Data-driven desirability function to measure patients' disease progression in a longitudinal study.

Area of Science:

• Biostatistics
• Clinical Trial Methodology
• Chronic Disease Management

Background:

• Assessing chronic disease progression often involves multiple outcome measures.
• Unequal contributions of outcomes and subjective judgments can bias patient response assessments.
• Standardized methods are needed for robust evaluation of treatment efficacy.

Purpose of the Study:

• To introduce a data-driven approach using desirability functions for assessing overall patient response to treatment.
• To minimize bias in outcome assessment by estimating function shapes and weights from a gold standard.
• To provide a meaningful and interpretable overall progression score for chronic disease patients.

Main Methods:

• Utilized desirability functions to integrate multiple outcome measures into a single patient assessment.
• Developed a data-driven strategy to estimate desirability function parameters, mitigating subjective bias.
• Extended the methodology for longitudinal data analysis with multiple time points.

Main Results:

• Demonstrated the utility of desirability functions for creating an overall patient response score.
• The data-driven approach successfully minimized bias in the assessment of chronic disease progression.
• The method was validated using longitudinal data from a scleroderma clinical trial.

Conclusions:

• Desirability functions offer a robust framework for evaluating patient response using multiple outcomes in chronic diseases.
• The proposed data-driven method enhances objectivity and interpretability in clinical trial data analysis.
• This approach facilitates better comparison and clinical decision-making for patients with chronic conditions.