BERKELEY, CA (UroToday.com) - We undertook this study to see if the recently changing practice pattern for PCNL, namely a shift toward teaching and urban centers, had led to changes in the volume-outcome relationship for the operation.
Volume has been linked to outcome for a variety of other surgical procedures, but studies are heterogeneous and often heavily biased. For PCNL, the Clinical Research Office of the Endourological Society (CROES) recently published a study in which the authors extracted volume-outcome data from a large international data pool that was collected over a one-year period. They reported that higher volumes tended toward better outcomes (higher stone-free rates, lower rates of complication), with high-volume centers defined as those centers performing > 77 cases per year. Interestingly, complication rates were slightly higher at the highest-volume centers, which the authors felt may have been due to a referral effect wherein more complex (a difficult adjective to objectify in a study) patients were clustered at the referral centers.
In our study, we used the Nationwide Inpatient Sample (NIS), which is a large publicly available administrative dataset that describes outcomes in the United States. The advantage of the NIS is the very large volume of data that can be captured. The primary disadvantages are related to case capture and complication capture; in addition, the dataset does not provide information on “urological” variables or outcomes such as stone size, operative time, or stone-free rate.
We categorized volume by center (i.e., hospital), which is the only way to capture volume in the NIS. One cannot extract surgeon-specific volume. We did our first analysis by median volume (for the entire cohort), but in the end felt it was more appropriate to classify the volume by quartile. In that portion of the paper, we did comparative statistics looking at differences in the frequencies of certain outcomes -- for instance, comparison in mortality rate and comparison in complication rate. For the multivariate analysis, we looked at things slightly differently. There, our outcomes were binary -- for instance, presence of a complication versus absence of a complication -- as opposed to continuous, i.e., complication rates in percentages. The variables tested were categorical or ordinal -- for instance, volume quartiles, or CCI category.
If we were to repeat the study, we felt that we might have limited the complication capture to those that were highly significant, such as renovascular complication, or thromboembolism, or urinary infection. This would have weeded out some of the complications that are less clinically significant. Unfortunately, the NIS doesn't really afford that sort of specificity, and this was certainly the primary limitation of our study.
In the end, these sorts of studies are essentially descriptive. We would caution readers against drawing hard-and-fast conclusions about volumes and outcomes, but rather use the data for academic discussion, hypothesis generation, and study design.
Written by:
Adam O. Kadlec, MD as part of Beyond the Abstract on UroToday.com. This initiative offers a method of publishing for the professional urology community. Authors are given an opportunity to expand on the circumstances, limitations etc... of their research by referencing the published abstract.
Department of Urology, Loyola University Medical Center, Maywood, Illinois USA
More Information about Beyond the Abstract