Ⅰ. Introduction

Defibrillation is a vital intervention that increases survival rates in cardiac arrest patients (Wigginton et al., 2025). With each minute of delay in delivering defibrillation, the patient's survival rate decreases by approximately 10% (Kolte et al., 2015). However, delays in delivering prompt early defibrillation to in-hospital cardiac arrest (IHCA) patients frequently occur (Bodempudi et al., 2023; Stærk et al., 2022-b). Additionally, inappropriate defibrillation is sometimes administered (Cho et al., 2018), further compromising patient outcomes.

In many IHCA situations, the first responder is a nurse (Lin, Lin, Ho, Fu, & Koo, 2017). Therefore, if a nurse initiates defibrillation early in the cardiac arrest, it can be delivered to the patient before the code team arrives. Despite the critical importance of early defibrillation by nurses for the survival of cardiac arrest patients, nurses frequently demonstrate reluctance to perform defibrillation by themselves due to insufficient knowledge about cardiac arrest rhythms, lack of confidence, concerns regarding legal issues, or psychological burden (Heng & Wee, 2017; Uhm & Jung, 2021; Yun & Kim, 2020). Furthermore, nurses hesitate when performing defibrillation using manual defibrillators, which are primarily used in hospital settings (Yun & Kim, 2020). To address these issues, some hospitals have adopted the use of automated external defibrillators (AEDs) for IHCA response (Vaillancourt et al., 2024).

However, introducing a new AED to every nursing unit within a hospital entails significant costs and administrative burdens. The initial expense of purchasing dozens of AEDs can be considerable, and in addition to the upfront cost, managing and replacing each device’s batteries and electrode pads before their expiration, as well as performing regular functionality checks, can create substantial financial and manpower burdens for hospital operations (Smith et al., 2017). From this perspective, utilizing the automatic defibrillation mode of a manual defibrillator can allow for an efficient response to cardiac arrest patients in the early stages without adding additional costs or management burdens. Moreover, when the advanced resuscitation team arrives, it is possible to immediately switch to manual defibrillation (Manual-D) mode and continue cardiopulmonary resuscitation (CPR) without replacing the pads.

Modern monitor-defibrillators often include an automatic defibrillation (Auto-D) mode that functions similarly to an AED. In this mode, the device automatically analyzes the patient’s cardiac rhythm through pre-connected adhesive pads and determines whether a shockable rhythm is present. If a shockable rhythm is detected, the defibrillator selects a preset biphasic energy level, charges, and either delivers the shock autonomously (fully automatic mode) or prompts the operator to press the shock button (semi-automatic mode). Throughout this process, the device provides audio and/or visual prompts instructing rescuers to pause chest compressions for rhythm analysis, avoid patient contact during shock delivery, and promptly resume high-quality cardiopulmonary resuscitation afterward.

According to the CPR and Emergency Cardiovascular Care (ECC) guidelines of American Heart Association (AHA), it is recommended to use either the automatic or manual defibrillation mode depending on the operator’s proficiency (Panchal et al., 2020). It is known that the accuracy of determining the necessity for defibrillation using the Auto-D mode does not significantly differ from that of skilled healthcare providers (HCPs) in Manual-D mode (Nettinger, Wittig, Riis, Løfgren, & Lauridsen, 2025). However, because Auto-D mode requires certain duration of time to assess whether defibrillation is needed, the actual time to deliver a shock may be delayed compared to Manual-D mode. Therefore, for skilled HCPs, using the manual defibrillator allows for more rapid defibrillation (Panchal et al., 2020). Nevertheless, if a protocol for using the Auto-D mode in IHCA response is introduced, even a person who is not a CPR specialist but is the initial responder to IHCA can easily determine the need for defibrillation and perform it. Ultimately, this can be an effective strategy to reduce the time to first defibrillation (Vaillancourt et al., 2024).

