Estimating Baseline Cutoffs for DHA Dosage in Preterm Birth Prevention: A Bayesian Personalized Change-Point Analysis

Area of Science:

• Obstetrics and Gynecology
• Nutritional Science
• Biostatistics

Background:

• Preterm birth (PTB) is a major cause of infant mortality, with early preterm birth (ePTB) posing the greatest risk.
• Docosahexaenoic acid (DHA) supplementation shows potential in reducing PTB, but effectiveness varies with baseline DHA levels.
• Current high-dose DHA strategies may not be cost-effective for all, necessitating personalized approaches.

Purpose of the Study:

• To develop and validate a novel Bayesian personalized change-point model for optimizing DHA supplementation strategies.
• To identify optimal baseline DHA thresholds and distributions for personalized ePTB prevention.
• To improve secondary analysis and clinical trial design by accounting for baseline DHA heterogeneity.

Main Methods:

• Developed a Bayesian personalized change-point model integrating dynamic linear and normal mixture models.
• Estimated optimal DHA baseline thresholds and their distributions.
• Applied the model to real-world data and simulated trials.

Main Results:

• The model effectively adjusts for baseline DHA heterogeneity in secondary analyses.
• Demonstrated improved clinical trial design through personalized DHA supplementation strategies.
• Identified potential for optimizing DHA doses based on individual patient profiles.

Conclusions:

• A personalized Bayesian approach can optimize DHA supplementation for ePTB prevention based on individual baseline DHA levels.
• This model offers a framework for tailoring interventions in other studies with quantifiable patient heterogeneity.
• Personalized medicine strategies, like optimized DHA dosing, are crucial for improving maternal and infant health outcomes.