Evaluating the performance of pediatric risk of mortality (prism) score as a tool to predict mortality in children admitted to paediatric medical wards of Kenyatta National Hospital

Abstract

Background: With the growth in paediatric critical care, there is need to have a standard way of assessing severity of illness and the risk of mortality. PRISM score is based on physiological derangement by the disease process and is determined within 24 hours of admission. It has 14 variables both clinical and laboratory measures and has been used in several paediatric critical units and has been found to have good assessment of the risk of mortality. Mortality is affected by many factors other than the disease severity. Therefore it is necessary to assess the ability of this scoring system to predict mortality in a setting that is different from the original population it was developed from. Objective: To determine the prediction of probability of death at various PRISM scores Methodology: A longitudinal survey carried out in Kenyatta National Hospital targeting children aged 1 month to 12 years admitted in acute rooms within the paediatric medical wards. A focused physical examination was done and blood samples drawn within 24 hours of admission to assess the PRISM score variables. The PRISM score was tabulated and the risk of mortality calculated using a logistic regression equation Results: A total of 210 patients were enrolled for the study with a median age of 10 months with 55% being males. 61 (29%) patients died during the study. There was a 3% mortality in PRISM score between 0-9 and 80% in >29 PRISM score. The probability of death increased with increase in the PRISM score with it being 76% in PRISM score of 30 compared to 2.2% at a PRISM score of 5. Conclusion: There is increasing probability of death with increasing PRISM score with the rise being exponential from a score of 15.