%0 Journal Article %1 Calster2012 %A Calster, Ben Van %A Vergouwe, Yvonne %A Looman, Caspar W N %A Belle, Vanya Van %A Timmerman, Dirk %A Steyerberg, Ewout W %D 2012 %J European journal of epidemiology %K DataInterpretation DiscriminantAnalysis Female Humans LogisticModels Male Models OvarianNeoplasms OvarianNeoplasms:diagnosis OvarianNeoplasms:epidemiology Prevalence Prognosis ROCCurve RiskAssessment Statistical TesticularNeoplasms TesticularNeoplasms:diagnosis TesticularNeoplasms:epidemiology %N 10 %P 761-70 %R 10.1007/s10654-012-9733-3 %T Assessing the discriminative ability of risk models for more than two outcome categories. %U http://www.ncbi.nlm.nih.gov/pubmed/23054032 %V 27 %X The discriminative ability of risk models for dichotomous outcomes is often evaluated with the concordance index (c-index). However, many medical prediction problems are polytomous, meaning that more than two outcome categories need to be predicted. Unfortunately such problems are often dichotomized in prediction research. We present a perspective on the evaluation of discriminative ability of polytomous risk models, which may instigate researchers to consider polytomous prediction models more often. First, we suggest a "discrimination plot" as a tool to visualize the model's discriminative ability. Second, we discuss the use of one overall polytomous c-index versus a set of dichotomous measures to summarize the performance of the model. Third, we address several aspects to consider when constructing a polytomous c-index. These involve the assessment of concordance in pairs versus sets of patients, weighting by outcome prevalence, the value related to models with random performance, the reduction to the dichotomous c-index for dichotomous problems, and interpretation. We illustrate these issues on case studies dealing with ovarian cancer (four outcome categories) and testicular cancer (three categories). We recommend the use of a discrimination plot together with an overall c-index such as the Polytomous Discrimination Index. If the overall c-index suggests that the model has relevant discriminative ability, pairwise c-indexes for each pair of outcome categories are informative. For pairwise c-indexes we recommend the 'conditional-risk' method which is consistent with the analytical approach of the multinomial logistic regression used to develop polytomous risk models. %@ 1573-7284; 0393-2990

@article{Calster2012, abstract = {The discriminative ability of risk models for dichotomous outcomes is often evaluated with the concordance index (c-index). However, many medical prediction problems are polytomous, meaning that more than two outcome categories need to be predicted. Unfortunately such problems are often dichotomized in prediction research. We present a perspective on the evaluation of discriminative ability of polytomous risk models, which may instigate researchers to consider polytomous prediction models more often. First, we suggest a "discrimination plot" as a tool to visualize the model's discriminative ability. Second, we discuss the use of one overall polytomous c-index versus a set of dichotomous measures to summarize the performance of the model. Third, we address several aspects to consider when constructing a polytomous c-index. These involve the assessment of concordance in pairs versus sets of patients, weighting by outcome prevalence, the value related to models with random performance, the reduction to the dichotomous c-index for dichotomous problems, and interpretation. We illustrate these issues on case studies dealing with ovarian cancer (four outcome categories) and testicular cancer (three categories). We recommend the use of a discrimination plot together with an overall c-index such as the Polytomous Discrimination Index. If the overall c-index suggests that the model has relevant discriminative ability, pairwise c-indexes for each pair of outcome categories are informative. For pairwise c-indexes we recommend the 'conditional-risk' method which is consistent with the analytical approach of the multinomial logistic regression used to develop polytomous risk models.}, added-at = {2023-02-03T11:44:35.000+0100}, author = {Calster, Ben Van and Vergouwe, Yvonne and Looman, Caspar W N and Belle, Vanya Van and Timmerman, Dirk and Steyerberg, Ewout W}, biburl = {https://www.bibsonomy.org/bibtex/24ad801716adde7a7e5475b8dabed4e42/jepcastel}, city = {Department of Development and Regeneration, KU Leuven, University of Leuven, Herestraat 49 Box 7003, 3000, Leuven, Belgium, ben.vancalster@med.kuleuven.be.}, doi = {10.1007/s10654-012-9733-3}, interhash = {b99d4b954e24875b153049dae531efbf}, intrahash = {4ad801716adde7a7e5475b8dabed4e42}, isbn = {1573-7284; 0393-2990}, issn = {1573-7284}, journal = {European journal of epidemiology}, keywords = {DataInterpretation DiscriminantAnalysis Female Humans LogisticModels Male Models OvarianNeoplasms OvarianNeoplasms:diagnosis OvarianNeoplasms:epidemiology Prevalence Prognosis ROCCurve RiskAssessment Statistical TesticularNeoplasms TesticularNeoplasms:diagnosis TesticularNeoplasms:epidemiology}, month = {10}, note = {7087<br/>JID: 8508062; 2012/05/14 [received]; 2012/09/14 [accepted]; 2012/10/07 [aheadofprint]; ppublish;<br/>Models predictius; Avaluació de riscs}, number = 10, pages = {761-70}, pmid = {23054032}, timestamp = {2023-02-03T11:44:35.000+0100}, title = {Assessing the discriminative ability of risk models for more than two outcome categories.}, url = {http://www.ncbi.nlm.nih.gov/pubmed/23054032}, volume = 27, year = 2012 }