February 29, 1996Statistics in Medicine
Multivariable Prognostic Models: Issues in Developing Models, Evaluating Assumptions and Adequacy, and Measuring and Reducing Errors
Authors
Discussion
Loading...
Member takes
Implication
Randomized trial evaluates predictive accuracy in multivariable regression models, suggesting improved outcomes in clinical studies.
Key Points
- This research aims to address the challenges in developing and validating multivariable regression models for clinical outcomes.
- Discusses the measurement of predictive accuracy for survival data with censoring.
- Introduces an index of predictive discrimination and calibration methods.
- Emphasizes the use of bootstrapping and cross-validation for unbiased validation.
- Models were demonstrated to improve predictive accuracy, reducing risks of poor fit and overfitting.
- The proposed modeling strategy showed effective results specifically in survival analysis contexts.
- Cox regression illustrated the advantages of calibration and discrimination measures.
Cite This Study
Harrell et al. (1996) studied this question.