2.1. Study design and participants

This qualitative study, guided by the TPB, employed semi-structured interviews for data collection conducted between January 1, 2022, and May 20, 2022 [22]. To ensure credibility and rigor, the study adhered to the Consolidated Criteria for Reporting Qualitative Research (COREQ) checklist [23].

A purposive sampling strategy, utilizing snowball sampling through personal networks, was employed to recruit participants, with the aim of maximizing diversity in terms of gender, age, occupation, and education level. The sample size was determined based on the principle of data saturation, reached when interview responses became repetitive and no new themes emerged during data analysis. Consequently, 29 interviewees were included in this study [24]. For the purposes of this study, a lay rescuer was defined as an individual without specialized healthcare training or one whose training was limited to basic life support or the use of a portable defibrillator. Participants were included if they demonstrated normal communication (spoken Mandarin) and comprehension skills. Exclusion criteria encompassed interruptions in the interview process exceeding 10 min, whether due to personal matters, emotional distress, or internet connectivity issues that impeded the interview’s continuation. Moreover, individuals with serious physical or psychological conditions that impaired their ability to perform CPR were excluded from the study. The demographic information of the participants is presented in Table 1, S1 Fig.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Sociology of participant population.

https://doi.org/10.1371/journal.pone.0327439.t001

2.2. Interviews

This study was guided by the TPB. A comprehensive interview guide was developed through an extensive review of existing literature and consultation with experts. Following the initial interviews with the first two lay rescuers, the interview outline was revised and refined through team discussions. The finalized interview outline, depicted in Fig 2, was utilized to structure the subsequent interviews.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Interview outline based on the TPB.

https://doi.org/10.1371/journal.pone.0327439.g002

A total of 29 respondents from diverse age groups, occupations, and regions were recruited through both online and offline purposive sampling methods. Informed consent was obtained from all participants, either in written or electronic form. For minor participants, two-person written consent was acquired, involving both the participant and their guardian, following a phone conversation and an online consent form. Interviews were conducted either in person or online via video, with all interviewers undergoing systematic training. Two researchers were present during each interview, one conducting the interview and the other observing nonverbal cues. Prior to the interview, participants were briefed on the process, assured of its comfort, and given the option to pause or terminate at any time. Upon completion of the interview, participants received a small token of appreciation worth 10 RMB. Each interview lasted approximately 20–30 min.

2.3. Data analysis

The audio data were compiled and summarized within 24 h of each interview. Data analysis was conducted by a team of three researchers: A nursing professor (Z₁) and two postgraduate nursing students (Z₂ and L) trained in systematic qualitative research methods. Researcher Z₂ transcribed the audio recordings into text, which were subsequently reviewed and modified by both Z₂ and L to minimize information loss during the transcription process. The final transcribed texts, together with the original audio files, were imported into Nvivo 10.0, a qualitative research software, for analysis and coding. To maintain data integrity, all raw data were saved and documented for future reference. Initial coding was independently conducted by two researchers using thematic analysis [25], followed by a comparison and discussion of the identified themes. A third researcher subsequently analyzed any controversial aspects and derived conclusions directly from the raw audio recordings.

2.4. Credibility and rigor

To enhance the study’s credibility and rigor, COREQ guidelines [23] were followed. Interviews were conducted in quiet, participant-preferred settings to promote a relaxed and comfortable environment. To establish rapport and foster trust, researchers contacted participants prior to the interviews. During the sessions, researchers actively listened, paraphrased key points, and observed nonverbal cues to gain deeper insights. To ensure data quality, researchers employed a consistent interview guide and recording equipment while engaging in ongoing data analysis throughout the study. Upon completing the analysis, three participants were randomly selected for member checking, during which researchers shared the findings and sought feedback to validate the interpretations. Furthermore, reflective journaling was employed to document the researchers’ coding and analysis processes, as well as the influence of their conceptual perspectives on decision-making, thereby enhancing the study’s trustworthiness.

2.5. Ethical consideration

Informed consent was obtained from all interviewees prior to their participation in this study. To ensure participant confidentiality, all identifying information was anonymized. This study was approved by the Ethics Committee of Wuhan University People’s Hospital (approval number: WDRY2021-K054).

3.1. Attitude

3.2 Perceived norms

3.3. Perceived behavior control

3.4. Actual control

4.1. Integrating emotional and cultural factors into the TPB attitude dimension

Our study underscores the critical need to integrate emotional responses and cultural influences into CPR training frameworks, thereby advancing a culturally informed redefinition of the “attitude” component within the TPB. Traditional CPR programs, while effective in teaching technical skills, systematically neglect emotional barriers such as guilt associated with unsuccessful resuscitation attempts or fear of confronting mortality [31]. These affective reactions—evidenced by participants’ self-reports of guilt, such as concerns about “causing someone to die faster” and death-related anxiety—can significantly disrupt rational decision-making in emergency contexts [32]. This underscores the necessity of training models that integrate both cognitive assessments and emotional dynamics to improve real-world responsiveness.

