Centralization of paediatric intensive care: are critically ill children appropriately referred to a regional centre?

Objectives: To evaluate the appropriateness of emergency referrals for inter hospital transfers by local physicians in hospitals without intensive,e care facilities to a regional tertiary paediatric intensive care unit (PICU). Design: A prospective, descriptive study in a tertiary PICU and hospitals without paediatric intensive care facilities in and around the London area, UK. Patients: All patients (< 18 years) referred for emergency admission to the PICU from district hospitals (n = 436) as well as those admitted through other modes of admission (n = 286) between 1 October 1998 and 30 June 1999 were prospectively studied. Admissions and transfers were deemed appropriate if the risk of mortality using the Paediatric Risk of Mortality (PRISM II) score was greater than 1 %, and/or if the patient required a unique ICU-dependent therapy Effectiveness was estimated using PRISM II derived observed-to-expected mortality ratio. Of the 436 emergency referrals 398 (91.3%) were retrieved and transported to the PICU. Of these, 38 referrals were thought to be inappropriate after telephone consultation and were not transferred. Of the emergency referrals 376 (94.4 %) had a mortality risk greater than 1 % or required an ICU-dependent therapy on admission day Thus 86.2 % (376/436) of the referrals and 94.4 % (376/398) of transfers were considered appropriate. The PRISM II derived standardized mortality rate was 0.694 (95% CI 0.517-0.913) in the overall population and 0.613 (95 % CI 0.434-0.843) amongst the emergency referrals. Conclusion: Physicians at local hospitals within a centralized system of delivering paediatric intensive care were able to maintain adequate assessment skills in recognition and requesting for appropriate transfers of the most ill and efficiently utilized resources available at the regional centre.