Sample size reduction in preclinical experiments: A Bayesian sequential decision-making framework.

Area of Science:

• Preclinical research methodology
• Animal welfare in research
• Biostatistics and experimental design

Background:

• Ethical concerns regarding animal welfare are paramount in preclinical research.
• The 3Rs principles (Replacement, Reduction, Refinement) guide humane animal research.
• Current methods often lack efficient strategies for reducing animal numbers.

Purpose of the Study:

• To propose and evaluate a novel sequential framework for preclinical experiments.
• To enhance the reduction aspect of the 3Rs by enabling early stopping for efficacy or futility.
• To minimize the number of animals required in preclinical studies.

Main Methods:

• A sequential experimental design is proposed, allowing for interim analyses.
• Decision-making at each stage involves examining experimental outcomes.
• Loss functions are specified to guide early stopping decisions based on posterior expected loss.

Main Results:

• The proposed framework demonstrates potential for further reducing animal numbers compared to traditional methods.
• Simulations involving two-arm experiments with binary and continuous endpoints support the methodology.
• The design is applicable to real-world preclinical experimental scenarios.

Conclusions:

• The sequential framework offers a statistically sound approach to reduce animal usage in preclinical research.
• Implementing early stopping for efficacy or futility optimizes resource allocation and animal welfare.
• This methodology provides a valuable tool for more humane and efficient preclinical study design.