Previous studies on AED use have primarily been retrospective studies, focused on out-of-hospital cardiac arrest situations, or targeted at emergency medical technicians (Kramer-Johansen et al., 2007; Lee & Hwang, 2017; Loma-Osorio et al., 2018). Several studies have examined nurses’ implementation of defibrillation using AEDs or the Auto-D mode (Hui, Low, & Lee, 2011; Vaillancourt et al., 2024; Zafari et al., 2004). However, since these are all cases from other countries, it is difficult to apply the research findings directly to domestic hospitals due to differences in clinical environments and sociocultural contexts. Furthermore, in Korea, negative perceptions about nurses performing defibrillation persist (Heo et al., 2022; Yun & Kim, 2020). Therefore, to introduce the Auto-D mode as a protocol for initial hospital cardiac arrest response, it is necessary to conduct basic research targeting HCPs, especially nurses, who comprise the largest proportion of initial responders to cardiac arrest in Korean hospitals.

Simulation is a widely used method in CPR training, enabling HCPs to practice their skills in a safe environment without actual risks to patients. Cardiac arrest remains the highest-priority emergency occurring within hospitals. HCPs participating in CPR must possess sufficient knowledge and demonstrate highly proficient skills in clinical settings. Because early defibrillation by nurses using Auto-D mode is not yet a commonly implemented technique in Korea, conducting research in actual cardiac arrest situations is problematic. Therefore, simulation was determined to be the most appropriate research method for the study.

Although it is known that introducing the use of AED or Auto-D mode for the initial response to IHCA patients can help facilitate rapid defibrillation, this approach has rarely been discussed in the context of Korean hospitals. Therefore, the overall clinical usefulness of Auto-D mode cannot be determined based solely on quantitative simulation results. Cardiac arrest response is the highest priority and occurs in very chaotic situations in clinical practice. When introducing and implementing new methods such as AEDs or Auto-D mode, it is important to identify potential barriers and facilitators (Stærk et al., 2022-a). For this reason, it is also important to assess participants’ perceptions of the Auto-D mode through debriefings after the simulation.

This study compared the use of Auto-D mode and Manual-D mode by nurses acting as initial responders to IHCA. Specifically, the aim was to compare the accuracy of determining the need for defibrillation (ECG interpretation) and the time taken to perform defibrillation in each mode within IHCA simulation scenarios. In addition, participants’ perceptions of implementing Auto-D mode in clinical practice were explored.

Ⅱ. Methods

1. Study Design

This study was a randomized controlled simulation trial in which, during IHCA scenarios, the experimental group used Auto-D mode, and the control group used Manual-D mode, comparing the accuracy of determining the need for defibrillation and the time taken to deliver defibrillation.

2. Study Participants

A total of 60 nurses were recruited from a single university hospital. The hospital where this study was conducted is a tertiary teaching hospital with a total of 1,616 beds and employs approximately 3,000 nurses. The hospital serves approximately 2,000,000 outpatients and 378,000 inpatients and around 400-500 IHCAs occur per year.

Eligible participants were certified with advanced cardiac life support (ACLS) course of the AHA or Korean Advanced Life Support (KALS) course of the Korean Association of Cardiopulmonary Resuscitation (KACPR). Participation was voluntary, and after providing a full explanation of the study’s purpose, written informed consent was obtained from all participants. The following baseline characteristics were collected: age, gender, ward assignment, job position, work experience, experience with CPR, experience with defibrillation, and experience with Auto-D mode.

The required sample size was calculated using G*Power 3.1 software, with a two-tailed test, a significance level of .05, an effect size of .80, and statistical power of .80, based on previous studies (Oh, Kim, Kim, & Lee, 2015; Park, 2017). These parameters indicated that at least 27 participants were needed per group. To account for potential participant attrition during the study period, a total of 60 participants (30 per group) were recruited. All 60 participants completed the simulation and data collection procedures; thus, data from all 60 participants were included in the final analysis.

3. Ethical Consideration

Approval of this study was obtained from the Institutional Review Board of the participating hospital (IRB No. H-2205-090-1324). Recruitment was conducted by posting a notice on the hospital's intranet bulletin board from July 1 to July 31, 2022, and nurses who wished to participate could independently apply by scanning the QR code included in the notice. On the simulation day, participants were provided with an information sheet describing the study purpose, procedures, estimated duration, and compensation for participation. Sufficient time was allotted for participants to review the information, after which written informed consent was obtained from those wishing to proceed. The information sheet specified that only essential personal data would be collected, that all data would be used exclusively for research purposes, and that participants could withdraw from the study at any time without penalty or disadvantage.

