Changing efficacies of coronary revascularization

Our results suggest that coronary artery bypass grafting prolongs life in patients with coronary artery disease. How much depends upon their underlying coronary anatomy and the underlying medical risk of the patient. These results provide a consistent framework for interpreting the results of the randomized clinical trials and the observational results reported from the CASS registry. All of the reported information is consistent with this interpretation. Our results further suggest that improvements in the efficacy of coronary artery bypass grafting are present. This improvement appears due not only to improvements in operative survival but also to improvements in the long-term survival of operative survivors (21), (23). Thus, the results at Duke would suggest that some of the decline observed nationally in coronary heart disease mortality is due to an improvement in survival of patients undergoing coronary artery bypass grafting. While the numbers of procedures are insufficient to account for a large part of the trend (1), (23), (29) our results do suggest that an increasing proportion of it may be due to coronary artery bypass grafting. Whether the results found at Duke can be generalized more broadly is an issue. Both the insurance industry and medical practice have a great stake in understanding the prognostic risk of an individual and how that risk is modified by treatment. There is also a great need to understand how that risk may be changing as a result of changes in technology. Where will this information come from? Randomized trials typically are performed for only limited periods of time and are not well suited to capture the moving target of technology as it changes. Databases like the one at Duke are useful in capturing the performance of the technology in an academic center but whether or not these results can be translated to the effectiveness of a given therapy in the overall community is not clear. Administrative databases that capture the broad experience of the community often fail to capture sufficient clinical detail to permit examination of such issues. There is a great need for large representative databases of information that include reasonably detailed clinical profiles, descriptions of therapies patients have received, and the long-term outcomes patients have experienced. Without such information, the influence on mortality and morbidity of therapeutic interventions will be difficult to detect and improvements in the performance of the technologies virtually impossible to recognize. It will not be easy to create such an information resource but it is even more difficult to justify not having it.