Model-based network meta-analysis: Joint estimation of dose-response and time-course relationships

Area of Science:

• Biostatistics
• Clinical Trial Analysis
• Pharmacometrics

Background:

• Standard network meta-analysis (NMA) has limitations in synthesizing evidence across multiple doses or timepoints.
• Model-based network meta-analysis (MBNMA) offers a framework to address these limitations for dose or timecourse synthesis.
• Integrating dose-response and time-course data is crucial for comprehensive evidence synthesis.

Purpose of the Study:

• To propose a joint dose-response and time-course MBNMA (DT-MBNMA) framework.
• To enable evidence synthesis across multiple timepoints and dose levels simultaneously.
• To validate the DT-MBNMA framework and demonstrate its utility in drug development.

Main Methods:

• Development of a joint dose-response and time-course MBNMA (DT-MBNMA) model.
• Validation through a simulation study to assess parameter recovery and precision.
• Application to a real-world dataset of randomized clinical trials (RCTs) for obesity treatment.

Main Results:

• The DT-MBNMA framework successfully integrated data from early and late-stage clinical studies.
• Simulation studies demonstrated unbiased recovery of drug effect parameters.
• The methodology showed increased statistical efficiency for indirect treatment comparisons (ITC) compared to standard NMA.
• Application to GLP-1 RAs for obesity demonstrated greater precision in treatment effects.

Conclusions:

• The proposed DT-MBNMA framework effectively synthesizes evidence across multiple doses and timepoints.
• This approach enhances statistical efficiency and precision in indirect treatment comparisons.
• DT-MBNMA supports robust clinical decision-making throughout the drug development process.