Evaluation of the Impact of a Less-Invasive Trunk and Pelvic Trauma Protocol on Mortality in Patients with Severe Injury by Interrupted Time-Series Analysis.

Abstract

Background and Objectives: Minimally invasive trauma management, including interventional radiology and non-operative approaches, has proven effective. Consequently, our hospital established a trauma IVR protocol called "Ohta Nishinouchi Hospital trauma protocol (ONH trauma protocol) in 2013, mainly for trunk trauma. However, the efficacy of the ONH trauma protocol has remained unverified. We aimed to assess the protocol's impact using interrupted time-series analysis (ITSA). Materials and Methods: This retrospective cohort study was conducted at Ohta Nishinouchi hospital, a tertiary emergency hospital, from January 2004 to December 2019. We included patients aged ≥ 18 years who presented to our institution due to severe trauma characterized by an Abbreviated Injury Scale of ≥3 in any region. The primary outcome was the incidence of in-hospital deaths per 100 transported patients with trauma. Multivariable logistic regression analysis was conducted with in-hospital mortality as the outcome, with no exposure before protocol implementation and with exposure after protocol implementation. Results: Overall, 4558 patients were included in the analysis. The ITSA showed no significant change in in-hospital deaths after protocol induction (level change -1.49, 95% confidence interval (CI) -4.82 to 1.84, p = 0.39; trend change -0.044, 95% CI -0.22 to 0.14, p = 0.63). However, the logistic regression analysis revealed a reduced mortality effect following protocol induction (odds ratio: 0.50, 95% CI: 0.37 to 0.66, p < 0.01, average marginal effects: -3.2%, 95% CI: -4.5 to -2.0, p < 0.01). Conclusions: The ITSA showed no association between the protocol and mortality. However, before-and-after testing revealed a positive impact on mortality. A comprehensive analysis, including ITSA, is recommended over before-and-after comparisons to assess the impact of the protocol.