Changes in the profile of paediatric intensive care associated with centralisation

Objectives: To compare intensive care admissions from a defined population of children in 1991 and 1999, during a period of organisational change and centralisation of paediatric intensive care. Design: Two 12-month population-based audits were compared. Data were colle cted from hospitals in Birmingham and the surrounding districts. Denominato r data were obtained from the Office for National Statistics. The place and rate of intensive care admission, the use of mechanical ventilation at adm ission, mortality and length of stay were compared. Setting: Hospitals in the West Midlands. Participants: All children (< 15 yrs) living in Birmingham who received int ensive care during the study periods. Measurements and results: The number of Birmingham resident children admitt ed for intensive care increased from 277 to 510 (p < 0.0001) i.e. from 1.3 to 2.3 admissions per 1,000 children per year. The proportion of admissions to the principal paediatric intensive care unit increased from 60 % to 90 % (p < 0.0001) in association with its expansion from 6 to 18 beds. Length of ICU stay decreased from 103 to 74 h (difference 219 h, 95 % CI, 4.78-54. 2 h, p = 0.0117). Child mortality fell over this period by 34 deaths per 10 0,000 children (95 % CI 16-51, p < 0.000 1). The proportion of children req uiring mechanical ventilation at admission to intensive care was unchanged. Conclusions: Centralisation by expansion of the lead centre was associated with a large increase in the numbers of children receiving intensive care c onsistent with an unmet need for paediatric intensive care in 1991, which m ay still exist. Centralisation of paediatric intensive care may have contri buted to the fall in child mortality over this time period.