Pre-ICU PRISM and PIM both discriminate well but calibrate poorly when applied to a cohort of children with meningococcal sepsis. Both scores provide an effective means of stratification for clinical trial purposes. The main advantage for PIM appears to be ease of data collection.