Quality-adjusted survival estimation with periodic observations.

Area of Science:

• Biostatistics
• Clinical Trials
• Health Economics

Background:

• Quality-adjusted survival integrates longevity and quality of life.
• Analyzing quality-adjusted survival with incomplete, periodically collected data is challenging.
• Existing methods may not adequately address data sparsity in longitudinal studies.

Purpose of the Study:

• To develop and validate a statistical method for estimating quality-adjusted survival from incomplete longitudinal data.
• To provide a robust estimator for use in clinical studies where quality of life is assessed periodically.
• To demonstrate the practical application of the proposed method in a real-world clinical scenario.

Main Methods:

• Partitioning the time axis into disjoint intervals based on observed time points.
• Assuming a Markovian model for patient health status transitions.
• Estimating expected quality-adjusted survival by summing the product of quality of life and mean sojourn time within each health state and interval.
• Utilizing asymptotic normality for variance calculation.

Main Results:

• The proposed estimator for quality-adjusted survival is shown to be asymptotically normal.
• A simple variance calculation method is provided for the estimator.
• Simulation studies confirm the estimator's behavior.
• The method is successfully illustrated using a stroke clinical study.

Conclusions:

• The developed method offers a statistically sound approach to estimating quality-adjusted survival with incomplete longitudinal data.
• The estimator is practical for clinical trials and provides reliable results.
• This methodology enhances the analysis of patient-centered outcomes in health research.