Simultaneous confidence bands for comparisons to placebo, with application to detecting the minimum effective dose

Area of Science:

• Biostatistics
• Pharmacometrics
• Clinical Trial Design

Background:

• Dose-response studies are crucial for determining drug efficacy and safety.
• Accurate estimation of the minimum effective dose (MED) is essential for clinical decision-making.
• Existing methods for confidence bands and MED estimation have limitations, particularly with monotone dose-response relationships.

Purpose of the Study:

• To develop a max-min approach for constructing simultaneous confidence bands in dose-response studies.
• To propose a step-down testing procedure for estimating the minimum effective dose (MED) using these bands.
• To evaluate the performance of the proposed method against existing procedures.

Main Methods:

• Utilized a max-min approach to construct two-sided and one-sided simultaneous confidence bands.
• Assumed a monotone non-decreasing dose-response curve.
• Developed a step-down testing procedure leveraging lower confidence bands for MED estimation.
• Conducted simulation studies to assess the procedure's performance.

Main Results:

• The proposed max-min approach effectively constructs simultaneous confidence bands for mean response contrasts.
• The step-down procedure accurately estimates the minimum effective dose (MED).
• Simulation results demonstrate the superiority of the proposed procedure over competitors when the MED is the lowest dose.

Conclusions:

• The novel max-min approach provides a robust method for confidence band construction in dose-response studies.
• The associated step-down procedure offers an improved way to estimate the minimum effective dose (MED).
• This methodology is particularly advantageous in scenarios where the minimum effective dose corresponds to the lowest tested dose.