There are two distinct yet related goals during preoperative evaluation. The first is to accurately predict the risk for postoperative complications, and the second is to modify risk factors to reduce the risk for complications. The main purpose of our study was to contribute to accomplishing the first goal by developing and validating a risk index for predicting postoperative pneumonia using information that is readily available to clinicians before surgery. We purposely did not include potential risk factors, such as intraoperative blood loss or postoperative pain management, because this information is not typically available before surgery and is therefore not useful for preoperative risk assessment. A risk index should not be considered "incomplete" simply because it does not include all possible risk factors. Rather, the value of a risk index should be judged based on its ability to accurately predict the outcome it was designed to predict.

Excluding intraoperative and postoperative risk factors limits the risk index's contribution to accomplishing the second goal of modifying risk. Risk reduction is usually accomplished through interventions proven useful in randomized, controlled trials. However, the ability to accurately select high-risk patients usually precedes the development of useful interventions because most interventions are targeted to high-risk patients. For example, preoperative cardiac risk indexes were used for more than 20 years before the usefulness of perioperative ß-blockers for high-risk cardiac patients was proven (1). It is our sincere hope that the postoperative pneumonia risk index will serve as a similar foundation for future patient-level and hospital-level interventional studies aimed at reducing the incidence of postoperative pulmonary complications.