To evaluate the ability of a Pediatric Early Warning Score to predict deterioration in different subspecialty patient populations.Design:
Single center, retrospective cohort study. Patients were classified into five groups: 1) cardiac; 2) hematology/oncology/bone marrow transplant; 3) surgical; 4) neurologic; and 5) general medical. The relationship between the Pediatric Early Warning Score and unplanned ICU transfer requiring initiation of specific ICU therapies (intubation, high-flow nasal cannula, noninvasive ventilation, inotropes, or aggressive fluid hydration within 12 hr of transfer) was evaluated.Setting:
Tertiary care, free-standing, academic children’s hospital.Patients:
All hospitalized acute care patients admitted over a 6-month time period (September 2012 to March 2013).Interventions:
None.Measurements and Main Results:
During the study time period, 33,800 patient-days and 136 deteriorations were evaluated. Area under the curve ranged from 0.88 (surgical) to 0.94 (cardiac). Sensitivities for a Pediatric Early Warning Score greater than or equal to 3 ranged from 75% (surgical) to 94% (cardiology) and number needed to evaluate to find one deterioration was 11.5 (neurologic) to 43 patients (surgical). Sensitivities for a Pediatric Early Warning Score greater than or equal to 4 ranged from 54% (general medical) to 79% (hematology/oncology/bone marrow transplant) and number needed to evaluate of 5.5 (neurologic) to 12 patients (general medical). Sensitivities for a Pediatric Early Warning Score of greater than or equal to 5 ranged from 25% (surgical) to 58% (hematology/oncology/bone marrow transplant) and number needed to evaluate of 3.5 (cardiac, hematology/oncology/bone marrow transplant, neurologic) to eight patients (surgical).Conclusions:
An elevated Pediatric Early Warning Score is associated with ICU transfer and receipt of ICU-specific interventions in patients across different pediatric subspecialty patient populations.