The Pediatric Index of Mortality as a Trigger Tool for the Detection of Serious Errors and Adverse Events.

Objectives: To test the hypothesis that patients who die in a PICU despite a low predicted mortality at PICU admission are affected by serious errors and adverse events. Design: Retrospective cross-sectional review of medical records for serious errors and adverse events. Setting: Tertiary interdisciplinary neonatal PICU. Patients: All admissions to our PICU who died despite a low expected mortality (Pediatric Index of Mortality) of less than 10% (trigger-positive admissions). They were compared with a random sample of 100 PICU admissions with a Pediatric Index of Mortality of less than 10% who survived (trigger-negative admissions). Interventions: None. Measurements and Main Results: There were 7,383 admissions (91%) with a Pediatric Index of Mortality 2 below 10%. Seventy-two trigger-positive admissions and 100 trigger-negative admissions met the criteria for detailed chart review. Forty-five serious errors and adverse events were identified, 0.47 per trigger-positive admission and 0.11 per trigger-negative admission (p < 0.001). Nineteen serious errors and adverse events (42%) were related to clinical sepsis acquired during the PICU stay, 17 (89%) in trigger-positive admissions and two (11%) in trigger-negative admissions (p < 0.001). A further 18 serious errors and adverse events (40%) were intervention related, nine (50%) in trigger-positive admissions and nine (50%) in trigger-negative admissions (p = 0.46). Eight serious errors and adverse events (18%) were associated with medication use, all of which occurred in trigger-positive admissions (p = 0.001). The median (interquartile range) age for admissions with and without serious errors and adverse events was 0.3 months (0.0-4.6 mo) and 7.4 months (0.4-58.4 mo) (p < 0.001), and their median (interquartile range) duration of invasive ventilation was 140 hours (50-451hr) and 2 hours (0-41hr) (p < 0.001), respectively. Conclusions: The records of PICU patients with a low expected mortality at admission and death in PICU should be reviewed routinely and/or discussed at morbidity and mortality meetings. These patients may have experienced more in-hospital safety-related events compared with PICU patients with a low Pediatric Index of Mortality who survived. Such adverse events may be amenable to system changes, thus improving patient care.