Prediction of a binary outcome:30-day mortality after acute myocardial infarction
Binary outcomes are encountered in many medical prediction problems, including diagnostic problems (presence of disease) and prognostic outcomes (occurrence of complications, short-term mortality). In this book, one key example of a binary outcome is 30-day mortality in patients suffering from an acute myocardial infarction. A prediction model was developed in over 40,000 patients from the GUSTO-I trial. We review the development of this model according to the seven steps of the checklist for developing valid prediction models presented in Part II. In addition we discuss design and results of a number of methodological studies that were performed in the GUSTO-I data set.