Pediatric Critical Care Medicine:

　　May 2014 - Volume 15 - Issue 4_suppl - p 63

　　doi: 10.1097/01.pcc.0000448994.50861.99

　　PERFORMANCE OF PIM AND PIM 2 SCORES IN THE PICU OF A TERTIARY CARE CHILDREN’S HOSPITAL IN INDIA

　　Ramachandran, B.; Managuli, A.; Ravikumar, K.G.; Sankar, J.

　　ABSTRACT

　　Background and aims: Pediatric Index of Mortality (PIM and PIM2) scores are common measures of outcomes prediction in the PICU.

　　Aims: To compare the performance of these scores in an Indian PICU in predicting mortality and determine the association of variables in these scores with outcome.

　　Methods: A prospective, observational, cohort study was undertaken in the PICU here from January - August 2012. A total of 262 consecutive admissions fulfilling inclusion criteria were enrolled. PIM and PIM2 score calculations were performed using published formulae after collecting necessary variables. Statistical analysis included Receiver Operating Characteristic curve (ROC), Standardised Mortality Ratio (SMR) and Hosmer-Lemeshow goodness of fit test. IRB waived consent.

　　Results: 306 children were admitted during the study period. 44 were excluded and 262 formed the study group, of which 41(15.6%) died. Analysis showed very good discrimination with area under ROC of 0.90 (95% CI 0.85–0.95) for PIM Score and 0.91 (95% CI 0.87–95) for PIM2 score but poor calibration by Hosmer-Lemeshow goodness of fit test (X=98.3, P=0.001 for PIM and X=100, P=0.001 for PIM2 score). Standardised Mortality Ratio was 1.55 for PIM and 1.72 for PIM2 score. Variables like dilated and fixed pupil, high base excess, mechanical ventilation, cardiac arrest prior to hospitalisation and ‘High risk diagnosis’ Group of PIM Score were significantly associated with mortality (P < 0.0001).

　　Conclusions: Both PIM and PIM2 Scores have very good discriminating power (ROC > 0.8) but poor calibration (P < 0.05). Both scores under predicted mortality (SMR >1).