Impact of hospital volume on complications and mortality following gastrectomy for gastric cancer: A nationwide Korean survey using propensity score weighting

Abstract

Purpose: This study aimed to investigate the impact of hospital volume on severe complications and mortality rates following gastrectomy for gastric cancer using nationwide survey data and propensity score weighting to adjust for baseline differences. Materials and Methods: Data were collected from 68 institutions through a nationwide survey by the Korean Gastric Cancer Association encompassing 14,076 patients who underwent surgery for gastric adenocarcinoma in 2019. After exclusion, data from 12,244 patients were included in the analysis. Hospitals were categorized into 3 volume groups: group A (<= 99 cases/year), group B (100-199 cases/ year), and group C (>= 200 cases/year). Propensity score weighting balanced the variables (age, sex, tumor-node-metastasis stage, comorbidities, American Society of Anesthesiologists score, approach methods, and extent of resection) among the hospital groups. Results: The overall complication rate was 14%, with a severe complication rate (Clavien-Dindo grade IIIa or higher) of 4.9% and a mortality rate of 0.2%. After propensity score weighting, there was no statistically significant difference in the severe complication rates among the 3 groups (P = .058). However, the mortality rates were significantly lower in groups B (0.2%) and C (0.2%) than in group A (0.7%) (P < .001). Specific complications such as intra-abdominal abscess and mechanical ileus were more prevalent in low-volume hospitals. Conclusions: Although hospital volume did not significantly affect severe postoperative complication rates after adjusting for patient characteristics, higher mortality rates were observed in lower-volume hospitals. These findings call for standardized protocols and training programs, along with the maintenance of more than 100 operations per hospital, to reduce mortality. (c) 2025 Elsevier Inc. All rights are reserved, including those for text and data mining, AI training, and similar technologies.