Interhospital variability of risk-adjusted mortality rates and associated structural factors in patients undergoing emergency laparotomy: England and Wales population-level analysis

Alexander R Darbyshire; Stuart J Mercer; Sonal Arora; Philip H Pucher

doi:10.1097/TA.0000000000004455

Interhospital variability of risk-adjusted mortality rates and associated structural factors in patients undergoing emergency laparotomy: England and Wales population-level analysis

J Trauma Acute Care Surg. 2024 Sep 27. doi: 10.1097/TA.0000000000004455. Online ahead of print.

Authors

Alexander R Darbyshire¹, Stuart J Mercer, Sonal Arora, Philip H Pucher

Affiliation

¹ From the Portsmouth Hospitals University NHS Trust (A.R.D., S.J.M., P.H.P.), Portsmouth; and Chelsea and Westminster Hospital NHS Trust (S.A.), London, England.

PMID: 39330934
DOI: 10.1097/TA.0000000000004455

Abstract

Background: Emergency surgical admissions represent the majority of general surgical workload. Interhospital variations in outcomes are well recognized. This analysis of a national laparotomy data set compared the best- and worst-performing hospitals according to 30-day mortality and examined differences in process and structural factors.

Methods: A retrospective multicenter cohort study was performed using data from the England and Wales National Emergency Laparotomy Audit (December 2013 to November 2020). The data set was divided into quintiles based on the risk-adjusted mortality calculated using the National Emergency Laparotomy Audit score risk prediction model. Primary outcome was 30-day mortality. Hospital-level factors were compared across all five quintiles, and logistic regression analysis was conducted comparing the lowest with the highest risk-adjusted mortality quintiles.

Results: Risk-adjusted 30-day mortality in the poorest performing quintile was significantly higher than that of the best performing (11.4% vs. 6.6%) despite equivalent predicted mortality (9.4% vs. 9.7%). The best-performing quintile was more likely to be a tertiary surgical (49.5% vs. 37.1%, p < 0.001) or medical school-affiliated center (26.4% vs. 18.0%, p < 0.001). In logistic regression analysis, the strongest associations were for surgery performed in a tertiary center (odds ratio, 0.690 [95% confidence interval, 0.652-0.731], p < 0.001) and if surgery was performed by a gastrointestinal specialist (0.655 [0.626-0.685], p < 0.001). Smaller differences were seen for postoperative intensive care stay (0.848 [0.808-0.890], p < 0.001) and consultant anesthetist involvement (0.900 [0.837-0.967], p = 0.004).

Discussion: This study has identified significant variability in postoperative mortality across hospitals. Structural factors such as gastrointestinal specialist delivered emergency laparotomy and tertiary surgical center status appear to be associated with improved outcomes.

Level of evidence: Original Research Article; Level II.