4. Study Procedures

1) Simulation settings

Simulations were conducted between August 8 and August 27, 2022, at the participants' affiliated hospital simulation center. The following equipment was utilized: Gaumard Scientific HAL S1000 simulator, equipped with an ECG rhythm and carotid pulse generator connected to the operating computer via a wireless network; ZOLL Medical Corporation R-series ALS Monitor/Defibrillator, featuring an integrated semi-automatic defibrillation mode with voice-guided prompts; and ZOLL Medical Corporation OneStep CPR A/A Electrode pads, used to establish the interface between the simulator and the defibrillator for ECG transmission and defibrillation delivery. The defibrillator was positioned on an emergency cart (E-cart) located near the entrance of the simulation room. All participants had completed at least one advanced life support (ALS) training course provided by the hospital prior to study participation, ensuring familiarity with the simulation center's location, environment, and equipment.

This simulation was designed as a 2-minute scenario, assuming an early stage of IHCA. Therefore, ECG presented in each simulation session was limited to a single type. Participants were informed that the ECG rhythms would be presented in a randomized order. However, to minimize the bias caused by the different order of ECGs, we maintained the same order of ECG presentations for all participants: pulseless electrical activity (PEA), ventricular fibrillation (VF-1), asystole, and VF-2.

Pulseless VT is also a shockable rhythm. But its incidence among unmonitored/unwitnessed IHCA patients is relatively low (Brady, Gurka, Mehring, Peberdy, & O'Connor, 2011; Hannen et al., 2023). In contrast to VF, pulseless VT requires pulse check, which can complicate the cardiac arrest response and, in the context of this study, risk prolonging the simulation and obscuring interpretation of the findings. This study represents a pioneering effort to introduce Auto-D mode into clinical practice in Korea, where there is a paucity of existing literature on the subject. Therefore, to simplify the simulation process and facilitate the interpretation of the study results, we decided to exclude VT from the scenarios.

One primary researcher and three research assistants were involved in conducting the simulations. The researcher served as the simulation operator, initiating and terminating each scenario and controlling the simulator software to modify the ECG rhythm displays. The first assistant assumed the role of the first rescuer, beginning each scenario by assessing the patient's (simulator) responsiveness in response to the operator's cue. Upon the first rescuer's request, the second rescuer (the study participant) brought the E-cart and defibrillator, attached it to the patient (simulator), and performed defibrillation when indicated. The second and third assistants were responsible for measuring and recording time using a stopwatch. To minimize errors in time measurement, the two timekeepers independently measured the time, and the average of the two recorded times was used as the result. To reduce bias from the measures, the assistants were not informed of the outcomes of this study.

2) Experimental group and control group

Participants were randomly assigned to one of two groups using an online sequence generator "http://www.random.org": the experimental group (Auto-D group) or the control group (Manual-D group).

In the experimental group, participants were instructed to use only Auto-D mode, whereas those in the control group were instructed to use only Manual-D mode. Participants were not informed of their group allocation or of the specific outcomes under investigation. This blinding was intended to minimize the risk that participants would consciously modify their performance based on prior knowledge of the study design, thereby enhancing the objectivity of the simulation-based evaluation.

Prior to simulation initiation, each participant was given verbal and written information about the study protocol and simulation scenario and was provided with a 5-minute hands-on period to get accustomed to the simulation environment, the equipment, and the defibrillation mode. And they were informed that they would assume the role of the provider responsible for defibrillation. They were instructed to perform the procedures according to what they could recall from their previous ALS training. As the second rescuer responding to the initial phase of IHCA, participants were directed to retrieve the E-cart and defibrillator, attach them to the patient, and perform defibrillation directly. To ensure participants' safety, the defibrillation energy was fixed at 10 joules throughout the simulations.

