Updating Markov models to integrate cross-sectional and longitudinal studies

Area of Science:

• * Medical Informatics
• * Computational Biology
• * Health Data Science

Background:

• * Cross-sectional studies offer disease snapshots but lack temporal dynamics for prognostic modeling.
• * Longitudinal studies capture disease progression over time but are often costly and time-limited.
• * Accurate modeling of disease progression is crucial for effective patient management and treatment.

Purpose of the Study:

• * To apply intelligent data analysis for building robust disease progression models using both cross-sectional and longitudinal data.
• * To learn disease trajectories from cross-sectional data, simulating progression from healthy to advanced stages.
• * To calibrate these models with longitudinal data using Baum-Welch re-estimation for enhanced dynamic parameter accuracy.

Main Methods:

• * Development of disease progression models using intelligent data analysis techniques.
• * Construction of realistic disease trajectories from healthy to advanced states using cross-sectional data.
• * Calibration of trajectory models with longitudinal data via Baum-Welch re-estimation.
• * Assessment of model improvement using Kullback-Leibler distance and Wilcoxon rank metrics.

Main Results:

• * Demonstrated the feasibility of learning disease trajectories from cross-sectional data.
• * Showcased successful calibration of learned models with longitudinal data.
• * Quantified improvements in model accuracy in reflecting underlying disease dynamics post-calibration.

Conclusions:

• * Intelligent data analysis can effectively model disease progression trajectories.
• * Model calibration using longitudinal data significantly enhances the accuracy of prognostic predictions.
• * This approach offers a powerful tool for understanding and predicting disease development.