| Abstract: | Background: It is unclear whether the use of emergency medical services (EMS) is associated with enhanced survival and decreased disability after hemorrhagic stroke and whether the effect size of EMS use differs according to the length of stay (LOS) in emergency department (ED). Methods: Adult patients (19 years and older) with acute hemorrhagic stroke who survived to admission at 29 hospitals between 2008 and 2011 were analyzed, excluding those who had symptom-to-ED arrival time of 3 h or greater, received thrombolysis or craniotomy before inter-hospital transfer, or had experienced cardiac arrest, had unknown information about ambulance use and outcomes. Exposure variable was EMS use. Endpoints were survival at discharge and worsened modified Rankin Scale (W-MRS) defined as 3 or greater points difference between pre- and post-event MRS. Adjusted odds ratios (AORs) with 95% confidence intervals (95% CIs) for the outcomes were calculated, including potential confounders (demographic, socioeconomic status, clinical parameter, comorbidity, behavior, and time of event) in the final model and stratifying patients by inter-hospital transfer and by time interval from symptom to ED arrival (S2D). ED LOS, classified into short (<120 min) and long (≥120 min), was added to the final model for testing of the interaction model. Results: A total of 2,095 hemorrhagic strokes were analyzed in which 75.6% were transported by EMS. For outcome measures, 17.4% and 41.4% were dead and had worsened MRS, respectively. AORs (95% CIs) of EMS were 0.67 (0.51–0.89) for death and 0.74 (0.59–0.92) for W-MRS in all patients. The effect size of EMS, however, was different according to LOS in ED. AORs (95% CIs) for death were 0.74 (0.54–1.01) in short LOS and 0.60 (0.44–0.83) in long LOS group. AORs (95% CIs) for W-MRS were 0.76 (0.60–0.97) in short LOS and 0.68 (0.52–0.88) in long LOS group. Conclusions: EMS transport was associated with lower hospital mortality and disability after acute hemorrhagic stroke. Effect size of EMS use for mortality was significant in patients with long ED LOS. |
