Statistical inferences on nonconstant relative potency with quantal response data

Area of Science:

• Toxicology
• Pharmacology
• Biostatistics

Background:

• Relative potency is crucial for comparing chemical potencies in toxicology and pharmacology.
• The classical concept assumes parallel dose-response curves, which often fails in practice, leading to misleading conclusions.
• Nonconstant relative potency requires advanced statistical methods beyond simple ratio calculations.

Purpose of the Study:

• To estimate relative potency at multiple response levels using relative potency functions.
• To develop and evaluate statistical inference procedures for assessing nonconstant relative potency.
• To address multiplicity issues arising from simultaneous assessment at various response levels.

Main Methods:

• Utilized the log-logistic dose-response model for estimation.
• Proposed and evaluated three approaches for calculating multiplicity-adjusted confidence limits.
• Employed Monte Carlo simulations to assess the characteristics of simultaneous confidence limits.

Main Results:

• The study provides methods for estimating relative potency functions.
• Developed and evaluated multiplicity-adjusted confidence limits for nonconstant relative potency.
• Simulation results offer insights into the performance of proposed statistical inference procedures.

Conclusions:

• Relative potency functions are valuable for handling nonconstant relative potency when dose-response curves are not parallel.
• The proposed multiplicity-adjusted methods enhance the reliability of assessing nonconstant relative potency at multiple response levels.
• These statistical approaches are essential for accurate toxicological and pharmacological risk assessment.