Summary

An observational study was conducted at 10 pediatric hospitals throughout the US to examine the ability of the pediatric component of the Comprehensive Severity Index (CSI^â) software system to predict length of stay, cost, morbidity, and mortality; triage critically ill patients; and profile physician performance in various patient populations.  Data were collected on 16,500 pediatric patients.

Primary Objectives

1. Test the ability of the pediatric component of the CSI software system to predict length of stay, cost, morbidity, and mortality.
   

2. Test the ability of the pediatric component of the CSI software system to serve as a pediatric intensive care unit triage instrument.
   

3. Test the ability of the pediatric component of the CSI software system to facilitate physician profiling.

Methods

Admission (first 24 hours), maximum (during hospitalization), and discharge severity data were collected on 16,500 pediatric patients from 10 pediatric hospitals throughout the US. 7,178 (43.5%) of these patients were admitted to a pediatric intensive care unit during their stay.

Results

1. The pediatric component of the CSI software system was a significantly better predictor of length of stay and cost than PRISM III, with an average additional R² of 18 percentage points for both length of stay and cost, and actual disease-specific additional R² of up 32 percentage points.
   

2. The pediatric component of the CSI software system was a significantly better predictor of mortality than PRISM III, with an average additional c statistic of .10, and an additional disease-specific c statistic of up to .20.
   

3. The pediatric component of the CSI software system facilitated physician profiling by severity-adjusting patients cared for by physician specialty groups.

Impact

Significant ability of the CSI software system to predict length of stay, cost, morbidity, and mortality demonstrates its effectiveness as a tool to improve quality of care and lower cost of care for pediatric patients.