Addressing Survivorship Bias in Neurocognitive Outcomes After Early Complex Cardiac Surgery Using Clustering and...

Area of Science:

• Pediatric Cardiology
• Neurodevelopmental Outcomes
• Congenital Heart Disease Surgery

Background:

• Advances in cardiac surgery have improved survival for congenital heart disease (CHD).
• Neurocognitive and functional outcomes in CHD survivors have not shown similar improvement.
• Survivorship bias was hypothesized as a potential reason for unchanged outcome scores.

Purpose of the Study:

• To investigate the impact of survivorship bias on neurocognitive and functional outcomes in single ventricle CHD.
• To adjust outcome trend lines using propensity score (PS) and k-mean clustering methods.

Main Methods:

• Prospective follow-up of 266 children with single ventricle CHD post-Norwood procedure (1997-2016).
• Application of PS and k-mean clustering for outcome adjustment.
• Analysis of crude and adjusted neurocognitive/functional outcomes using multiple linear regression.

Main Results:

• Key factors for PS calculation included age at surgery, ventilation days, and circulatory arrest time.
• Adjusted analysis revealed stable Full Scale Intelligence Quotient (FSIQ) and Visual Motor Integration (VMI) scores.
• General Adaptive Composite (GAC) scores showed inconsistent trends, with a potential decrease over time.

Conclusions:

• Propensity score adjustment with clustering did not confirm survivorship bias.
• PS-adjusted models indicated stable FSIQ and VMI scores, with variable GAC findings.
• Improved survival in complex cardiac surgery was not linked to worse long-term neurocognitive outcomes.