LAUSR

Inter-hospital transfer (IHT) may help reduce failure-to-rescue (FTR) by transferring patients to centers with a higher level of expertise than the index hospital. We sought to identify factors associated with an IHT and examine if IHT was associated with improved outcomes after complex gastrointestinal cancer surgery. Medicare Inpatient Standard Analytic Files were utilized to identify patients with >1 postoperative complication following resection for esophageal, pancreatic, liver, or colorectal cancer between 2013 and 2017. Multivariable logistic regression was used to examine the association of different factors with the chance of IHT, as well as the impact of IHT on failure-to-rescue (FTR) and expenditures. Among 39,973 patients with >1 postoperative complications, 3090 (7.7%) patients were transferred to a secondary hospital. The median LOS at the index hospital prior to IHT was 10 days (IQR, 6–17 days). Patients who underwent IHT more often had experienced multiple complications at the index hospital compared with non-IHT patients (57.7% vs. 38.9%) (p<0.001). Transferred patients more commonly had undergone surgery at a low-volume index hospital (n=218, 60.2%) compared with non-IHT (n=10,351, 25.9%) patients (p<0.001). On multivariate analysis, hospital volume remained strongly associated with transfer to an acute care hospital (ACH) (OR 5.53; 95% CI 3.91–7.84; p<0.001), as did multiple complications (OR 2.01, 95% CI 1.56–2.57). The incidence of FTR was much higher among IHT-ACH patients (20.2%) versus non-IHT patients (11.5%) (OR 1.51, 95% CI 1.11–2.05) (p<0.001). Medicare expenditures were higher among patients who had IHT-ACH ($72.1k USD; IQR, $48.1k–$116.7k) versus non-IHT ($38.5k USD; IQR, $28.1k–$59.2k USD) (p<0.001). Approximately 1 in 13 patients had an IHT after complex gastrointestinal cancer surgery. IHT was associated with high rates of FTR, which was more pronounced among patients who underwent surgery at an index low-volume hospital. IHT was associated with higher overall CMS expenditures.