Change in Ratio of Observed-to-Expected Deaths in Pediatric Patients after Implementing a Closed Policy in an Adult ICU That Admits Children
To evaluate the matching between workload in a paediatric cardiac intensive care unit (ICU) and the corresponding medical staffing levels over a 24-h period. A review of workload measured by: (a) admissions, (b) severity of illness in admissions using case-mix descriptors and mortality as a proxy, (c) cardiac arrests (CA) and (d) extracorporeal membrane oxygenation (ECMO) cannulations. An evaluation of matching between workload and medical staff schedules. A tertiary paediatric cardiac ICU. 2,799 admissions over a 49-month period. New admissions peaked in the evening, and the ratio of doctors’ hours to admissions was lowest between 1359 and 2000 h. Although only 515 (17.3%) cases were admitted between 2000 and 0759 h, these were more likely to be emergencies, to have higher Paediatric Index of Mortality 2 (PIM2) scores and to die (p < 0.001). There was an increased adjusted risk of death in admissions between 2000 and 0159 h (p = 0.021). There was no difference in the occurrence of either CA (p = 0.41) or ECMO (p = 0.95) between day and night. The ratio of doctors’ hours to CAs and ECMOs was lowest from 2000 to 0800 h. The conventional medical staffing roster generated the greatest concentration of staff in the morning, reducing to the lowest level between 0200 and 0759 h. Workload was most intense for the in-house team at night, in terms of sicker admissions, ECMOs and cardiac arrests. Conventional roster patterns may not offer ideal matching between staffing and workload. Data analysis of variable and urgent workload may be used to inform medical rosters.