If calibration, discrimination, lift gain, precision recall, F1, Youden, Brier, AUC, and 27 other accuracy metrics can’t tell you if a prediction model (or diagnostic test, or marker) is of clinical value, what should you use instead?

Traditional accuracy metrics such as discrimination and calibration, while extremely useful for the analyst in understanding how to improve a model, cannot answer the question of whether a model should be used in clinical practice, or which of two competing models should be used. This talk will discuss decision curve analysis, a now very widely-used method that can help determine the clinical utility of a prediction model.

Speaker: Andrew Vickers, PhD, Attending Research Methodologist, Department of Epidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center 
 

To obtain Zoom link please email: Liz Buggs