A probability based method for selecting the optimal personalized treatment from multiple treatments.

Area of Science:

• Biostatistics
• Medical Informatics
• Clinical Trial Design

Background:

• Personalized medicine requires effective treatment assignment strategies.
• Existing methods often rely on conditional means, limiting applicability across diverse models and error structures.

Purpose of the Study:

• To propose a novel method for optimal treatment assignment in a K-treatment scenario using individual patient covariate information.
• To develop a metric for treatment dominance that is robust and applicable to various statistical models.

Main Methods:

• The proposed method utilizes patient-specific scores derived from covariates to compare treatments.
• It employs Single Index Models to link outcome variables to covariates, offering flexibility beyond conditional mean approaches.
• The method assesses treatment dominance through surrogate quantities of conditional probabilities.

Main Results:

• Empirical investigations demonstrate high accuracy in correct treatment assignments.
• The method exhibits desirable large-sample properties and robustness against deviations from the Single Index Model structure.
• Treatment selection using the proposed metric shows no loss in average reward when treatments are closely ranked.

Conclusions:

• The proposed method provides an accurate and robust approach for optimal treatment assignment based on individual patient data.
• Real-world data analysis indicates potential improvements in average response and survival time.
• This strategy enhances personalized treatment decisions in clinical practice.