DOI: 10.1097/SLA.0b013e3181db33c7
,
Issn Print: 0003-4932
Publication Date: 2010/05/01
Facebook
Twitter
LinkedIn
Email
 |
Checking for direct PDF access through Ovid | |
Excerpt
The authors raise the question of whether volume is a dominant determinant of outcome or a confounding variable for other predictors. In our study, we demonstrate a robust association between volume and outcomes for colectomy, esophagectomy, gastrectomy, lung lobectomy, and pancreatectomy, and this has been shown in a large number of other studies. The question of subspecialty training cannot be answered directly using the HCUP-NIS database, but other published literature has shown that the relationship between surgeon specialty and outcomes does not appear to be as robust as surgeon volume.1-3
We agree that length of stay should not be overemphasized as an outcome measure, and we in no way intended to imply this; nevertheless, it is a salient outcome in the current health care reform environment in the United States. It is conventionally used as an outcome measure, and has been shown to be associated with clinically relevant outcomes, including complication rates. Our research group uses a log transformation to achieve a more normal distribution for continuous variables such as adjusted length of stay.
While it is certainly true that there is not consensus about the definition of a high volume surgeon, we chose to employ the best data to date from a systematic review of the literature, which synthesized this large body of literature.4 The authors' final point is interesting and poignant. However, if studies of volume are systematically biased towards one result or another (ie, if there is a publication bias), then meta-analyses of effect size might be biased as well. Since this study was a single data analysis and not a meta-analysis, however, it would seem that this cautionary note would not affect the veracity of our findings.
