Abstract 13004: Continuous Quality Review of ICU CPR Events: Performance, Outcomes and Effect of Systems-Level Changes

Abstract

Introduction: Outcomes of in-hospital cardiac arrest remain disappointing. Continuous quality review (CQR) has been used to improve provider, team and systems performance during CPR. Little data exists on the incidence of individual, team and systems errors during in-hospital CPR and the influence of errors and systems-level changes on outcomes are equally unexplored. Hypothesis: CQR will identify consistent errors in three performance domains (technical, team, systems). Environmental factors and systems changes will affect code quality and outcomes. Methods: Analysis of prospectively collected data describing all ICU CPR events from September 2009-December 2014 in a tertiary care children’s hospital. Monthly quality review by a multidisciplinary team focused on three quality metrics (noise, leadership and equipment failure) and was later expanded to six metrics (communication, technical or systems errors). Results: There were 243 arrests in 193 patients, most of whom had primary cardiac (39%) or pulmonary disease (23%). Most events occurred during the day (7am-7pm); 27% occurred on a weekend or holiday. Median compression time was 5 minutes (IQR 2-15min). Event survival was 80% (70% ROSC, 10% ECMO) and 48% of hospitalizations ended in death. Excess noise (19%), leadership confusion (11%) and equipment failure (6%) were common. Quality CPR (all 3 elements) occurred 74% of the time and was not associated with event survival. Excess noise (p<0.01) and leadership confusion (p=0.02) were more common during the day. A more recent cohort (n=34) demonstrated frequent communication, systems and technical errors (15%, 21% and 21% respectively). Only 38% of arrests had all 6 quality elements. In multivariable analysis, leadership confusion was associated with ICU survival (OR 2.2; 0.66-1). Statistical process control (SPC) only showed special cause variation in single quality metrics despite multiple systems level changes. Code incidence trended down after opening a new hospital and creation of a separate cardiac ICU service (p=0.07). Conclusion: Individual, team and systems errors are common during CPR in critically ill children. CQR incorporating SPC should be used to track performance and monitor the impact of QI initiatives and systems changes.