ObjectiveWe sought to evaluate the impact of a medical directive allowing nurses to use defibrillators in automated external defibrillator-mode (AED) on in-hospital cardiac arrest (IHCA) outcomes. MethodsWe completed a health record review of consecutive IHCA for which resuscitation was attempted using a pragmatic multi-phase before-after cohort design. We report Utstein outcomes before (Jan.2012–Aug.2013;Control) the implementation of the AED medical directive following usual practice (Sept.2013–Aug.2016;Phase 1), and following the addition of a theory-based educational video (Sept.2016–Dec.2017;Phase 2). ResultsThere were 753 IHCA with the following characteristics (Before n = 195; Phase 1n = 372; Phase 2n = 186): mean age 66, 60.0% male, 79.3% witnessed, 29.1% noncardiac-monitored medical ward, 23.9% cardiac cause, and initial ventricular fibrillation/tachycardia (VF/VT) 27.2%. Comparing the Before, Phase 1 and 2: an AED was used 0 time (0.0%), 21 times (5.7%), 15 times (8.1%); mean times to 1st analysis were 7 min, 3 min and 1 min (p < 0.0001); mean times to 1st shock were 12 min, 10 min and 8 min (p = 0.32); return of spontaneous circulation (ROSC) was 63.6%, 59.4% and 58.1% (p = 0.77); survival was 24.6%, 21.0% and 25.8% (p = 0.37). Among IHCA in VF/VT (n = 165), time to 1st analysis and 1st shock decreased by 5 min (p = 0.01) and 6 min (p = 0.23), and ROSC and survival increased by 3.0% (p = 0.80) and 15.6% (p = 0.31). There was no survival benefit overall (1.2%; p = 0.37) or within noncardiac-monitored areas (-7.2%; p = 0.24). ConclusionsThe implementation of a medical directive allowing for AED use by nurses successfully improved key outcomes for IHCA victims, particularly following the theory-based education video and among the VF/VT group.