Predicting PICU patient survival prognosis: Pediatric Logistic Organ Dysfunction vs Pediatric Index of Mortality scores

Abstract

Background Determining prognosis of patients using scoring
systems have been done in many pediatric intensive care units
(PICU). The scoring systems frequently used are pediatric logistic
organ dy sfunction (PELOD), pediatric index of mortality (PIM)
and pediatric risk of mortality (PRISM).
Objective To compare the performance of PELOD and PIM scores
in predicting the prognosis of survival vs death in PICU patients.
Methods A prognostic test in this prospective, cohort study was
conducted in the PICU of the Kariadi General Hospital, Semarang.
PELOD and PIM calculations were performed using formulae from
previously published articles. Statistical analyses included receiver
operating curve (ROC) characteristics to describe discrimination
capacity, sensitivity, specificity, positive predictive value, negative
predictive value and accuracy.
Results Thirty-three patients fulfilling the inclusion criteria were
enrolled in the study. PELOD score for area under the ROC
was 0.87 (95% CI 0.73 to 1.0; P=0.003), while that for PIM
was 0.65 (95% CI 0.39 to 0.90; P=0.2). PELOD scores showed
sensitivity 85.7% (95% CI 59.8 to 100), specificity 84.6% (95%
CI 70.7 to 98.5), positive predictive value 60.0% (95% CI 29.6
to 90.4) negative predictive value 95.6% (95% CI 87.3 to 100)
and accuracy 84.8%. PIM scores showed sensitivity 85.7% (95%
CI 59.8 to 100), specificity 50.0% (95% CI 30.8 to 69.2), positive
predictive value 31.6% (95% CI 10,7 to 52.5), negative predictive
value 92.9% (95% CI 79.4 to 100) and accuracy 57.6%.
Conclusion PELOD scoring had better specificity, positive predictive
value, negative predictive value, accuracy and discrimination capacity
than PIM scoring for predicting the survival prognosis of patients in
the PICU. [Paediatr Indones. 2012;52:165-9].