Fall2012 - Brown - Introduction To Survival Analysis v3
Fall2012 - Brown - Introduction To Survival Analysis v3
Fall2012 - Brown - Introduction To Survival Analysis v3
Analysis
DUSTIN BROWN
POPULATION RESEARCH CENTER
Objectives
The basic logic behind these methods is from the life table
Events
Repeatable vs. Non-Repeatable
Single vs. Multiple
Non-Repeatable Events
Transition can occur only once (absorbing state)
Examples: 𝐴𝐴𝐴𝐴𝐴 → 𝐷𝐷𝐷𝐷, 𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁 → 𝐹𝐹𝐹𝐹𝐹 𝐵𝐵𝐵𝐵𝐵
Repeatable Events
Transition can occur more than once (non-absorbing state)
Examples: 𝑀𝑀𝑀𝑀𝑀𝑀𝑀 ↔ 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷, 𝐻𝐻𝐻𝐻𝐻𝐻𝐻 ↔ 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷
1 Note that the unit of analysis does not necessarily have to be individuals.
Basic Concepts: Time (Exposure, Duration)
Source: Blossfeld & Rohwer. 2002. Techniques of Event History Modeling, 2nd Ed. (p. 40)
Objectives
𝑃 𝑡 + ∆𝑡 > 𝑇 ≥ 𝑡 𝑇 ≥ 𝑡) 𝑓(𝑡)
𝜆 𝑡 = lim =
∆𝑡→0 ∆𝑡 𝑆(𝑡)
With some additional math, you can get the Survival Function
𝑆 𝑡 = 𝑒𝑒𝑒 −𝜆𝜆
2 Note that strictly speaking the hazard rate is a probability only in discrete-time models.
Basic Concepts: Hazard & Survival Functions
Non-Parametric Models
No assumptions about the baseline hazard distribution
Pros: Imposes the least structure, easy to estimate and interpret
Cons: Difficult to incorporate predictors (mostly descriptive)
Examples: Kaplan-Meier, Nelson-Alan, “Classic” Life Table
Parametric Models
Baseline hazard assumed to vary in a specific manner with time
Pros: Easy to incorporate covariates, gives baseline hazard to
calculate rates, smoothes “noisy” data
Cons: Imposes the most structure, need to be sure that estimated
distribution matches the data
Examples: Weibull (decrease or increase), Gompertz (exponential
increase), Exponential (constant)
Basic Concepts: Hazard & Survival Functions
Semi-Parametric Models
Baseline hazard is not pre-determined, but it must be positive.
Pros: Covariates easily incorporated, less structure than
parametric, smoothes “noisy” data
Cons: Does not provide the baseline hazard
Cannot calculate rates (absolute differences)
Can only interpret in terms of relative differentials
Any specification errors are “absorbed” into the coefficients
Examples: Cox Proportional Hazards (most popular model)
3 You must "stset" the data before estimating survival models in Stata. Type "help st" for details.
Stata Example: Exponential Model
Stata Example: Cox PH Models
SAS Example: Cox Proportional Hazards Model