Quantitative methods for tracking cognitive change 3 years after coronary artery bypass surgery

Area of Science:

• Statistics
• Neuropsychology
• Cardiovascular Surgery

Background:

• Analyzing cognitive function changes after coronary artery bypass grafting (CABG) poses statistical challenges.
• Hierarchical linear models can estimate intervention effects on longitudinal biomarker data.
• Existing methods may not adequately address correlations or missing data in repeated cognitive assessments.

Purpose of the Study:

• To present and apply a novel statistical approach for analyzing cognitive function changes post-CABG.
• To estimate the effects of CABG on multiple cognitive test scores over a 36-month period.
• To provide a more precise analysis of cognitive trajectories by accounting for data correlations and dropouts.

Main Methods:

• Employed a hierarchical linear statistical model with an "analyze then summarize" strategy.
• Estimated intervention effects for each of 16 cognitive tests individually before pooling results.
• Implicitly imputed missing data using past scores and group patterns to handle dropouts.

Main Results:

• The "analyze then summarize" method identified differences in cognitive function between CABG patients and controls across individual tests and aggregate measures.
• The approach accounted for the correlation structure of the data, yielding more precise estimates.
• The model successfully handled dropouts, providing a comprehensive analysis of cognitive trajectories.

Conclusions:

• The proposed "analyze then summarize" statistical method effectively analyzes longitudinal cognitive data after CABG.
• This approach offers more accurate and precise results compared to summarizing before analysis.
• The methodology is broadly applicable to intervention studies involving multiple time-varying biomarkers.