Mortality Prediction After Cardiac Surgery in Children: An STS Congenital Heart Surgery Database Analysis

Area of Science:

• Cardiology
• Medical Informatics
• Public Health

Background:

• The Society of Thoracic Surgeons (STS) Congenital Heart Surgery Database (CHSD) provides risk-adjusted mortality rates for congenital heart centers.
• Current methods for case-mix adjustment in operative mortality risk prediction are not fully established.

Purpose of the Study:

• To develop and evaluate novel risk adjustment models for operative mortality in congenital heart surgery.
• To compare the performance of new models against the existing STS CHSD risk model.

Main Methods:

• A panel developed diagnosis-procedure combinations from CHSD encounters.
• Bayesian additive regression trees and lasso models were used to estimate operative mortality.
• New models incorporated diagnosis-procedure categories, procedure-specific factors, and syndromes.

Main Results:

• Across 98,825 encounters, 2.85% experienced operative mortality.
• New models demonstrated improved calibration for high-risk predictions, particularly with diagnosis-palliative procedure variables.
• Model discrimination varied, with differences noted for high-risk procedures, hospital volume, and hospitals.

Conclusions:

• Overall performance of new models was comparable to the STS CHSD model.
• Incorporating procedure-specific factors and diagnosis for palliative operations may enhance high-risk surgical procedure predictions.
• Further assessment of surgical quality evaluations using different risk adjustment models is recommended.