부산대학교 도서관

Study Objective: To assess the quality of care delivered during first-responder defibrillation and to determine the potential efficacy of modifying existing automated external defibrillator designs to improve first-responder performance.Design: Prospective case series.Setting: An urban emergency medical services system providing first-responder defibrillation and paramedic care.Type Of Participants: Firefighters who completed a four-hour (two-session) course in automated external defibrillator operation.Methods: Heartstart 2000 defibrillators (Laerdal Medical Corp, Armonk, New York) were used in 241 consecutive resuscitation attempts. Written reports and memory module printouts were abstracted to assess firefighter performance of 11 critical actions. The firefighter's response to each opportunity to perform a critical action was scored using explicit pass/fail criteria.Results: Records of 235 automated external defibrillator uses (97.5%) were submitted for analysis. Firefighters shocked within 15 seconds of a "shock indicated" message in 95% of opportunities and reanalyzed the rhythm within 90 seconds of the third consecutive shock (ie, after one minute of CPR) in 80% of cases. However, firefighters reanalyzed the patient's rhythm too soon in 75% of cases, thereby interfering with recommended intervals of CPR. Firefighters failed to reanalyze the patient's rhythm after device-initiated "check patient" prompts 62% of the time. Memory modules were left in the automated external defibrillator during practice sessions in 64 cases, decreasing available memory to monitor automated external defibrillator use in the field. Three instances of failure to withhold CPR during rhythm analysis resulted in a single inappropriate patient shock. No firefighter was shocked inadvertently.Conclusion: Current device algorithms result in effective delivery of the initial three shocks. However, firefighters often fail to interpose recommended intervals of CPR between further attempts at defibrillation. Modification of existing device algorithms to provide additional visual and auditory cues may be preferable to relying on the user to recall accurately all the steps in this infrequently performed procedure.