3) Defibrillation decision (ECG rhythm analysis) and defibrillation methods

Auto-D group: After pressing the "ANALYZE" button, the participants wait and then follow the next steps according to the defibrillator's voice prompt. Once the "ANALYZE" button is pressed, the defibrillator announces, "Stay clear" and automatically analyzes the rhythm, which takes approximately 7 seconds. If the analysis indicates a shockable rhythm, the defibrillator automatically charges and then prompts "Press shock"; at this point, the participants press the "SHOCK" button to deliver the shock. If a non-shockable rhythm is detected, the defibrillator announces "No shock advised. Start CPR", and the participants instruct the first rescuer to resume chest compressions.

Manual-D group: Participants were instructed to recall what they had learned in previous ALS training and, based on their memory, visually assess the rhythm displayed on the defibrillator monitor, analyze it themselves, and, if they determined that defibrillation was necessary, proceed to administer a shock.

From an ethical standpoint, all study participants were entitled to receive instruction in the new defibrillation method. Therefore, upon completion of all four simulation sessions, participants in Manual-D group received instruction on the use of Auto-D mode and were afforded the opportunity to practice the technique.

4) Simulation process

Each single session of simulation took about 2 minutes, and it took 10 to 15 minutes to complete four simulations.

The detailed process of the simulation was as follows:

5) Debriefing

When introducing a new and unfamiliar method, one that is not routinely used, into the treatment process for cardiac arrest patients, it is important to assess the perceptions of the staff who will use it and to explore potential barriers and facilitators. For this purpose, after each participant completed four simulation sessions, semi-structured debriefing lasting 10-15 minutes was conducted. The debriefings were facilitated by a single researcher who encouraged participants to discuss their emotional experiences during the simulation and to reflect on the perceived advantages and disadvantages of Auto-D mode. Participants were invited to share their feelings and perspectives openly, and the researcher deliberately avoided using leading questions that might elicit predetermined responses. Each debriefing concluded with the standardized open-ended question: "If you were required to perform defibrillation using automatic defibrillation mode on an actual cardiac arrest patient, what would be your thoughts regarding this intervention?" All debriefing sessions were audio-recorded for subsequent analysis.

5. Outcomes Measures

The primary outcomes of this study were the accuracy of defibrillation decision and defibrillation time in shockable rhythm scenarios. The secondary outcome was hands-off time in all scenarios.

The accuracy of defibrillation decision was evaluated by the researcher operating the simulations. If defibrillation was performed in a shockable rhythm or not performed in a non-shockable rhythm, it was recorded as “correct.” Conversely, if defibrillation was not performed in a shockable rhythm or performed in a non-shockable rhythm, it was recorded as “inappropriate.”

The defibrillation time and hands-off time was measured independently by two research assistants using stopwatches. Timing began simultaneously with the initiation of chest compressions by the first rescuer. The measured durations included: (1) the time from the start of chest compressions to when the second rescuer pressed the "SHOCK" button in shockable scenarios (defibrillation time), and (2) the time from when chest compressions were paused for ECG analysis to when chest compressions were resumed (hands-off time). We intended to exclude the time spent performing chest compressions during defibrillator charging from the hands-off time; however, none of the participants instructed compressions during the charging period. All measured times were recorded to the nearest 0.01 seconds.

The time required for the defibrillator to analyze the ECG and automatically complete charging in Auto-D mode was approximately 7 seconds, and the defibrillation time in Auto-D group included both the automatic analysis and charging durations.

6. Data Analysis

Statistical analysis was conducted using IBM SPSS 21.0 software. Descriptive statistics including frequency distributions and means were used to summarize the general characteristics of study participants. Homogeneity of baseline characteristics between two groups was assessed using the chi-square test for categorical variables and the independent t-test for continuous variables. The accuracy of defibrillation decision was compared between groups using chi-square test. Defibrillation time and hands-off time were evaluated for normality using Kurtosis, Skewness, and the Shapiro-Wilk test; as these variables did not satisfy the normality assumption, the Mann-Whitney U test was employed for between-group comparisons. All statistical tests used a significance level of .05.

