The discrimination of quick Paediatric Early Warning Scores in the pre-hospital setting

In our previous study, a Paediatric Early Warning Score could be calculated for only one-fifth of 102,993 children transported by ambulance to hospital, as components other than supplemental oxygen were not reliably measured: respiratory rate 90,358 (88%); Glasgow Coma Score 83,648 (81%); heart rate 83,330 (81%); time to capillary reperfusion 81,685 (79%); oxygen saturation 71,372 (69%); temperature 60,402 (59%); systolic blood pressure 37,088 (36%). We tested 12 abbreviated scores with 3-5 components. The discrimination of these 12 scores for the primary outcome (30-day mortality or admission to paediatric intensive care), as measured by the area under the receiving operator characteristic curve, ranged from 0.69 to 0.80. Scores could be calculated for at most 74,508 (72%) children when heart rate, conscious level and respiratory rate were measured, with or without supplemental oxygen: the discrimination of these two versions was 0.75 and 0.77, respectively. Optimal threshold scores of 3 and 2 for these two abbreviated versions discriminated an outcome rate of 2-3% in about one third of children from the other children who had < 1% rate of outcome.