Joint modeling of two longitudinal outcomes and competing risk data

Area of Science:

• Cardiovascular Medicine
• Biostatistics
• Medical Imaging

Background:

• Aortic gradient and aortic regurgitation are key echocardiographic markers of aortic valve function.
• These biomarkers are repeatedly measured in patients with valve abnormalities, suggesting a biological interrelation.
• Analyzing these outcomes jointly is crucial due to potential loss to follow-up from events like death or reoperation.

Purpose of the Study:

• To propose and apply a flexible joint statistical model for analyzing longitudinal echocardiographic data and time-to-event outcomes.
• To investigate the relationship between aortic gradient, aortic regurgitation, death, and reoperation in patients with human tissue aortic valves.
• To introduce a graphical approach for interpreting complex joint models with non-linear structures.

Main Methods:

• Development of a joint model incorporating two longitudinal outcomes (continuous aortic gradient, ordinal aortic regurgitation) and two time-to-event outcomes (death, reoperation).
• Utilized B-splines for flexible modeling of the average and subject-specific profiles of the continuous longitudinal outcome.
• Applied the proposed joint models within a Bayesian framework using serial echocardiographic data.

Main Results:

• The study demonstrates the application of a novel joint model for analyzing complex cardiovascular data.
• A graphical approach was proposed to facilitate the interpretation of results from non-linear joint models.
• The model was applied to a dataset of patients with human tissue aortic valves, examining echocardiographic markers and clinical outcomes.

Conclusions:

• Joint models offer a powerful framework for analyzing interrelated longitudinal and time-to-event data in cardiovascular research.
• The proposed flexible joint model and graphical interpretation method enhance the understanding of aortic valve disease progression and patient outcomes.
• This approach provides valuable insights for managing patients with aortic valve abnormalities.