Simulation study on regionalization of esophagectomy using data of 27,576 patients with esophageal malignant tumor.

Abstract

151 Background: The culture of evidence-based policy making in the field of cancer surgery is still at a developing stage. Although the volume-outcome relationship with esophagectomy has been recently recognized in Japan, there is no regionalization policy. This study was aimed to simulate regionalization of esophagectomy in Japan using data registered in a nationwide clinical database. Methods: The National Clinical Database (NCD) covers more than 95 per cent of all surgical procedures in Japan. The study used data of 27,476 patients with esophageal malignant tumor registered in the NCD as having undergone esophagectomy at 1040 hospitals between 2012 and 2016. The following four scenarios were tested; in scenario 1, 2, 3, and 4, a minimum volume standard was set as 2, 5, 10 and 15 cases per year. The risk-adjusted operative mortality rates after regionalization and travel distances according to patients’ place of residence were estimated for each scenario. Results: Current operative mortality rates according to hospital volume were < 2 cases: 4.8%, 2-4: 3.7%, 5-9: 2.4%, 10-14: 2.1%, and ≥15: 1.6%. In scenario 1 to 4, 598, 791, 896, and 939 hospitals, and 1,982, 4,740, 7,996, and 10,419 patients were affected by regionalization. The risk-adjusted operative mortality rates after regionalization in target patients were estimated to decrease to 2.7, 2.1, 1.8, and 1.7% in each scenario. The median travel distances after regionalization in metropolitan areas were calculated as 4 kilometer (km) in scenario 1 and 7 km in scenario 4. However, those in provincial cities and depopulated areas were 8 and 36 km in scenario 1, and 28 and 65 km in scenario 4. Conclusions: Regionalization of cancer surgery services could be simulated using the comprehensive clinical database covering almost all surgeries in the nation. The simulation showed that operative mortality rate could decrease to less than 2 % by regionalization but a minimum volume standard should be determined considering regional characteristics.