Increasing efficiency of preclinical research by group sequential designs

Area of Science:

• Preclinical experimental biomedicine
• Biostatistics
• Research methodology

Background:

• Preclinical studies predominantly utilize classical block designs, overlooking the potential of sequential designs.
• Group sequential designs are efficient and widely adopted in clinical trials but underutilized in preclinical research.

Purpose of the Study:

• Introduce group sequential designs to preclinical researchers.
• Demonstrate the efficiency gains of group sequential designs in preclinical settings.
• Advocate for reinvesting resource savings into increasing statistical power.

Main Methods:

• Simulated data analysis to compare group sequential designs with classical block designs.
• Evaluation of efficiency under varying effect sizes and sample sizes (n=18 and n=36 per group).
• Incorporation of futility stopping rules in simulations.

Main Results:

• Sequential frequentist analysis used approximately 80% of planned experimental units for large effect sizes (d=1, n=18).
• Futility stopping rules in larger trials (n=36) yielded resource savings up to 30% compared to block designs.
• Simulations confirmed efficiency improvements even with small sample sizes.

Conclusions:

• Group sequential designs enhance efficiency in preclinical experimental studies.
• Resource savings from sequential designs can bolster statistical power, addressing underpowered experiments.
• Adoption of group sequential designs can improve the value and predictiveness of preclinical research.