J. A. Mirrielees1, S. M. Weber1, C. C. Greenberg1, J. R. Schumacher1, J. E. Scarborough1 1University Of Wisconsin,Surgery,Madison, WI, USA
Introduction: Existing federal quality initiatives primarily target a number of nonspecific postoperative complications that are easy to measure without regard for their relative value, such as surgical site infection, venous thromboembolism, adverse cardiac events, and respiratory complications. The impact of these and other procedure-specific complications on the clinical and resource utilization of pancreatectomy patients is not currently known. We employ an empirical approach to examine the potential impact of a series of complications following pancreatectomy on mortality and resource utilization in order to identify the highest value targets for quality improvement interventions.
Methods: Patients from the 2014 ACS-NSQIP Pancreatectomy-Targeted Participant Use File were included for analysis. The frequency of 2 procedure-specific and 7 non-specific postoperative complications were determined. Multivariable poisson regression with log link and robust error variance was used to determine the independent associations between individual complications and subsequent 30-day clinical (mortality, end-organ dysfunction) and resource utilization (prolonged hospitalization, hospital readmission) outcomes. Adjusted relative risk estimates from these models were used to calculate adjusted population attributable fractions (PAFs) as a measure of complication impact. The PAF describes the estimated reduction in the incidence of an adverse outcome that would be anticipated if exposure to a specific postoperative complication had been completely avoided in the study population.
Results: There were 5,047 patients who underwent pancreatectomy in the study period. The most frequent complications included bleeding (18.3%), pancreatic fistula (18.1%), organ/space surgical site infection (11.4%), and delayed gastric emptying (11.3%). Bleeding and pneumonia were the complications with the largest overall impact on 30-day mortality in our study population (see Table). Complete prevention of these complications would have resulted in reduction in mortality of 29.7% and 26.4%, respectively.
Conclusion:
Bleeding, pneumonia, pancreatic fistula, delayed gastric emptying, and organ/space surgical site infection have relatively large impacts on the clinical and resource utilization outcomes of patients who undergo pancreatectomy. Most of the complications that are targeted by existing federal quality initiatives (urinary tract infection, venous thromboembolism, and surgical site infection) have comparatively small impacts on this patient population. Redirecting initiatives towards the postoperative complications which matter the most would likely improve their effectiveness.
