Correcting for misclassification and selection effects in estimating net survival in clinical trials

Area of Science:

• Biostatistics
• Clinical Trials
• Survival Analysis

Background:

• Net survival analysis compares patient survival, accounting for cancer-specific mortality.
• Traditional methods (cause-specific, population-based) have limitations in clinical trials due to potential misclassification of death causes and patient selection bias.

Purpose of the Study:

• To propose and evaluate a novel, flexible population-based model for estimating long-term net survival in clinical trials.
• To address limitations of existing methods regarding cause-of-death misclassification and patient selection effects.

Main Methods:

• Developed a new flexible population-based model for net survival estimation.
• Assessed model performance through simulation studies.
• Applied the model to prostate cancer clinical trial data for comparison with existing methods.

Main Results:

• Cause-specific methods showed significant bias with misclassified cancer deaths.
• Standard population-based methods yielded biased estimates influenced by patient selection.
• The proposed new model demonstrated superior accuracy in estimating long-term net survival in clinical trials.

Conclusions:

• The new flexible population-based model offers more accurate net survival estimates in clinical trials.
• This methodology advances net survival analysis, particularly for multicenter trials.
• The model accounts for crucial factors like misclassified deaths and patient selection bias.