Some intelligent commentary on clinical prediction rules from Chad Cook. My initial impression was similar to his points, the subject groups fit the description of patients with easier problems, so the prediction rule was of very limited value with any complex presentations.