Equivalence testing for parallelism in the four-parameter logistic model.

Area of Science:

• Pharmacology
• Biostatistics
• Drug Discovery

Background:

• Assessing parallelism in dose-response curves is vital for relative potency determination in drug discovery.
• Current methods often use approximate F tests, which may not be the most suitable statistical approach.

Purpose of the Study:

• To propose an alternative method for testing parallelism of four-parameter logistic response curves.
• To reframe parallelism testing as a practical equivalence problem using intersection-union tests.

Main Methods:

• Developed a novel parallelism testing method based on intersection-union test theory.
• Applied the method to analyze two real-world examples of drug dose-response data.
• Conducted a simulation study to compare the new method with the traditional F test.

Main Results:

• The proposed intersection-union test-based method provides a more appropriate statistical inference for parallelism.
• The new approach is intuitive, easy to implement with standard software, and demonstrated through practical examples.
• Simulation results indicate favorable empirical properties for the novel testing strategy.

Conclusions:

• The intersection-union test approach offers a statistically sound and practical alternative for parallelism testing in drug discovery.
• This method enhances the reliability of relative potency assessments between test and standard drug preparations.
• The findings suggest a shift towards equivalence testing frameworks for evaluating parallelism in dose-response studies.