In the medical field, prognostic prediction models are often constructed using Cox regression models for survival time data or logistic regression models for binary data. Internationally, various statements have been issued with the aim of ensuring the transparency of clinical research, and the TRIPOD Statement, "Statement for Transparent Reporting of Multivariate Predictive Models for Individual Prognosis and Diagnosis," recommends that all predictive model development studies should conduct some form of internal validation, especially when external validation is not performed. The generalized regression models available in JMP Pro not only allow the construction of predictive models involving variable selection, but also have a wealth of internal validation methods implemented, which may be useful for constructing prognostic prediction models. In this presentation, we will explain the theoretical background of the model construction method and internal validation method implemented in JMP Pro, and introduce how to build a prognostic prediction model based on actual data.
