Hypotheses and type I error in active-control noninferiority trials

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Drug Approval Processes

Background:

• Active-control noninferiority trials assume the control drug is effective.
• This assumption is often based on historical trials, which may yield false positive results.
• Failure to validate control effectiveness can lead to the approval of ineffective or harmful drugs.

Purpose of the Study:

• To present and suggest appropriate hypotheses for active-control noninferiority trials.
• To define and assess the false positive rate in noninferiority trial test procedures.
• To evaluate the impact of control ineffectiveness on drug approval via noninferiority trials.

Main Methods:

• Review and presentation of various noninferiority trial hypotheses.
• Definition of the false positive rate within the context of noninferiority testing.
• Simulation studies to demonstrate the false positive rate inflation.

Main Results:

• Current noninferiority trial procedures do not account for the false positive rate of historical control assessments.
• A defined false positive rate is associated with the combined testing of historical control efficacy and current noninferiority.
• Simulation results indicate a significant inflation of the false positive rate.

Conclusions:

• The effectiveness of the active control is a critical, yet often unaddressed, assumption in noninferiority trials.
• Current noninferiority trial designs may lead to an unacceptably high false positive rate, potentially approving ineffective drugs.
• Revised hypotheses and assessment methods are needed to ensure the validity of noninferiority trial outcomes.