Analysis of Response Data for Assessing Treatment Effects in Comparative Clinical Studies

Area of Science:

• Clinical Trials
• Biostatistics
• Oncology

Background:

• Traditional treatment effect assessment often relies on binary outcomes and fixed time points.
• Current methods may not accurately capture sustained treatment response or duration of response (DOR).
• Existing statistical methods for analyzing DOR data often exclude non-responders.

Purpose of the Study:

• To introduce and illustrate a method for constructing and analyzing the probability of being in response (PBIR) curve.
• To demonstrate how the PBIR curve visualizes treatment response dynamics, including rapidity and duration.
• To advocate for the adoption of PBIR curve analysis as a robust tool in clinical studies.

Main Methods:

• Constructing a curve of current response rates (PBIR) over the entire study period.
• Calculating the area under the PBIR curve to determine the mean DOR.
• Applying the method to data from a renal cell carcinoma treatment study.

Main Results:

• The PBIR curve effectively visualizes the time course of treatment response.
• The area under the PBIR curve provides a measure of mean DOR.
• The method incorporates all patients, offering a more inclusive analysis than traditional DOR methods.

Conclusions:

• The PBIR curve offers a clinically interpretable and robust method for assessing treatment effects.
• This approach provides valuable information on both the speed and sustainability of therapeutic response.
• Clinical trialists are encouraged to utilize PBIR curve analysis for enhanced evaluation of treatment efficacy.