These include: Other considerations include assessing the robustness of the model to influential observations and outliers, studying possible interaction between predictors, deciding whether and how to adjust the final model for over-fitting (so called shrinkage),5 and exploring the stability (reproducibility) of a model.7Studies often measure more predictors than can sensibly be used in a model, and pruning is required.
The AUC on the validation cohort was 0.66 and 0.76, respectively. Blood biomarkers, including lactate dehydrogenase, C-reactive protein, osteopontin, carbonic anhydrase IX, interleukin (IL) 6, IL-8, carcinoembryonic antigen (CEA), and cytokeratin fragment 21-1, were measured.
Conclusions: The performance of the prognostic model for survival improved markedly by adding two blood biomarkers: CEA and IL-6. A multivariate model, built on a large patient population (N = 322) and externally validated, was used as a baseline model.
We assume here that the available data are sufficiently accurate for prognosis and adequately represent the population of interest.
Before starting to develop a multivariable prediction model, numerous decisions must be made that affect the model and therefore the conclusions of the research.
A multivariate model, built on a large patient population (N = 322) and externally validated, was used as a baseline model.