Discounting phase 2 results when planning phase 3 clinical trials

Area of Science:

• Clinical Trial Design
• Biostatistics
• Pharmaceutical Development

Background:

• Sample size planning is critical for phase 3 clinical trials.
• Phase 2 trial data is often used for phase 3 sample size planning.
• Bias can arise from selecting favorable phase 2 results and population heterogeneity.

Purpose of the Study:

• To enhance phase 3 trial sample size planning using phase 2 data.
• To investigate bias adjustment methods for phase 2 treatment effect estimates.
• To evaluate 'launch criteria' for progressing development to phase 3.

Main Methods:

• Utilized an approach based on the concept of assurance.
• Adjusted (discounted) phase 2 treatment effect estimates using multiplicative and additive methods.
• Conducted simulations to investigate launch criteria with and without bias adjustment.
• Supplemented simulation results with empirical evidence.

Main Results:

• Bias adjustment of phase 2 estimates is necessary when planning phase 3 sample sizes.
• Simulations demonstrated the impact of bias adjustment on sample size calculations.
• Empirical evidence supports the need to discount phase 2 results for phase 3 planning.

Conclusions:

• Discounting phase 2 results is recommended for robust phase 3 sample size planning.
• The proposed bias adjustment methods improve the reliability of clinical trial design.
• Recommendations are provided based on simulation and empirical investigations.