Statistical Analysis

Data are expressed as numbers and percentages, or as mean values ± standard deviations. Statistical analyses were performed using the statistical software SPSS version 11.0 (SPSS Inc., Chicago, IL, USA) and R version 3.5.1 (http://www.r-project.org). For the primary analysis, multiple linear regression models were used to assess independent associations between the treatment groups (TCS-based CS, phone-based CS, non-CS) during NBH and key task processing times (D2C time, door-to-CBC time, DTN time for IV-tPA, door-to-puncture (D2P) time for EVT) with adjustments for large vessel occlusion (LVO), oral anticoagulants, and the initial National Institutes of Health Stroke Scale (NIHSS) score as a confounder. For the secondary analysis, effects of the treatment (phone-based CS, non-CS) outside of NBH on key task processing times were also assessed using a multiple linear regression model adjusted for the abovementioned confounders. To assess the effects of TCS on IV-tPA and EVT for CS, a multivariable logistic regression model was used adjusted for LVO, oral anticoagulants, and the initial NIHSS score as a confounder. To assess the dispersion of key time metrics, the variance estimators and their confidence intervals (CIs) were used. Distributions of key time metrics were distorted to the left, and outcomes were thus natural log-transformed to ensure normality in regression residuals. An F-test was performed to compare the variance between the two groups. A p-value <0.05 was considered statistically significant for all analyses.