Bayesian predictive probability for binary outcomes in neurodegenerative diseases.

Area of Science:

• Clinical Trials
• Biostatistics
• Neuroscience

Background:

• Adaptive clinical trials allow design modifications based on interim data.
• Bayesian predictive probability estimates the likelihood of trial success using current data.

Purpose of the Study:

• To estimate the predictive probability of success for binary outcomes in Alzheimer's disease and Ataxia patients treated with NeuroEPO plus.
• To evaluate the utility of Bayesian predictive probability in adaptive clinical trial design.

Main Methods:

• Retrospective Bayesian analysis using Phase II trial data as priors for Phase III trials.
• Calculation of predictive probabilities at interim analysis points with varying sample sizes (50, 100, 150, 176).

Main Results:

• The study indicated potential for early trial cessation due to high probabilities of success or failure.
• Adaptive design using Bayesian predictive probability could optimize sample size and trial duration.

Conclusions:

• Bayesian predictive probability is valuable for rare disease clinical trials, especially with limited treatments or heterogeneous outcomes.
• This approach enhances interim evaluations, enabling more accurate and efficient trial designs by incorporating prior information.