A novel risk classification system for 30-day mortality in children undergoing surgery

Area of Science:

• Pediatric Surgery
• Surgical Outcomes Research
• Clinical Risk Stratification

Background:

• Accurate preoperative risk assessment is crucial for children undergoing surgery.
• Existing systems may lack precision for diverse pediatric surgical populations.
• A need exists for a simple, objective, and validated risk classification system.

Purpose of the Study:

• To develop and validate a preoperative risk classification system for 30-day postsurgical mortality in children.
• To identify key preoperative variables associated with postsurgical mortality.
• To create and test a multi-level risk stratification tool for pediatric surgical patients.

Main Methods:

• Analysis of the National Surgical Quality Improvement Project-Pediatric (NSQIP-Ped) dataset (2012-2014).
• Classification tree analysis to identify significant preoperative predictors of 30-day mortality (D30).
• Validation using 2015 NSQIP-Ped data and application to high-risk subpopulations (neonates, specific CPT codes).

Main Results:

• A five-level risk classification system was developed, demonstrating high accuracy.
• Key predictors included need for ventilation, oxygen support, inotropic support, sepsis, emergent surgery, and do-not-resuscitate orders.
• Observed D30 mortality ranged from 0.075% (Very Low Risk) to 38.6% (Very High Risk).
• The system maintained predictive accuracy in neonates and across different surgical complexity levels.

Conclusions:

• The developed five-level system provides an objective and accurate method for risk stratification in pediatric surgery.
• This classification system effectively differentiates risk levels, aiding in clinical decision-making and resource allocation.
• The validated tool is applicable across a wide range of pediatric operations, including high-risk cases and neonates.