Abstract
BackgroundMultivariable prediction models are important statistical tools for providing synthetic diagnosis and prognostic algorithms based on patients’ multiple characteristics. Their apparent measures for predictive accuracy usually have overestimation biases (known as ‘optimism’) relative to the actual performances for external populations. Existing statistical evidence and guidelines suggest that three bootstrap-based bias correction methods are preferable in practice, namely Harrell’s bias correction and the .632 and .632+ estimators. Although Harrell’s method has been widely adopted in clinical studies, simulation-based evidence indicates that the .632+ estimator may perform better than the other two methods. However, these methods’ actual comparative effectiveness is still unclear due to limited numerical evidence.MethodsWe conducted extensive simulation studies to compare the effectiveness of these three bootstrapping methods, particularly using various model building strategies: conventional logistic regression, stepwise variable selections, Firth’s penalized likelihood method, ridge, lasso, and elastic-net regression. We generated the simulation data based on the Global Utilization of Streptokinase and Tissue plasminogen activator for Occluded coronary arteries (GUSTO-I) trial Western dataset and considered how event per variable, event fraction, number of candidate predictors, and the regression coefficients of the predictors impacted the performances. The internal validity of C-statistics was evaluated.ResultsUnder relatively large sample settings (roughly, events per variable ≥ 10), the three bootstrap-based methods were comparable and performed well. However, all three methods had biases under small sample settings, and the directions and sizes of biases were inconsistent. In general, Harrell’s and .632 methods had overestimation biases when event fraction become lager. Besides, .632+ method had a slight underestimation bias when event fraction was very small. Although the bias of the .632+ estimator was relatively small, its root mean squared error (RMSE) was comparable or sometimes larger than those of the other two methods, especially for the regularized estimation methods.ConclusionsIn general, the three bootstrap estimators were comparable, but the .632+ estimator performed relatively well under small sample settings, except when the regularized estimation methods are adopted.
Highlights
Multivariable prediction models are important statistical tools for providing synthetic diagnosis and prognostic algorithms based on patients’ multiple characteristics
A multivariable prediction model is usually constructed using an adequate regression model based on a series of representative patients, but it is well known that their “apparent” predictive performances such as discrimination and calibration calculated in the derivation dataset are better than the actual performance for an external population [2]
Note that we focused on the evaluation of the C-statistic [16] in this work, since it is the most popular discriminant measure in clinical prediction models, because the simulation data should be too rich for the extensive studies; the generalizations to other performance measures would be discussed in the Discussion section
Summary
Multivariable prediction models are important statistical tools for providing synthetic diagnosis and prognostic algorithms based on patients’ multiple characteristics Their apparent measures for predictive accuracy usually have overestimation biases (known as ‘optimism’) relative to the actual performances for external populations. Practical guidelines (e.g., the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) statements [2, 3]) recommend optimism adjustment using principled internal validation methods, e.g., split-sample, cross-validation (CV), and bootstrapbased corrections Among these validation methods, split-sample analysis is known to provide a relatively imprecise estimate, and CV is not suitable for some performance measures [4]. Bootstrap-based methods are a good alternative because they provide stable estimates for performance measures with low biases [2, 4]
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have