P145: Orthomageddon: An epidemiological analysis of weather-dependent mass-casualty incidents in a Canadian city

Introduction: Unique weather patterns on March 16th, 2017 led to 3 times the number of emergency department (ED) visits due to fall-injuries (FIs) on snow or ice compared to winter averages. The objective of the study was to identify weather-dependent differences in demographics, length-of-stay (LOS) predictors, and volume of ED presentations for winter FIs. We placed emphasis on Chinook phenomenon (rapid freeze-thaw cycles) common east of the Rocky Mountains. Methods: Patients with extremity injury due to fall on snow or ice were identified from the Alberta Health Services ED database from November 1st 2013 to March 31st 2018. We conducted regressions, chi-square analysis, bivariate correlations, and t-tests to identify differences in post-Chinook, high-volume, and regular winter patient cohorts. High-volume dates included any date with more than 25 FI presentations, representing a 400% increase from the daily average of 5. Results: We identified 3478 patients, with females more likely to present, X2 (1, N = 3480) = 443.266, p < 0.001, making up 67.8% of the total cohort. Mean age was 48.2 (SD ± 19.9) in all patients, and 48.4 (SD ± 20.0) among the post-Chinook cohort. Looking at ED LOS in the full patient cohort, age over 65 predicted longer ED LOS (mean = 4.23, SD ± 3.06) compared to younger age groups (mean = 3.42, SD ± 2.39), t(3478) = -7.37, p < 0.001]. Patients with fractures to the wrist or hand had shorter ED LOS (mean = 2.50, SD ± 5.83) than those without (mean = 10.95, SD ± 92.54), t(3478) = 2.64, p = 0.008. Among admitted patients, results were similar, with elevated inpatient LOS for patients over the age of 65 (mean = 171.71, SD ± 508.35) compared to younger patients (mean = 45.45, SD ± 39.53), t(3478) = −3.41, p = 0.001. Patients with radius fractures had shorter LOS (mean = 61.87, SD ± 210.37) compared to those without (mean = 288.83, SD ± 632.29), t(3478) = 3.87, p < 0.001. With respect to volume and weather, night-freezing events (below-freezing temperatures the preceding day, followed by freezing temperatures prior to 0600 hours the following day) were more likely to result in high FI volume (OR, 8.08; 95% CI, 5.14, 12.07; p < 0.001) as were recent Chinook events (OR, 1.39; 95% CI, 1.06, 1.81; p = 0.017). Conclusion: Chinook-induced meteorological mass-casualty events can be severe, but do not target populations distinct from winter averages. They can be predicted based on forecasted weather variations and should be considered for population-level alerts utilizing cellular technology.