The resulting ROI measures were retained for inclusion into a final multiple regression model and the discrimination accuracy for the training cohort was derived.
For the validation cohort, only those ROI measures that best discriminated the MCI group in the training cohort were entered into a multiple regression model and the discrimination accuracy was derived.
Although large, internal evaluations have shown that the SRC has good accuracy (model discrimination and calibration), external validations have been inconsistent and tend to favor a conclusion of inadequate performance.
The current study also only focused on problems with model discrimination (the risk adjuster accuracy), but poor model calibration could also cause problems in certain situations, especially if a study focused on a patient subgroup that was poorly predicted by the model.
In the paper we demonstrate how to formulate the model for a few real examples, and how to use widely available or popular statistical software (like SAS, R or S-Plus, and Stata) to fit the models, and estimate the discrimination accuracy of tests.
In the validation cohort, the discrimination accuracy of the model was 0.825 955% CI, 0.736−0.895; Figure 3B).
The data were analysed using multiple logistic regression models and summarized using measures of discrimination, accuracy, and fit.
Procedures for assessing model predictive performance or accuracy are discussed, including model discrimination, calibration, and power.
In Model 2, the effect of G × E interaction was excluded and the fitness index of the model was changed [fitness index of this model, χ = 61.216 (P <0.001), -2log likelihood = 37.706, discrimination accuracy = 96.9%].
In the POR model, one uses LOR as the measure for diagnostic discrimination accuracy, and builds statistical test based on the LOR-ratio, hence the test corresponds to whole ROCs (of general form).
