Discrimination

Discrimination tests the ability of a method to determine patients who live from patients who die, based on the estimates of the probability of hospital death. A number of arbitrary cut points from zero to 1.0 (zero to 100 per cent) are applied, and for each cutpoint the expected number of deaths (those patients with a probability of hospital death greater than the cutpoint) is compared with the actual number of deaths. The results from each cutpoint applied can be plotted to give a receiver operating characteristic (ROC) curve. The greater the true-positive rate (the proportion of patients predicted to die who actually die) relative to the false-positive rate (the proportion of patients predicted to die who live), the greater the area under the ROC curve. The area may range from 0.5 (the method is no better than chance at determining patients who live from patients who die) to 1.0 (the method is perfect).

Healthy Fat Loss For A Longer Life

Healthy Fat Loss For A Longer Life

What will this book do for me? A growing number of books for laymen on the subject of health have appeared in the past decade. Never before has there been such widespread popular interest in medical science. Learn more within this guide today and download your copy now.

Get My Free Ebook


Post a comment