813Risk Factors for Predicting Mortality—G H Tan et al Risk Factors for Predicting Mortality in a Paediatric Intensive Care Unit

Rapid advances in critical care technology and rising cost of medical care have spurred the development of outcome analysis including mortality risk prediction. The main objective of this study was to assess the risk factors contributing to mortality in our paediatric intensive care unit (PICU). This is a cohort study, consisting of consecutive admissions to the PICU from 1 January to 31 December 1997. The factors studied included multi-organ dysfunction syndrome (MODS), Pediatric Risk of Mortality III (PRISM III) scores in the first 24 hours (PRISM III-24), mechanical ventilation, renal replacement therapy, age, and diagnosis-related groups. Univariate and multivariate statistical methods were used. Univariate analysis showed that need for mechanical ventilation, renal replacement therapy, presence of MODS involving 3 or more organs and PRISM III-24 scores were significantly associated with outcome (P <0.0005). Relative risk of mortality in the presence of MODS and PRISM III-24 scores>8 were 11.3 (95 % CI: 3.3 to 38.3) and 15.8 (95 % CI: 2.0 to 127.8), respectively. Using Cox Proportional Hazards model, the relative risk of mortality for any new admission could be calculated by the equation RR = e0.1032*P, where P = PRISM III-24 scores