Monitoring potential adverse event rate differences using data from blinded trials: the canary in the coal mine

Area of Science:

• Pharmacovigilance and Drug Safety
• Clinical Trial Methodology
• Biostatistics

Background:

• New drugs may have unexpected toxicities beyond their intended use.
• FDA regulations allow unblinding for specific serious adverse events (SAEs) in blinded trials.
• Unblinding individual SAEs can threaten overall study validity and integrity.

Purpose of the Study:

• To develop a method for promptly identifying unexpected potential toxicities during blinded clinical trials.
• To provide a strategy for assessing elevated risk of adverse events without unblinding the entire trial.
• To support routine safety monitoring in drug development programs with large blinded studies.

Main Methods:

• A Bayesian approach is described for determining the likelihood of elevated risk.
• The method utilizes external data on adverse event rates in similar patient populations.
• It accommodates binomial or Poisson likelihoods, applicable to various risk metrics.

Main Results:

• The proposed Bayesian method allows for risk assessment without compromising study blinding.
• It offers a statistically sound approach to evaluating potential safety signals.
• The method is particularly suitable for ongoing monitoring of large-scale blinded clinical trials.

Conclusions:

• A Bayesian statistical framework can effectively monitor drug safety in blinded trials.
• This approach addresses concerns about compromising study validity when investigating SAEs.
• The method facilitates early detection of potential drug-induced toxicities, enhancing patient safety.