Evaluation of clinical scoring systems in critically ill infants and children.

Four scoring systems, the Acute Physiologic Score for Children (APSC), the Physiologic Stability Index (PSI), the Paediatric Risk of Mortality (PRISM) and the Therapeutic Intervention Scoring System (TISS), were evaluated for 103 critically ill infants and children according to the Clinical Classification System (CCS) class IV. The admission scores were higher for children who died than those who lived (APSC, PSI, PRISM p less than 0.001, TISS p <0.025). In addition, the mean APSC and PSI showed significant differences (p less than 0.01) between survivors (S) and nonsurvivors (NS) in all patients, mean PRISM showed significant differences (p less than 0.01) between S and NS in all but renal failure patients and the mean TISS showed only significant differences (p less than 0.01) between S and NS with primary cardiovascular and respiratory diseases. The mortality rate was 30%. Using the 0.5 predicted risk rate, total correct prediction of admission APSC, PSI, and PRISM was 80%, 80.5% and 80% respectively. Receiver Operating Characteristic (ROC) curves drawn for each severity index were in a discriminating position. There were no significant differences between the areas under the ROC curves of the physiological scores. However, there was a significant difference between the physiologic scores and TISS (p less than 0.001). Admission APSC, PSI and PRISM excellently describe severity of illness and give prognostic information in critically ill paediatric patients. In addition, TISS gives information about the therapeutic support needed.