Methods: A retrospective matched comparison group study was performed between Oct. 2012 – Nov. 2015. Data was extracted from unique cases of severe sepsis, MSDRGs (870, 871, 872). Our historical blood culture system was the comparator (cohort A), while the BD BACTEC Plus Aerobic and Lytic Anaerobic media (Becton, Dickinson and Company, Sparks, MD) was designated as the intervention (cohort B). The downstream impact was determined for the following variables: a) TTD, 2) 30 day all-cause raw mortality, substratified by MSDRG, 3) all-cause mortality ratios (observed to expected, O:E, Premier, Inc.), and 4) inpatient length of stay ratios (O:E, Premier, Inc.). Data analysis was performed using JMP, 12.0.1 (Cary, NC); data was analyzed by Wilcoxian Rank Sum or Analysis of Means, alpha = 0.05
Results: Unique positive blood cultures were identified (n=216). Cohort demographics were well matched; no statistical differences were observed for age, gender, compliance with sepsis bundles, phlebotomy practices, MSDRG distribution, microbial diversity, or sample size. The overall mean TTD was 21.0 ± 15.2 hr for cohort A and 16.8 ± 13.2 hrs for B. Crude all-cause mortality for ICU patients with MSDRG 870 (Septicemia or Severe Sepsis with Mechanical Ventilation for > 96 hrs) was statistically decreased in cohort B (p = 0.04). Among sepsis survivors, statistical improvement in TTD for positive cultures in cohort B was observed (p = 0.003). Finally, the O:E mortality and mean inpatient LOS O:E ratios were significantly improved in cohort B, (p=0.02) and (p=0.002), respectively.
Conclusion: Decreases in TTD for bloodstream pathogens were associated with significant improvements in crude all-cause mortality for ICU subjects with MSDRG 870 and with improved O:E ratios for mortality and LOS.
Prompt reporting of blood culture Gram stains has been associated with decreased patient mortality, yet original findings have not yet been widely replicated. Our aim was to document the positive impact for patients with severe sepsis after improvements in the time to detection (TTD) of positive blood cultures.
J. Riley, None
D. Hernandez, None
D. Wolk, Becton Dickinson: Grant Investigator , Research grant
bioMerieux: Grant Investigator , Research grant