Related Experiment Videos
Predictive survival model with time-dependent prognostic factors: development of computer-aided SAS Macro program.
Li-Sheng Chen1, Ming-Fang Yen, Hui-Min Wu
1Institute of Public Health and Institute of Health Informatics and Decision Making, School of Medicine, National Yang-Ming University, Taipei, Taiwan.
Journal of Evaluation in Clinical Practice
|April 9, 2005
Summary
A new SAS Macro program predicts survival using time-dependent Cox regression, offering real-time cumulative survival predictions based on dynamic patient data and time-varying covariates for improved clinical decision-making.
Area of Science:
- Biostatistics
- Medical Informatics
- Survival Analysis
Background:
- Time-dependent Cox regression models are crucial for survival analysis but rarely addressed by computational tools.
- Predicting survival with time-varying covariates requires specialized statistical approaches.
Purpose of the Study:
- To develop a SAS Macro program for time-dependent Cox regression predictive modeling.
- To enable accurate survival prediction using empirical survival data with time-dependent covariates.
Main Methods:
- Explicitly delineated time-dependent proportional hazards regression and partial likelihood.
- Incorporated baseline hazard using Andersen's method for dynamic cumulative survival prediction.
- Developed SAS Macro programs using SAS IML for predictive modeling and receiver operative characteristic (ROC) validation.
Main Results:
- The program was successfully applied to clinical surveillance data for small hepatocellular carcinoma (HCC).
- Time-varying predictors such as alpha-feto protein (AFP) were utilized in the model.
- Demonstrated the program's utility in analyzing outcomes for treatments like percutaneous ethanol injection (PEI) and transcatheter arterial embolization (TAE).
Conclusions:
- The developed SAS Macro program is highly effective for real-time survival prediction.
- The tool facilitates accurate cumulative survival estimations based on time-dependent covariates.
- Enhances clinical utility by providing dynamic survival insights for patient management.