0

.Hello,everyone. I am studying the influence of one biomarker on multiple disease characteristics, and I would like to calculate its cutoff point. I created univariate ROC curves to investigate the diagnostic accuracy of this biomarker for each characteristis, and it's obviously that I have obtained different cutoff values. Could I combine this AUCs in order to obtain one general, optimal cutoff point? (something like multivariate analysis with multiple outcomes?)

Thank you in advance for you help!

Thank you for your attention and your answers! But your answers consider the case when we have MULTIPLE independent variables (predictors) and ONE dependent variable (outcome). However, I have MULTIPLE outcomes.

Julia
  • 1
  • 1
  • You may use a [logistic regression](https://www.sciencedirect.com/science/article/pii/S0929664611000696). See also a [similar approach with logistic regression](https://www.researchgate.net/post/How_do_I_create_ROC_curve_for_combined_biomarkers_in_SPSS). – Ertxiem - reinstate Monica Oct 15 '19 at 20:42
  • Looking for cutoffs in a single predictor is typically not the best way to proceed. If you have a quantitative biomarker, model it along with the other characteristics in a combined model. See [this page](https://stats.stackexchange.com/q/67560/28500) for example. Note the the "optimal" cutoff even of a complete risk model based on all relevant predictors will depend on the relative costs of false-positive and false-negative identifications, which aren't always straightforward to determine in clinical studies. – EdM Oct 15 '19 at 20:53

0 Answers0