Historical controls and modern survival analysis

Area of Science:

• Biostatistics
• Epidemiology
• Medical Statistics

Background:

• Mortality rate comparisons have a long history.
• Modern regression models allow for individualized standardization using patient covariates.
• Accounting for random variation in standard groups is sometimes necessary.

Purpose of the Study:

• To survey critical choices and pitfalls in individualizing mortality standardization.
• To emphasize the application of Cox regression models in survival data analysis.
• To illustrate methods by comparing liver patient survival data.

Main Methods:

• Utilizing Cox regression models for survival data analysis.
• Individualizing standardization of mortality rates using patient covariates.
• Comparing observed mortality with standard rates, accounting for random variation.

Main Results:

• The study identifies key considerations and potential errors in individualized mortality comparisons.
• Cox regression models provide a framework for analyzing survival data with covariates.
• Liver patient survival post-transplantation versus conservative treatment is used as a case study.

Conclusions:

• Individualized standardization of mortality rates is increasingly important with advanced statistical models.
• Careful application of methods like Cox regression is crucial to avoid pitfalls.
• The presented methods offer a framework for robust survival data analysis in clinical settings.