I need to adjust my biomarker results with comorbidities, age and sex. Hopefully I will be able to enter the results into the ROC curve and compare it to original ROC curve (without adjustment).

I have been wondering if u could use linear regression (with age, sex and/or comorbidities being dependent and biomarker being an independent) with adjusted predicted value checked and then insert the obtained adjusted value into ROC curve as a test value? Is this the right method?

More Danica Markovic's questions See All
Similar questions and discussions