There is no English translation for this web page.
Complex gastric surgery in Germany - is centralization beneficial? Observational study using national hospital discharge data.
U, Haist T, Gockel I, Mansky T, Lorenz
Arch Surg 2019; 404(1):93-101|
Purpose This observational study explored the association between hospital volume and short-term outcome following gastric resections for non-bariatric indication, aiming to contribute to the discussion on centralization of complex visceral surgery in Germany.
Methods Based on complete national hospital discharge data from 2010 to 2015, the association between hospital volume and in-hospital mortality was evaluated according to volume quintiles and volume deciles. Case-mix differences regarding surgical indication, age, sex, and comorbidities were considered for risk adjustment. In addition, rates of major complications and failure to rescue were analyzed across hospital volume categories.
Results Inpatient episodes (72,528) with gastric resection were analyzed. Risk-adjusted mortality in patients treated in very low volume hospitals (median volume of 5 surgeries per year) was higher (12.0% [95% CI 11.4 to 12.5]) compared to those treated in very high volume hospitals (50 surgeries per year; 10.6% [10.0 to 11.1]). Failure to rescue patients with complications was 28.1% [27.0 to 29.3] in very low volume hospitals and 22.7% [21.6 to 23.8] in very high volume hospitals. Differences were similar within the subgroup of patients operated for gastric cancer.
Conclusions Treatment in very high volume hospitals is associated with a lower in-hospital mortality compared to treatment in very low volume hospitals. This effect seems to be determined by the ability to rescue patients who experience complications. As the observed benefit is only related to very high volumes, the results do not clearly indicate that centralization may improve short-term results substantially, unless a very high degree of centralization would be achieved. Possibly, further research focusing on other outcome measures, such as clinical processes or long-term results, might lead to divergent conclusions.