Survival models based upon the logistic and log--logistic distributions

Area of Science:

• Biostatistics
• Survival Analysis
• Statistical Software

Background:

• Censored survival data analysis is crucial in many scientific fields.
• Accurate statistical modeling is essential for interpreting survival data.
• Existing methods may require specialized software for complex regression models.

Purpose of the Study:

• To introduce LOGC and LOGLOGC, interactive FORTRAN programs for survival data analysis.
• To implement logistic and log-logistic regression models for censored data.
• To provide comprehensive statistical outputs and diagnostic tools.

Main Methods:

• Utilized maximum likelihood estimation for parameter fitting.
• Incorporated calculation of standard errors and log-likelihood values.
• Developed residual plotting for goodness-of-fit assessment.

Main Results:

• Generated maximum likelihood estimates for logistic and log-logistic models.
• Provided estimates for survivorship functions and median remaining lifetimes.
• Included confidence intervals for key survival estimates and functions.

Conclusions:

• LOGC and LOGLOGC offer robust tools for fitting logistic and log-logistic regression models.
• The programs facilitate detailed analysis of censored survival data.
• Integrated diagnostic features aid in model validation and interpretation.