Calster, B. van

category discrimination measure model c-statistic polytomou scenario value outcome index prediction category i approach 0.25 distribution van calster performance polytomous discrimination index tumor probability cancer calster 0.017 score patient number pair approaches wiley k cases outcome categories copyright prevalence pairwise 2610–2626 application polytomous discrimination 0.5 prediction models result fication level prediction model table roc curve event curve example leuven 0.75 category-speci fic pdis pevent pair approach pairwise c-statistics discrimination performance fidence intervals medicine steyerberg testicular metastatic research problem 0.95 1 vsr 0.05 dataset comparison journal polytomous c-statistics statistic 0.45 diagnosis concordance ci pn 1 testicular cancer model behavior target marker van calster b evaluation validation