Y. Song1, A. D. Tieniber1, R. E. Roses1, D. L. Fraker1, R. R. Kelz1, G. C. Karakousis1 1Hospital Of The University Of Pennsylvania,Philadelphia, PA, USA
Introduction: An increasing body of literature has demonstrated improved cancer-related outcomes for complex oncologic surgeries performed at specialized centers. We studied a national sample to determine whether these data have led to a shift in hospital referral patterns for such cases over time, using teaching hospitals (TH) as a surrogate for specialized centers.
Methods: Patients who underwent major elective resections for esophageal, gastric, liver, and pancreas cancers between 2003 and 2015 were identified using the National Inpatient Sample (NIS). Patients undergoing ventral hernia repairs (VHR) and appendectomies were used as a reference group. Primary outcome was the proportion of operations performed at TH over time. Secondary outcomes measured included trends in patient age, number of comorbidities, and inpatient mortality rate. Average annual percent change (AAPC) was calculated using weighted annual estimates. When there was a significant change in trend, segmented linear regression was performed to calculate AAPC for each time period. Statistical analyses were performed using R version 3.5.1.
Results: Between 2003 and 2015, an estimated 227,311 resections for esophageal (19,351), gastric (58,408), primary (22,743) and secondary (49,130) liver, and pancreas (77,712) cancers were performed. In the same time period, there were 338,223 VHR and 1,910,884 appendectomies. The proportion of cancer operations occurring at TH increased over time (AAPC=2.2, P<0.001, adjusted P*<0.001), with a concurrent decline at both rural (AAPC= -12, P<0.001, P*=0.004) and non-TH centers (AAPC= -10, P<0.001, P*<0.001). The largest increase in referral to TH occurred for gastrectomy (AAPC=3.5, P<0.001, P*=0.006), followed by pancreatectomy (AAPC=2.0, P<0.001, P*<0.001), esophagectomy (AAPC=1.8, P<0.001, P*=0.048), and hepatectomy (AAPC=0.98, P=0.002, P*=0.001). Notably, adjusted trends in the proportion of VHR and appendectomies performed at TH was not statistically significant. For cancer operations, there was an increase in number of patient comorbidities (AAPC=0.92, P=0.029), but decrease in inpatient mortality rate (AAPC= -5.6, P<0.001) over time. Mean patient age and number of comorbidities increased significantly for both VHR and appendectomies with no significant change in mortality rate.
Conclusions: The proportion of major operations for gastrointestinal and hepatobiliary cancers performed at TH has increased over time, driven primarily by an increase in gastrectomies. Additionally, inpatient mortality rates for these operations decreased. Further studies are needed to identify barriers that may impede appropriate referrals of complex oncologic surgeries to specialized centers.
