Are high-volume surgeons and hospitals the most important predictors of in-hospital outcome for colon cancer resection?

Abstract

Background: Although recent studies have reported that high-volume surgeons and hospitals have better outcomes for colon cancer resections, it remains unclear whether there are other factors that are more important than volume. This study aims to evaluate the importance of the volume variables relative to other factors in an attempt to target specific areas for improving outcomes.

Methods: Using nationwide data from the Healthcare Cost and Utilization Program, full-model logistic regression was performed on all patients undergoing colon cancer resection. In hospital mortality was regressed against more than 30 different independent variables, including demographic factors (eg, age, gender, race, ethnicity, and socioeconomic status), burden of morbid and comorbid disease (prevalence and severity), and provider variables (eg, hospital size, location, teaching status, hospital and surgeon volume). A separate baseline probability analysis was then performed to compare the relative importance for all predictor variables.

Results: The sample size for this analysis was 22,408; 622 in-hospital deaths occurred (2.8%). Average age was 70 years old, 51% of particIpants were male, and 60% had at least 1 comorbid disease. An operation was elective (64%), urgent (19%), or emergency (15%). The significant predictors for mortality (at P <.05) included age, gender, comorbid disease (ie, cardiovascular, pulmonary, liver), operation severity (ie, emergency, urgent), and volume (both hospital and surgeon). The baseline probability analysis shows that the mortality for a baseline case is 12/1000. If this baseline case goes to a high-volume hospital or surgeon, the mortality will decrease to 11/1000 and 10/1000, respectively. If a patient with a baseline case of colon cancer also has coexistent liver disease or requires an emergency operation, mortality increases to 44/1000 and 45/1000, respectively. Overall, the volume variables, although statistically significant, have a relatively smaller effect on outcome compared with other factors.

Conclusions: Whereas other factors have a stronger association with outcome than volume, volume is the only acutely mutable variable. Although the regionalization controversy (ie, using only high-volume surgeons or hospitals) is not solved with our findings, this study elucidates and compares the relative importance of several different factors on outcome, which is essential when considering the conclusions and implications of this type of policy-relevant outcomes research.