1) Qualitative analysis

To analyze the recorded debriefing data, we employed inductive content analysis following the process of open coding, grouping, categorization, and abstraction (Elo & Kyngäs, 2008). All authors independently reviewed the transcribed interviews to familiarize themselves with the data. Through repeated careful reading, meaningful units were identified and extracted. During the initial coding phase, meaningful units were examined and assigned conceptual labels or phrases that captured the essential meaning of each unit, resulting in a comprehensive preliminary coding scheme. Microsoft Excel was used as a coding platform. In the subsequent phase of categorization and abstraction, all authors systematically reviewed the preliminary codes, analyzing similarities and differences among the assigned concepts and phrases. Codes with similar conceptual meanings were grouped and synthesized to form subthemes, which were assigned descriptive labels reflecting their content. Subthemes were then further analyzed to identify overarching conceptual patterns and relationships, and closely related subthemes were integrated and abstracted into higher-order themes. Throughout the analytical process, any disagreements in interpretation or categorization among authors were addressed through discussion until consensus was achieved.

We employed several strategies to ensure the trustworthiness of the study. First, investigator triangulation was used during the data analysis phase. All authors had participated in data analysis independently and crosschecked the emerging codes and themes to validate the findings. Whenever inconsistencies in coding arose, the research team engaged in a collaborative review of the raw transcripts. This iterative process of discussion and data re-examination continued until a final consensus regarding the most suitable codes was achieved. Furthermore, the findings were shared with several participants for member checking to confirm that the results accurately reflected their lived experiences and perspectives.

Ⅲ. Results

1. Characteristics of Participants

No significant differences were observed between two groups based on homogeneity testing for general characteristics (Table 1).

2. Primary and Secondary Outcomes

A comparison of the accuracy of defibrillation decision, time to defibrillation, and hands-off time between two groups is presented in Table 2.

No statistically significant differences were found between the two groups in the accuracy of defibrillation decision. Auto-D group achieved 100% accuracy, whereas Manual-D group achieved 95% accuracy with four inappropriate shocks in PEA and two inappropriate shocks in asystole (6 inappropriate shocks out of 120 simulations).

Regarding defibrillation time, there was no statistically significant difference between two groups in VF-1 (47.90±8.99 vs. 49.54±8.77, p=.344), but Auto-D group had a significantly shorter time in VF-2 (43.98±10.93 vs. 48.05±11.78, p=.038).

There were no statistically significant differences in hands-off times between two groups in PEA (12.31±1.97 vs. 17.83±15.13, p=.728), VF-1 (11.47±3.82 vs. 11.51±4.30, p=.631), and VF-2 (10.65±3.66 vs. 12.42±9.07, p=.584). The hands-off time of Auto-D group in asystole was significantly longer than Manual-D group (12.14±1.78 vs. 8.07±4.93, p<.001). Looking at the hands-off time in detail, Auto-D group showed little variation, with times ranging from a minimum of about 10 seconds to a maximum of about 12 seconds across all rhythm scenarios. In contrast, Manual-D group showed a relatively larger difference, with hands-off times ranging from a minimum of about 8 seconds to a maximum of about 17 seconds.

3. Qualitative Data

Qualitative analysis of the debriefing data yielded five major themes (Table 3). Although many studies employing inductive content analysis typically present results organized hierarchically as themes and subthemes, this study did not employ such a hierarchical structure due to the limited number of distinct themes identified during analysis. The following section describes the meaning of each theme.

Accuracy of rhythm analysis: Participants acknowledged that the chaotic nature of IHCA situations could compromise the accuracy of manual ECG rhythm interpretation. However, they expressed confidence that automated rhythm analysis by the defibrillator would consistently provide accurate determinations.

Possible error in analyzing rhythm: Participants expressed concern that mechanical malfunctions or technical errors could result in inaccurate automated ECG rhythm analysis by the defibrillator.

Reduced time to defibrillation: Participants anticipated that automated ECG analysis would enable rapid determination of shock necessity, thereby reducing the total time required to deliver defibrillation to the patient.

Extended analysis duration: Participants expressed concern that the automated analysis phase, which necessarily requires a certain period of time for ECG rhythm assessment, could result in prolonged intervals before HCPs could deliver necessary interventions to the patient.

Decreased psychological burden: Participants indicated that the automated defibrillator's capacity to analyze the ECG rhythm and provide definitive feedback regarding shock indication substantially reduced the psychological burden associated with manual rhythm interpretation and defibrillation decisions. This automated decision of the Auto-D mode decreased decisional hesitation, enabling nurses to confidently initiate defibrillation without fear of error.

