Admissions for critically ill children: where and why?

INTRODUCTION: Planning services for critically ill children requires identification of overall critical care activity as well as an assessment of population needs. METHOD AND OBJECTIVES: This prospective needs assessment took a census approach to estimating population-based admission rates for paediatric critical care irrespective of where care was provided. A survey form was completed for every child in the study population for all of their admissions.

CRITERIA: The need for tracheal intubation was used as a proxy for defining need for intensive care in this study. Critical illness was defined by clinical criteria adapted from the Advanced Paediatric Life Support Guidelines.

STUDY POPULATION: All children under 17 years resident in south-east England (Thames regions) who required care for a critical illness in any inpatient setting between 1 December 1996 and 30 November 1997. Critical illness was the presence of acute body-system or multi-system failure.

RESULTS: A wide variation in the rates of critical care admission to different types of care settings was reported ranging from 1.2 admissions per 1000 resident children per year for PIC units (general and cardiothoracic units) to 0.02 admissions per 1000 children per year for children admitted from the community to neonatal units. The age of children and their rates of admission were directly related to the type of ward or unit where children received care. Tracheal intubation occurred in all care settings. The proportion of intubated children transferred to paediatric intensive care units (PICUs) or paediatric cardiothoracic intensive care units (PCICUs) varied according to the type of referring unit. The proportion of episodes involving tracheal intubation where the child was not transferred to a PICU or PCICU was 52% for children in stand-alone neurosurgical units, 41% for those in adult intensive care units (AICUs), and 4% for those in children's wards.

CONCLUSIONS: This baseline study shows a significant number of critically ill children who are never cared for in PIC units. With national changes in UK policy to regionalise care for these children, monitoring care in all locations by cause of admission remains important. While the data were collected in 1997, the findings from this study remain relevant and provide the basis for planning regional critical care services for children. Results are also relevant to other geographical areas in that measuring the use of services for critically ill children must go beyond documenting admission to ICUs for children and adults. All settings for critical care must be identified, the activity documented, and the use of services measured against existing resources. Clear clinical criteria are needed to identify children who can be cared for appropriately on high dependency units.