Survival kit

Survival Kit performs survival analysis using proportional hazards and frailty models implemented in Fortran 90 to model time-to-event biomedical data and account for unobserved heterogeneity.

Key Features:

• Implementation (Fortran 90): Implemented in Fortran 90.
• Proportional Hazards Models: Implements proportional hazards models to assess covariate effects on the hazard of an event over time.
• Frailty Models: Includes frailty models that incorporate random effects to account for unobserved heterogeneity.
• Response Time Handling: Supports analysis with a single response time recorded per subject.
• Hazard Function Specification: Defines the hazard as the product of a baseline hazard and an exponential function of covariates that may be fixed, random, or time-dependent.
• Model Flexibility: Supports Stratified Cox models, grouped data models, and Weibull models.
• Random Effects Distribution: Allows random effects distributed as log-gamma or normal.
• Pedigree Structure: Accounts for pedigree structures to model familial relationships.
• Bayesian Estimation: Estimates variance parameters using Bayesian methods.
• Correlation of Random Effects: Allows specification or estimation of correlation coefficients between two random effects.
• Biomedical Data Characteristics: Designed to handle complex biomedical datasets with smaller file sizes and lower effect levels compared to fields like agronomy.

Scientific Applications:

• Clinical survival analysis: Analysis of time-to-event outcomes such as time until disease recurrence or patient survival times.
• Genetic and familial studies: Modeling survival data with pedigree structures and frailty effects to account for familial correlations and unobserved heterogeneity.

Methodology:

Fits proportional hazards models extended with frailties (random effects), estimates variance parameters via Bayesian methods, and allows specification or estimation of correlations between random effects.