Ⅳ. Discussion

This study compared the accuracy of defibrillation decision, time to defibrillation, and hands-off time between Auto-D mode group and Manual-D mode group in simulated IHCA scenarios.

Although differences in the accuracy of defibrillation decision between two groups did not reach statistical significance, Manual-D group demonstrated six inappropriate defibrillations in non-shockable scenarios, compared to 100% decision accuracy in Auto-D group. These findings are consistent with prior studies demonstrating that accuracy of automated rhythm analysis was not inferior to the manual evaluation by HCPs (Nettinger et al., 2025; Cheskes et al., 2015). Unnecessary defibrillation adversely affects cardiac arrest patient survival by prolonging interruptions in chest compressions. Consequently, utilizing Auto-D mode might help reduce the incidence of inappropriate shocks, potentially resulting in decreased hands-off times and improved patient outcomes. The observation that IHCA patient survival rates improved following implementation of Auto-D protocol (Vaillancourt et al., 2024; Zafari et al., 2004) demonstrates the clinical utility of Auto-D mode.

The current CPR guidelines recommend that the hands-off time should not exceed 10 seconds when checking the pulse or performing defibrillation (Kleinman et al., 2025). In general, the hands-off time in Manual-D mode is known to be shorter than that in Auto-D mode (Cheskes et al., 2015; Kramer-Johansen et al., 2007). However, although the results were not statistically significant, not only did Manual-D mode show longer hands-off times in both VF and PEA scenarios, but also, it exceeded the hands-off time of Auto-D mode by more than 5 seconds in the PEA scenario. Asystole is a rhythm with no electrical signal at all, so in Manual-D mode, it can be intuitively recognized visually that defibrillation is not needed. However, in Auto-D mode, regardless of the type of rhythm, a certain amount of time is required for analysis, which is presumed to have caused this difference. On the other hand, in cases where the rhythm is PEA, the process includes pulse checking, so compared to the asystole scenario, the hands-off time appeared to be relatively longer. Furthermore, in Manual-D group, in four cases, inappropriate defibrillation was performed, which suggests that participants may have had more difficulty judging whether defibrillation was necessary for PEA rhythm. Some nurses perceive performing defibrillation on cardiac arrest patients as a physician's domain (Vincelette, Lavoie, Fortin, & Quiroz-Martinez, 2018), and many nurses have little experience on independently carrying out defibrillation (Yun & Kim, 2020). It is presumed that this perception and lack of clinical experience make it difficult for nurses to analyze cardiac arrest rhythms and determine the need for defibrillation. Considering the results of Vaillancourt et al. (2024), which demonstrated a progressive reduction in time to first defibrillation after introducing Auto-D mode for IHCA, together with those of this study, it can be expected that nurses with limited experience in cardiac arrest rhythm analysis will be able to perform defibrillation more easily by using the Auto-D mode.

In the debriefing, participants' perceptions of Auto-D mode resembled ambivalence. On one hand, there were concerns about the possibility of rhythm analysis errors and the longer time required for analysis. On the other hand, participants also expected that the defibrillator's rhythm analysis would be more accurate than their own and that, as a result, the time needed for defibrillation could be reduced. Among the participants, nurses who were confident in ECG interpretation and defibrillation decisions likely felt that the Auto-D mode took too much time to make decisions, while those who lacked confidence felt the opposite. Based on the simulation results of this study, concerns regarding potential errors or time delays associated with Auto-D mode can be attributed to a lack of familiarity with the new protocol.

Nevertheless, the most noteworthy implication is the potential alleviation of psychological burden. Although nurses often report confidence in performing defibrillation itself, they frequently lack certainty about the accuracy of their ECG rhythm interpretation (Yun & Kim, 2020). Consequently, Auto-D mode can reduce nurses’ psychological burden associated with rhythm analysis, thereby decreasing hesitation during defibrillation. Ultimately, this could shorten the time to first defibrillation when non-expert CPR providers respond during the early phase of IHCA (Hui et al., 2011; Vaillancourt et al., 2024).

