Commentary: Evaluation of specialist paediatric retrieval teams

In 1993 a British Paediatric Association working party concluded that care for critically ill children should be concentrated in paediatric intensive care units.1 This report was criticised in a response by the NHS Centre for Reviews and Dissemination for relying too heavily on opinion rather than evidence,2 but two independent reviews commissioned by the centre,2 while critical of some recommendations, did agree on the need for the most severely ill children to be cared for in dedicated paediatric intensive care units. Despite a lack of randomised controlled trials it seems likely that the weight of professional and public opinion will drive the NHS towards centralisation, resulting in many extremely ill children being transported long distances.

In 1991, only around a third of paediatric intensive care units provided specialised retrieval services.1 A study at Birmingham Children's Hospital, which relied on referring hospitals to provide transfer to the intensive care, found that a significant adverse clinical event occurred during transfer in three out of four children, with events described as life threatening in 20%.3 By contrast, Britto et al report in this issue that adverse events occurred in only two of 51 children transferred to a London intensive care unit by their specialist team.4

On the basis of these and other reports5 the need for specialised retrieval teams seems incontrovertible. Closer examination of the studies, however, reveals the difficulties associated with the interpretation of observational data. Nearly two thirds of patients in the Birmingham study3 were neonates, most with major congenital anomalies, while around 80% included in the London study had meningococcal disease or other infections. These differences in the patient population and the lack of easily interpretable measures of severity make any estimate of the magnitude of the benefits likely to be unreliable.

Health planners need to estimate the magnitude of potential benefits of expensive interventions to carry out cost-benefit calculations that take account of opportunity costs. Although better care during transfer may lead to decreased morbidity and mortality, the benefits are difficult to quantify from present data. There are those who would argue that this points to the need for a randomised controlled trial to compare specialist retrieval teams with transfer by doctors from referring hospitals. It is difficult to suggest in the face of these studies or indeed on the basis of clinical experience, however, that specialist teams do not provide some advantage. A trial would require parents to agree to randomisation not because there is genuine uncertainty as to which intervention is better for their child but because there is an overall benefit to the community in estimating the magnitude of the benefit provided by the more expensive option.

In these circumstances a trial would be impractical and indeed unethical. A more appropriate way forward would be to ensure that the introduction of specialised retrieval teams is accompanied by an evaluation planned at the outset. The development of more sophisticated severity scores for critically ill children to allow better comparison between units and careful costing and evaluation of newly established teams could lead to a more evidence based approach to policy making in this subject.