Introduction: In-hospital cardiac arrest (IHCA) most commonly occurs in non-monitored areas, where we observed a 10min delay before defibrillation (Phase I). Nurses (RNs) and respiratory therapists (RTs) cannot legally use Automated External Defibrillators (AEDs) during IHCA without a medical directive. We sought to evaluate IHCA outcomes following usual implementation (Phase II) vs. a Theory-Based educational program (Phase III) allowing RNs and RTs to use AEDs during IHCA. Methods: We completed a pragmatic before-after study of consecutive IHCA. We used ICD-10 codes to identify potentially eligible cases and included IHCA cases for which resuscitation was attempted. We obtained consensus on all data definitions before initiation of standardized-piloted data extraction by trained investigators. Phase I (Jan.2012-Aug.2013) consisted of baseline data. We implemented the AED medical directive in Phase II (Sept.2013-Aug.2016) using usual implementation strategies. In Phase III (Sept.2016-Dec.2017) we added an educational video informed by key constructs from a Theory of Planned Behavior survey. We report univariate comparisons of Utstein IHCA outcomes using 95% confidence intervals (CI). Results: There were 753 IHCA for which resuscitation was attempted with the following similar characteristics (Phase I n = 195; II n = 372; III n = 186): median age 68, 60.0% male, 79.3% witnessed, 29.7% non-monitored medical ward, 23.9% cardiac cause, 47.9% initial rhythm of pulseless electrical activity and 27.2% ventricular fibrillation/tachycardia (VF/VT). Comparing Phases I, II and III: an AED was used 0 times (0.0%), 21 times (5.6%), 15 times (8.1%); time to 1st rhythm analysis was 6min, 3min, 1min; and time to 1st shock was 10min, 10min and 7min. Comparing Phases I and III: time to 1st shock decreased by 3min (95%CI -7; 1), sustained ROSC increased from 29.7% to 33.3% (AD3.6%; 95%CI -10.8; 17.8), and survival to discharge increased from 24.6% to 25.8% (AD1.2%; 95%CI -7.5; 9.9). In the VF/VT subgroup, time to first shock decreased from 9 to 3 min (AD-6min; 95%CI -12; 0) and survival increased from 23.1% to 38.7% (AD15.6%; 95%CI -4.3; 35.4). Conclusion: The implementation of a medical directive allowing for AED use by RNs and RRTs successfully improved key outcomes for IHCA victims, particularly following the Theory-Based education video. The expansion of this project to other hospitals and health care professionals could significantly impact survival for VF/VT patients.