2014 and December 2015. The length of hospitalization, number of empirical therapies, deaths and information related to microbiological isolation were analyzed. The sample was obtained by convenience. Pearson's Chi-square and Student's t-tests were used to compare outcomes. The program used was the Stata release, version 11, being considered significant values of p<0.05. Results: A total of 472 patients were evaluated. There was no reduction in the empirical prescription of antimicrobials (54.7% vs 45.3%; p=0.33), ICU stay (14.5 days vs 15.8 days; p=0.78) and mortality (54.4% vs 45.6%; p=0.36). Similarly, profile of isolated agents in both methods did not appear to be discrepant, however, there was an increase of 44.7% in the number of microbial isolates (76 vs 110) and a better characterization of them. Conclusion: The microbiology laboratory automation did not modify the length of stay, ICU mortality and the number of empirical therapies. However, identification and isolation of microorganisms was improved.