Scheer, S. et al. Critical Care Medicine. Published online: September 22 2016
Objective: To investigate the impact of a quality improvement initiative for severe sepsis and septic shock focused on the resuscitation bundle on 90-day mortality. Furthermore, effects on compliance rates for antiinfective therapy within the recommended 1-hour interval are evaluated.
Design: Prospective observational before-after cohort study.
Setting: Tertiary university hospital in Germany.
Patients: All adult medical and surgical ICU patients with severe sepsis and septic shock.
Intervention: Implementation of a quality improvement program over 7.5 years.
Measurements: The primary endpoint was 90-day mortality. Secondary endpoints included ICU and hospital mortality rates and length of stay, time to broad-spectrum antiinfective therapy, and compliance with resuscitation bundle elements.
Main Results: A total of 14,115 patients were screened. The incidence of severe sepsis and septic shock was 9.7%. Ninety-day mortality decreased from 64.2% to 45.0% (p < 0.001). Hospital length of stay decreased from 44 to 36 days (p < 0.05). Compliance with resuscitation bundle elements was significantly improved. Antibiotic therapy within the first hour after sepsis onset increased from 48.5% to 74.3% (p < 0.001). Multivariate analysis revealed blood cultures before antibiotic therapy (hazard ratio, 0.60-0.84; p < 0.001), adequate calculated antibiotic therapy (hazard ratio, 0.53-0.75; p < 0.001), 1-2 L crystalloids within the first 6 hours (hazard ratio 0.67-0.97; p = 0.025), and greater than or equal to 6 L during the first 24 hours (hazard ratio, 0.64-0.95; p = 0.012) as predictors for improved survival.
Conclusions: The continuous quality improvement initiative focused on the resuscitation bundle was associated with increased compliance and a persistent reduction in 90-day mortality over a 7.5-year period. Based on the observational study design, a causal relationship cannot be proven, and respective limitations need to be considered.
Read the abstract here