How is the child score calculated in pediatric patients?

Child Score Calculation in Pediatric Patients

In pediatric patients, three main scoring systems are available for risk assessment: the Pediatric Risk of Mortality (PRISM) score, the Pediatric Index of Mortality (PIM/PIM2), and the Clinical Risk Index for Babies (CRIB II) for newborns. 1

PRISM Score

• PRISM score consists of 14 routinely measured physiologic variables across 23 variable ranges, calculated from data collected within the first 24 hours of PICU admission 2
• The original PRISM has been updated to PRISM III, which uses 72 worst-in-24-hour variables to calculate mortality risk 1
• PRISM score requires collection of the worst physiological values over a 24-hour period, making it more labor-intensive but potentially more accurate for predicting mortality 3
• The score has demonstrated good calibration in multiple studies, with the area under ROC curve of 0.667-0.78, showing adequate discriminatory capacity between survivors and non-survivors 4, 5

Pediatric Index of Mortality (PIM)

• PIM2 (updated in 2003) and PIM3 are calculated using data collected at the time of admission, requiring only ten variables for risk calculations 1
• PIM scores are generally easier to calculate than PRISM as they don't require tracking physiological variables over 24 hours 3
• Studies show PIM2 and PIM3 have slightly better discrimination ability than PRISM III with areas under ROC curve of 0.728 and 0.726 respectively 3
• PIM may perform better in populations with high mortality rates and pre-existing chronic disorders 6

Clinical Risk Index for Babies (CRIB)

• CRIB II (updated in 2003) is specifically designed for newborns less than 1 month of age, as neither PRISM nor PIM correctly predict mortality in this population 1
• This scoring system is tailored to the unique physiological parameters relevant to neonatal intensive care 1

Important Considerations for Child Score Calculation

• Accurate data collection is critically important when using these mortality prediction models 1
• Sufficient resources must be available so all information is collected and checked by a small number of properly trained staff 1
• A common pitfall is failing to collect data for children who should have received intensive care but never made it to the unit 1
• These scoring systems help evaluate the quality (effectiveness and efficiency) of pediatric intensive care and allow comparison between different units 1
• Mortality prediction models enable investigation of best practices in organizing intensive care, monitoring effects of changes in practice, and assessing relationships between severity-of-illness and length-of-stay or cost 1

Clinical Application

• These scoring systems should not be used in isolation but as part of a comprehensive assessment of the child's condition 4
• Serial assessments using these scores provide more valuable information than single measurements for detecting deterioration 4
• For accurate risk stratification, the appropriate score should be selected based on the patient's age (CRIB II for neonates, PRISM or PIM for older children) 1