V. V. Simianu1, A. Basu2, R. Alfonso-Cristancho3, A. D. Flaxman4, D. R. Flum1,3 1University Of Washington,Department Of Surgery,Seattle, WA, USA 2University Of Washington,Department Of Health Services,Seattle, WA, USA 3University Of Washington,Surgical Outcomes Research Center (SORCE),Seattle, WA, USA 4University Of Washington,Institute For Health Metrics And Evaluation,Seattle, WA, USA
Introduction: Breakdown of a colorectal anastomosis is a rare but potentially life-threatening complication. Pressure testing the anastomosis by submerging it in water as air is injected (leak testing) can identify leaks intra-operatively and reduces the risk of leaks after surgery by up to 50%. Surgeons have varying opinions about the value of leak testing, and the field of behavioral economics predicts that perceived value drives behavior. We evaluated the impact of having a surgical leak on a surgeons’ leak-testing behavior during subsequent cases, to test the hypothesis that a recent leak would influence the perceived value of leak testing.
Methods: Using a prospectively gathered cohort from the Surgical Care and Outcome Assessment Program (SCOAP) in Washington State, we quantified leak testing during elective colorectal procedures with testable anastomoses (left colectomy, low anterior resection, and total abdominal colectomy) and assessed for adverse events related to leak. We describe patterns of leak testing and leaks, stratified by surgeon volume. Higher volume surgeons were defined as performing 5 or more procedures per year. To test the hypothesis of behavior change, we explored a difference-in-difference non-parametric model to compare leak testing before and after a leak.
Results: From 2008 to 2013, surgeons performed 7,497 elective colorectal operations across 46 hospitals, with a leak rate of 2.6% (n=195). Higher-volume surgeons accounted for 83.2% of the cases (n= 6,234) in the time period. Mean leak testing rate for all surgeons was 85.9%. While leaks occur more often in untested cases (3.5% vs 2.5%, p=0.05), leak events and leak testing were not different between lower- and higher-volume surgeons. The overall rate of leak testing increased for both lower-volume (76 to 88%, p=0.007) and higher-volume (82 to 88%, p=0.002) surgeons over the study. Lower-volume surgeons seem to increase their testing after a leak, as shown in Table 1. However, our difference-in-difference analytic model was limited by small sample size at the individual surgeon level. Several hundred unique surgeons’ data would be needed in each strata to detect significant differences.
Conclusion: Intraoperative leak testing appears to increase the most for lower-volume surgeons who experienced a leak, suggesting that these surgeons may attribute higher value to leak testing after a leak. For higher-volume surgeons, it may be that surgeon-specific preferences and practice style are more influential in the uptake of leak testing rather than exposure to adverse events. These insights may help in crafting quality improvement initiatives around colorectal surgery that require clinician behavior change.
