A time-dependent Poisson-Gamma model for recruitment forecasting in multicenter studies.

Area of Science:

• Biostatistics
• Clinical Trial Management
• Epidemiology

Background:

• Accurate forecasting of patient recruitment is essential for the successful completion of multicenter studies.
• The conventional Poisson-Gamma recruitment model, a Bayesian approach, assumes constant enrollment rates over time.
• This constant-rate assumption is a significant limitation in real-world clinical trial settings.

Purpose of the Study:

• To present a flexible generalization of the Poisson-Gamma recruitment model.
• To allow for time-varying enrollment rates in forecasting models for multicenter studies.
• To address the limitations of the constant-rate assumption in existing recruitment forecasting techniques.

Main Methods:

• The study proposes a novel Bayesian methodology for recruitment forecasting.
• Enrollment rates are modeled using B-splines to capture temporal variations.
• The generalized model is evaluated through simulation studies and real-world data analysis.

Main Results:

• The proposed B-spline approach effectively models time-varying enrollment rates.
• Simulations demonstrate the suitability of the method across diverse recruitment patterns.
• Application to the Canadian Co-infection Cohort shows accurate estimation of recruitment progression.

Conclusions:

• The flexible Bayesian model incorporating B-splines offers a significant improvement over traditional constant-rate recruitment forecasting.
• This enhanced methodology provides more realistic and accurate predictions for multicenter studies.
• The approach is valuable for optimizing clinical trial management and resource allocation.