Gram-negative pathogens take 24-72 hours to be identified (ID) and for antimicrobial susceptibilities (AS) to be obtained from blood cultures. Rapid molecular diagnostic tests (RDT) can shorten time to pathogen identification and antibiotic optimization. We compared our current processes with the predicted impact of an RDT system, to assist with an institutional decision to invest in RDT. The Accelerate PhenoTest™ BC Kit, which provides pathogen ID within 90 minutes of positive growth and AST within 7 hours, was selected as an example of a commercially available system for study purposes.
A retrospective review of adult patients between January 2016 and September 2017 with ≥ 1 positive blood culture with a gram-negative organism detectable by the RDT system was conducted. Subject characteristics and organisms identified were recorded. Primary endpoints of the study were potential change in times to ID and AS with use of RDT. Standard laboratory procedures were used for ID and AS (pre-intervention period). The same subject population was used to calculate a theoretical time to ID and AS if RDT were used (modeled post-intervention). Patients were excluded if they expired or were discharged prior to culture results, on hemodialysis, were an outpatient at the time of + culture, or if time of ID or AS was not reported in the electronic record.
156 subjects met inclusion criteria. The most common organisms isolated were E. coli (45%) and K. pneumoniae (22%). The pre-intervention mean time to ID and AS in the medical record were identical at 56 hours (using VITEK®). The mean times to effective (covering) and optimal (targeted/consolidated) therapy for the pre-intervention group were 8 hours and 75 hours, respectively. For the modeled post-intervention period, RDT could decrease time to ID by 54 hours (95% CI: 50.5-59.1, p < .001) and AS by 49 hours (95% CI: 45.1-53.5, p < .001).
Time to optimal therapy of gram-negative bloodstream infections at our facility was ~ 3 days (within a day of final organism ID and AS). This demonstrates excellent communication protocols between our microbiology and clinical departments, suggesting that our modeled benefit to RDT for organism ID and AS has good potential to be translated into clinical benefits.
F. Perez, None
U. Akpoji, None
A. Hirsch, None
C. Burant, None
M. Navas, None
S. Sims, None