A prospective observational study was undertaken to examine time series ICU data of pressure variables (mean arterial pressure (MAP), intracranial pressure (ICP) and cerebral perfusion pressure (CPP)) and relate their variability (SD) to outcome, together with simple graphical displays which could be useful at the ICU bedspace. Forty-three children (aged < 1-15 years) were admitted to the intensive care unit for Regional Neurosurgical Service, Edinburgh, following traumatic brain injury (TBI). The standard deviations from 221,291 validated pressure data measurements (representing three variables) were calculated for the duration of ICP monitoring (and in 48 h epochs from the time of injury). Data were displayed on polygraphs, and several well-defined 'patterns' were described. The standard deviations of MAP, ICP and CPP for the total duration of monitoring were found to be significantly related to survival (p = 0.003, <0.001 and 0.005, respectively), while the SD of ICP alone was strongly related to global recovery (p = 0.008) in the first 48 h post-injury. Patterns in 104 epochs (each of 48 h) were identified. Ninety-two were of the type I (MAP > CPP > ICP) pattern and 12 were of the non-type I pattern. Glasgow Outcome Scale scores at 12 months were significantly related to the dichotomized pattern type (Fisher's exact test p < 0.001 for both alive versus dead and independent versus dependent outcomes). Only one patient with type I pattern died in this series. While variability of ICP during the first 48 h post-injury is predictive of the outcome, the pattern behaviour of three pressure signals gives useful outcome prediction information throughout monitoring. These displays may help interpret some of the plethora of data produced at the bedside.