| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 2008;135:1259-1260
© 2008 The American Association for Thoracic Surgery
Invited Commentary |
I believe that most of us are happy to learn that the high-volume surgeon in a low-volume environment does like others in that environment. The main concern is, of course, that if a problem occurs, the surgeon gets most of the blame, and much less goes to the other team members. I am wondering whether further subgroup analysis is possible and whether there is a difference between the high-volume generalist cardiac surgeon compared with the high-volume CABG specialist? For the future, it might be important not to be focused on CABG alone. I am also wondering whether the low-volume surgeons are treated unfavorably in the statistical analysis here. If you have a surgeon doing 10 CABG procedures a year and he loses 1 patient, his mortality will be 10%, and if he does not lose a patient in the second year, it will be 0%, and therefore the mean will be 5%. Even if he loses no patient in the following 8 years, his mortality in this analysis will be 5% compared with that of others who lose 1 of 100 patients, where the mortality will be 1%.
I have the following questions. How does a high-volume CABG surgeon perform when he does a few procedures in another high-volume program compared with those who do a lot in the other high-volume program? How does a high-volume surgeon do in CABG, valve, and other procedures compared with a high-volume surgeon doing CABG alone in a high-volume environment, and what about the same for a low-volume environment?
Thank you for the privilege to comment on this paper.
Dr Carey. Thank you very much for those comments. We did not look at how the numbers would compare if those low-volume procedures done by high-volume surgeons were done in a high-volume institution. We did not break it down. That would have been a third dimension that got me kind of confused. Therefore we might have to look at that some point along the way, but that has not been examined.
We do not have valve data or other data for the surgeons. This is an unfortunate choice that has been made by a lot of these states. They like to look at coronary bypass surgery because it is the highest-volume operation, and they tend to ignore the valves, although I believe that they are doing some valve reporting in some of the other states now. Therefore it will eventually happen, but we do not have it in California at this time.
Dr Robert A. Guyton (Atlanta, Ga). I enjoyed the paper, and this has obviously been a controversy for a long time. The Coronary Bypass Guidelines Group looked hard at this and was unable to find convincing evidence that volume was related to performance, particularly because it is very common to find low-volume surgeons who are among the very best in the cohort.
The question I have is whether you were able to separate out the patients undergoing isolated coronary bypass because about 15% or 20% of your patients undergo combined operations, presumably patients undergoing coronary valve operations. I presume that even though only 20% of your patients are combined patients, probably 40% of the deaths are in that category. Were you able to sort that out because that might be more meaningful in that the combined patients might tend to be focused on a certain group of surgeons compared with the patients undergoing isolated coronary bypass.
Dr Carey. We do not have any mortality rates for the nonisolated cases in this group. We just used the total CABG numbers to break down the high- and low-volume surgeons because there were a few surgeons who performed a lot of nonisolated CABG procedures and not very many isolated procedures. However, the risk adjustment was only for patients undergoing isolated coronary bypass. Therefore we have no information on that other group.
Dr Guyton. The risk-adjusted mortality data are only for the isolated patients in your series, even though you collected the data for the entire series? Therefore the data that you presented are only for isolated coronary bypass?
Dr Carey. Right.
Dr Guyton. I am sorry. That was not clear. Thank you.
Dr Carey. The reason for collecting all the CABG procedures is to try to plug the loophole of moving patients back and forth from isolated and nonisolated procedures and thereby gaming our data. Therefore all of those nonisolated cases were audited.
Dr R. Scott Mitchell (Stanford, Calif). Thanks a lot for the information. I think the most worrisome feature that I see when I look at these data is the California performance compared with performance in the 3 other states. I wonder whether you had insight into that difference. Do you know the typical volume for cases per surgeon in the other states that report versus California, which has multiple programs and multiple surgeons?
Dr Carey. Well, they are higher, on average, and all of the states pretty much have, or at least all of the bigger states have, higher numbers of cases per program, and to some extent they have higher numbers of procedures per surgeon. It was interesting to me to see that the vast majority of procedures in California were actually performed by what we defined as high-volume surgeons. These are surgeons performing an average of 121 CABG procedures per year, and most of them are performing a lot of other operations as well. Therefore these are busy surgeons, and therefore the fact that we have a higher mortality rate is kind of hard to blame on anything related to volume. It might be related to something to do with percutaneous coronary intervention. We reported that some years ago at the Western—you might remember that—and we thought that the mortality rate for CABG was related to aggressive percutaneous coronary intervention performance. It is a little hard to prove that. In New York there are some data to suggest that they do not do a lot of aggressive percutaneous coronary interventions. That has been reported by some of the people from Michigan; they looked at that.
The data from California, or the percutaneous coronary intervention data anyway, are pretty similar to those from the rest of the country, whereas in New York they are not. Percutaneous coronary intervention mortality in New York is almost 0%, and it is 1
% in the rest of the country.
Dr Ralph J. Damiano (St Louis, Mo). I had one quick question for you. What struck me most about your data was the tremendous variation in results in the low-volume programs, whereas most of your high-volume programs clustered very closely around the same mortality. Have you looked at the variation in mortality rates by volume of the centers; that is, could you give us an idea of the standard deviation? If there is a wide variability, can you really accept the fact that you are taking patients at the low-volume centers and subjecting them to potentially getting a surgeon with a good mortality rate but also potentially getting surgeon with an extremely high mortality rate either because of the low volume of the hospital or the low volume of the surgeon? Is that something we can live with as a profession?
Dr Carey. The variation from year to year is not that much. The better performers tend to be better performers, and the worse performers tend to be worse performers.
Dr Damiano. But I am just looking at your scatter plots. There looks like a huge variation in the low-volume centers, whereas all of your high-volume centers were clustered around the same mortality rate.
Dr Carey. To some extent that is true, although there was one slide in there with hospital volumes that showed there was a fair scatter in the higher-volume hospitals. The scatter in the lower-volume hospitals is partly related to what the discussant brought up, that one mortality makes a difference. But for the most part, there is some consistency from year to year, even in the low-volume programs, and that was looked at by Dr Birkmeyer, who has done a lot of this volume-outcome work. He found that mortality in previous years would predict mortality in future years. We are running some statistics on that.
What was the second part of your comment?
Dr Damiano. It was just my own impression, but I do not know whether you tried to quantify the variation by volume.
Dr Carey. We have not really looked at the details of that yet.
Related Article
J. Thorac. Cardiovasc. Surg. 2008 135: 1254-1260.
| ||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
