Introduction: Wide variability exists in emergency department (ED) syncope management. The Canadian Syncope Risk Score (CSRS) was derived and validated to predict the probability of 30-day serious outcomes after ED disposition. The objective was to identify barriers and facilitators among physicians for CSRS use to stratify risk and guide disposition decisions Methods: We conducted semi-structured interviews with physicians involved in ED syncope care at 8 Canadian sites. We used purposive sampling, contacting ED physicians, cardiologists, internists, and hospitalists until theme saturation was reached. Interview questions were designed to understand whether the CSRS recommendations are consistent with current practice, barriers and facilitators for application into practice, and intention for future CSRS use. Interviews were conducted via telephone or videoconference. Two independent raters coded interviews using an inductive approach to identify themes, with discrepancies resolved through consensus. Our methods were consistent with the Knowledge to Action Framework, which highlights the need to assess barriers and facilitators for knowledge use and for adapting new interventions into local contexts. Results: We interviewed 14 ED physicians, 7 cardiologists, and 10 hospitalists/internists across 8 sites. All physicians reported the use of electrocardiograms for patients with syncope, a key component in the CSRS criteria. Almost all physicians reported that the low risk recommendation (discharge without specific follow-up) was consistent with current practice, while less consistency was seen for moderate (15 days outpatient monitoring) and high risk recommendations (outpatient monitoring and/or admission). Key barriers to following the CSRS included a lack of access to outpatient monitoring and uncertainty over timely follow-up care. Other barriers included patient/family concerns, social factors, and necessary bloodwork. Facilitators included assisting with patient education, reassurance of their clinical gestalt, and optimal patient factors (e.g. reliability to return, support at home, few comorbidities). Conclusion: Physicians are receptive to using the CSRS tool for risk stratification and decision support. Implementation should address identified barriers, and adaptation to local settings may involve modifying the recommended clinical actions based on local resources and feasibility.

Introduction: The Canadian Syncope Risk Score (CSRS) is a validated risk tool developed using the best practices of conventional biostatistics, for predicting 30-day serious adverse events (SAE) after an Emergency Department (ED) visit for syncope. We sought to improve on the prediction ability of the CSRS and compared it to physician judgement using artificial intelligence (AI) research with modern machine learning (ML) methods. Methods: We used the prospective multicenter cohort data collected for the CSRS derivation and validation at 11 EDs across Canada over an 8-year period. The same 43 candidate variables considered for CSRS development were used to train and validate the four classes of ML models to predict 30-day SAE (death, arrhythmias, MI, structural heart disease, pulmonary embolism, hemorrhage) after ED disposition. Physician judgement was modeled using the two variables, referral for consultation and hospitalization. We compared the area under the curve (AUC) for the three models. Results: The proportion of patients who suffered 30-day SAE in the derivation cohort (N = 4030) was 3.6% and in validation phase (N = 2290) was 3.4%. Characteristics of the both cohorts were similar with no shift. The best performing ML model, a gradient boosting tree-based model used all 43 variables as predictors as opposed to the 9 final CSRS predictors. The AUC for the three models on the validation data were: best ML model 0.91 (95% CI 0.87–0.93), CSRS 0.87 (95% CI 0.83–0.90) and physician judgment 0.79 (95% CI 0.74 - 0.84). The most important predictors in the ML model were the same as the CSRS predictors. Conclusion: A ML model developed using AI method for risk-stratification of ED syncope performed with slightly better discrimination ability though not significantly different when compared to the CSRS. Both the ML model and the CSRS were better predictors of poor outcomes after syncope than physician judgement. ML models can perform with similar discrimination abilities when compared to traditional statistical models and outperform physician judgement given their ability to use all candidate variables.

Introduction: Emergency department (ED) syncope management is extremely variable. We developed practice recommendations based on the validated Canadian Syncope Risk Score (CSRS) and outpatient cardiac monitoring strategy with physician input. Methods: We used a 2-step approach. Step-1: We pooled data from the derivation and validation prospective cohort studies (with adequate sample size) conducted at 11 Canadian sites (Sep 2010 to Apr 2018). Adults with syncope were enrolled excluding those with serious outcome identified during index ED evaluation. 30-day adjudicated serious outcomes were arrhythmic (arrhythmias, unknown cause of death) and non-arrhythmic (MI, structural heart disease, pulmonary embolism, hemorrhage)]. We compared the serious outcome proportion among risk categories using Cochran-Armitage test. Step-2: We conducted semi-structured interviews using observed risk to develop and refine the recommendations. We used purposive sampling of physicians involved in syncope care at 8 sites from Jun-Dec 2019 until theme saturation was reached. Two independent raters coded interviews using an inductive approach to identify themes; discrepancies were resolved by consensus. Results: Of the 8176 patients (mean age 54, 55% female), 293 (3.6%; 95%CI 3.2-4.0%) experienced 30-day serious outcomes; 0.4% deaths, 2.5% arrhythmic, 1.1% non-arrhythmic outcomes. The serious outcome proportion significantly increased from low to high-risk categories (p < 0.001; overall 0.6% to 27.7%; arrhythmic 0.2% to 17.3%; non-arrhythmic 0.4% to 5.9% respectively). C-statistic was 0.88 (95%CI0.86–0.90). Non-arrhythmia risk per day for the first 2 days was 0.5% for medium-risk, 2% for high-risk and very low thereafter. We recruited 31 physicians (14 ED, 7 cardiologists, 10 hospitalists/internists). 80% of physicians agreed that low risk patients can be discharged without specific follow-up with inconsistencies around length of ED observation. For cardiac monitoring of medium and high-risk, 64% indicated that they don't have access; 56% currently admit high-risk patients and an additional 20% agreed to this recommendation. A deeper exploration led to following refinement: discharge without specific follow-up for low-risk, a shared decision approach for medium-risk and short course of hospitalization for high-risk patients. Conclusion: The recommendations were developed (with online calculator) based on in-depth feedback from key stakeholders to improve uptake during implementation.