Reassessing mechanism as a predictor of pediatric injury mortality

Area of Science:

• Pediatric trauma research
• Injury outcome prediction
• Public health surveillance

Background:

• Mechanism of injury (MOI) data is often incomplete or inaccurate in trauma databases.
• Predicting injury outcomes is crucial for effective trauma care and resource allocation.
• Understanding MOI's role in pediatric trauma mortality is essential for improving survival rates.

Purpose of the Study:

• To evaluate the significance of MOI as a predictor of mortality in injured children.
• To compare the predictive value of MOI across different injury coding methodologies.

Main Methods:

• Utilized data from the National Trauma Data Bank for children (<15 years) with blunt injuries.
• Developed predictive models for injury mortality using MOI and two injury coding schemes: abbreviated injury scale (AIS) post-dot values (low-dimensional) and International Classification of Diseases, Ninth Revision (ICD-9) codes (high-dimensional).
• Assessed model performance and MOI's variable importance with and without its inclusion in both coding schemes.

Main Results:

• Analysis of 62,569 records revealed a 0.9% mortality rate.
• MOI inclusion improved model performance with low-dimensional (AIS) coding, contributing 28% to model variance.
• MOI inclusion showed no performance improvement with high-dimensional (ICD-9) coding, contributing <1% to model variance.

Conclusions:

• MOI's predictive importance for pediatric blunt trauma mortality is contingent upon the injury coding approach.
• When detailed injury characteristics are captured via high-dimensional coding (e.g., ICD-9), MOI becomes a less critical predictor of mortality.
• Advanced injury coding schemes can effectively characterize blunt pediatric trauma, potentially rendering MOI data non-essential for outcome prediction.