The results of previous studies on the advantages of Auto-D mode, as well as the findings of this study, have important implications for CPR education targeting nurses. In Korea, most accredited CPR-related education programs for HCPs are the Korean BLS (KBLS) and KALS courses offered by KACPR, and the BLS and ACLS courses offered by the AHA. In the KBLS and BLS courses, only AEDs are used for training, while in the KALS and ACLS courses, only manual defibrillators are used. Auto-D mode is not utilized in any of these training programs. There may be hospitals that educate the use of Auto-D mode in their own internally developed CPR training courses, but this could not be confirmed in the literature. If simulation-based education, similar to the KALS and ACLS courses, where trainees directly analyze ECGs and perform manual defibrillation, were provided more frequently and regularly, it could help improve overall CPR response competency. However, simulation training is limited by the number of participants, as well as issues such as time and cost, making it difficult to offer frequently. Therefore, considering the number of training sessions and the costs that can be provided, if protocols incorporating Auto-D mode are integrated into existing CPR training courses, there is a possibility that actual defibrillation time could be reduced even with limited resources (Vaillancourt et al., 2024).

To bring about such changes in training courses, it is first necessary to revise the guidelines or recommendations. The 2025 Korean Guidelines for CPR and ECC recommended considering the installation of AEDs in places where manual defibrillators are not frequently used or where HCPs lack ECG interpretation skills. However, no mention is made of using the Auto-D mode of existing manual defibrillators (Korea Disease Control and Prevention Agency & Korean Association of Cardiopulmonary Resuscitation, 2026). While it is reasonable to install AEDs in outpatient clinics or diagnostic departments where manual defibrillators are not available, as mentioned in the introduction of this study, most hospital wards are already equipped with manual defibrillators, and installing additional AEDs would increase costs and management burdens. Therefore, in areas of the hospital where manual defibrillators are already available, it would be a more rational strategy to actively consider the use of Auto-D mode. Thus, if CPR guidelines were to recommend the use of Auto-D mode in this context, it could serve as an important starting point for changes in education and practice.

This study yields important implications from both practical and academic perspectives. Although prior studies and the present findings consistently demonstrate the advantages of Auto-D mode, its implementation in real-world IHCA responses may introduce complex challenges, including team communication and leadership dynamics (Stærk et al., 2022-a). Prior to full adoption of Auto-D mode in IHCA protocols, potential barriers should be proactively identified and addressed through approaches such as in-situ simulation training conducted in the actual hospital work environment. Such preparation would maximize the benefits of Auto-D mode implementation. Moreover, given that the goal of adopting Auto-D mode is to improve IHCA patient survival rates, ongoing measurement and comparison of patient outcomes before and after implementation are essential to verify its effectiveness.

Ⅴ. Conclusions

This study compared the accuracy of defibrillation decision and the time required to deliver defibrillation using the Auto-D mode of manual defibrillators versus traditional manual defibrillation. The results demonstrated that Auto-D mode achieved reliably high accuracy in determining the need for defibrillation, and defibrillation time did not increase; on the contrary, it was slightly shorter. Additionally, although not statistically significant, the hands-off time with Auto-D mode tended to be shorter in PEA and two VF scenarios.

These findings suggest that Auto-D mode represents a cost-effective and practical strategy for optimizing the initial response to IHCA and may contribute to enhancing patient survival outcomes. As a foundational study, this work provides evidence to support the introduction of Auto-D mode into domestic hospital setting and establishes its potential role in nursing care during the early phase of IHCA response.

This study has several limitations that should be considered in its interpretation, and future research is recommended based on these findings.

First, this study enrolled only participants who had completed the KALS or ACLS course. Therefore, the results may differ if nurses or other HCPs without such training were to use Auto-D mode. Future research should include HCPs who have not completed an ALS course in simulation-based studies, which would allow for greater generalization of the findings regarding Auto-D mode utility.

Second, the results of this study were obtained in a highly controlled simulation environment. Consequently, direct application of these findings to real-world IHCA response, in which multiple factors influence the process and patient outcomes, warrants caution. Therefore, research is needed to evaluate the utility or limitations of Auto-D mode in actual IHCA response, either through in-situ simulation training or by implementing Auto-D mode in real CPR situations and assessing the outcomes.