The rationale for developing scoring systems for clinical practice.

Risk scoring systems are used infrequently in thoracic surgery. Current systems accurately predict infrequent complications such as operative mortality, but most systems have not been evaluated adequately for prediction of nonfatal complications. The extant systems are well suited to risk stratification in large patient populations. However, they have not yet reached the degree of refinement that enables accurate prediction for individual patients. Efforts at improving clinical and biological predictive systems for postoperative complications are important in improving patient selection for surgery, enhancing informed consent processes, stratifying patient groups for research and internal audit, and appropriately assigning resources for postoperative management.