How about a test? Isn't there some test we can do to tell who is going to have a problem? Let's examine what we would like in the test. First, it must be able to predict who is going to have perioperative cardiac morbidity (a.k.a. death, MI, etc). Second, the test must have low risk and high predictive value because we are going to apply it to 6 million patients per year in the U.S.. Thirdly, the cost of the test times 6 million, plus the therapy times some fraction of 6 million, must be affordable. Finally, the risk of the test, plus the risk of the therapy, must be less than some other approach.