A Quality Improvement Collaborative Improves Care for Pediatric Septic Shock*

BACKGROUND: Sepsis is a leading cause of morbidity in children worldwide. Barriers exist for timely recognition and management. We sought to improve timely management for patients with sepsis and improve mortality through a one-year quality improvement (QI) collaborative. METHODS: 15 hospitals initially enrolled through this Children’s Hospital Association (CHA) collaborative from June 2012-2013. Children from the ED, ICU and floor with sepsis, severe sepsis and septic shock were included. Definitions of sepsis were institutionally based but were encouraged to model the International Sepsis Consensus Conference guidelines by Goldstein et al. Time of sepsis onset was institutionally determined and could include ED arrival, first abnormal vital signs (VS), clinician determination, order set or best practice alert (BPA) initiation or an elevated pediatric early warning system (PEWS) score. Process metrics included adherence to first clinical assessment including a full set of VS within 20 minutes, administration of first crystalloid bolus within 15 minutes, and third bolus and antibiotics within 60 minutes. Outcome measures included 3 and 30 day mortality. Multidisciplinary teams attended virtual monthly learning sessions to share intervention strategies that informed Plan Do Study Act (PDSA) cycles, coordinate data collection efforts and overcome barriers to care. Data reports of site specific and aggregated metrics were disseminated monthly as run charts to drive rapid cycle improvement. RESULTS: 10 hospitals submitted results throughout the collaborative timeframe. 1737 pediatric patients with sepsis, severe sepsis and septic shock were treated. 656 patients presented with severe sepsis or septic shock. Quarterly data demonstrated a mean improvement in initial clinical assessment from 46% to 60% (p < 0.001) and in adherence to administration of first fluid bolus from 38 to 46% (p < 0.015). (Figure 1) There was no statistically significant improvement in other process metrics. Patients with severe sepsis or septic shock had a reduction in 30 day mortality from 11% to 3% (p=0.029), with a trend towards improved 3 day mortality (4% to 1%, p=NS). (Figure 2) There was no statistically significant improvement in 3 or 30 day mortality for the total cohort (sepsis plus severe sepsis and septic shock). Reported barriers to improvement efforts included difficulties with defining sepsis, retrospectively defining time of sepsis onset, implementing electronic health record (EHR) strategies and automating data capture. CONCLUSIONS: A QI collaborative focused on improving timely recognition and management of pediatric sepsis and septic shock improved some process metrics for care. There was an observed decrease in mortality for the severe sepsis and septic shock cohort, likely due to multifactorial causes. Future national efforts should standardize sepsis and time zero definitions, data abstraction methods and processes of care for management of patients with sepsis. Collaboratives should focus upon shared platforms for maximizing EHR capabilities across institutions.