The imperative to contextualize training extends to cultural dimensions. Cultural values, particularly moral obligations and the intrinsic fulfillment derived from aiding others, profoundly shape individuals’ willingness to perform CPR. These emotions are not isolated psychological reactions, they are deeply shaped by cultural norms. For instance, in Confucian-influenced societies, deeply ingrained death-related taboos exacerbate emotional barriers, as individuals may fear violating cultural norms when interacting with a dying body—a concern poorly addressed by Western-centric TPB frameworks [33]. By aligning training content with culturally resonant motivators—such as framing CPR as a Confucian moral obligation in East Asian contexts or emphasizing communal responsibility in collectivist cultures— programs can amplify lay rescuers’ perceived rewards and mitigate action hesitancy [34]. Notably, while cost-benefit analyses remain part of decision-making, emotional factors such as guilt and fear frequently override purely rational calculations, especially in cultural contexts where death carries stigma or spiritual significance.

To address these gaps, our refined TPB model explicitly incorporates two interrelated mechanisms: (1) The interaction between emotion-laden evaluations (guilt and fear) and cognitive processes (PBC), and (2) the moderating influence of cultural context in shaping these emotional reactions. This dual emphasis enhances both the model’s explanatory power and its cross-cultural applicability. From a practical standpoint, embedding culturally specific values—such as moral obligation in collectivist societies or personal fulfillment in individualist contexts—into CPR advocacy efforts can yield more effective interventions. For example, training programs that integrate traditional narratives emphasizing heroism or community solidarity may counteract death-related taboos, thereby reframing emotional barriers as motivational drivers.

4.2. Contextual risks and dynamic control: Reconstructing TPB’s PBC dimension

Our findings demonstrate that macro-contextual factors, particularly legal disputes and online harassment [35], significantly influence PBC in the context of bystander CPR. In contrast to traditional static views, our results reveal that PBC dynamically fluctuates based on real-time assessments of situational risks. Legal uncertainties stemming from inconsistent interpretations of Good Samaritan laws and lengthy litigation processes undermine rescuers’ confidence [36]. Meanwhile, threats of online blame heighten psychological barriers by associating personal failure with public humiliation [37]. These findings highlight two key mechanisms: (1) Legal-institutional friction, where perceived control is redefined as a calculation of legal risks, and (2) sociotechnical amplification, through which digital platforms magnify reputational risks.

Conversely, the availability of external resources positively moderates both perceptions and behaviors. The presence of automated external defibrillators (AEDs) and telephone-guided CPR (T-CPR) reduces reliance on individual skills, shifting dependency toward systematic external support. AEDs enable even untrained rescuers to confidently perform lifesaving interventions by simplifying the necessary technical skills [38]. Similarly, T-CPR provides immediate professional guidance, significantly reducing uncertainty and bridging perceived skill gaps during emergency situations [39,40]. These resources not only enhance rescuers’ self-efficacy but also integrate individual actions into broader, coordinated emergency response systems. Accordingly, the availability and effective use of these external resources redefine PBC as a dynamic construct shaped through continuous interactions between individuals and support systems.

Importantly, our contextualized model of PBC explicitly integrates micro-level personal agency with macro-level structural enablers, thereby addressing criticisms of the individualistic nature of TPB. This integrated approach effectively mitigates barriers by clearly defining legal responsibilities, simplifying complex procedures, enhancing public awareness, improving legal assistance, and leveraging intelligent big-data strategies [41,42]. Moreover, systematic promotion and deployment of AEDs and T-CPR further strengthen PBC.

4.3. Normative conflicts and multistage intention enactment: Reconstructing TPB’s subjective norm and intention-behavior pathways

Our findings underscore the inherent normative conflicts within the bystander effect, arising from the tension between competing descriptive and injunctive norms. Specifically, descriptive norms, demonstrated by others’ inaction, directly contradict injunctive norms that dictate societal expectations for intervention. This conflict reveals the limitation of the TPB, which assumes normative homogeneity, thereby emphasizing the necessity for a dual-norm framework. Clearly distinguishing between these two types of norms allows for a more precise analysis of how normative conflicts affect bystander decision-making and intentions to perform CPR [43]. From a practical perspective, educational interventions should explicitly convey societal expectations, demonstrate positive bystander behaviors, and strategically leverage social pressures (roles, rewards, or punishments) to reconcile conflicting norms and facilitate CPR performance [44].

However, strong intentions alone are insufficient; practical barriers such as inadequate CPR skills, situational ambiguity, and concerns regarding the patient’s condition often hinder action. These findings suggest that the TPB does not adequately account for the complexities involved in translating intentions into actual bystander rescue behaviors. Therefore, we propose a revised two-stage TPB model that explicitly distinguishes between the processes of intention formation and intention enactment [45]. This clear separation enables the development of targeted interventions: Norm-alignment strategies address normative conflicts during the intention formation stage, while ongoing CPR skills training mitigates practical barriers by improving self-efficacy during intention enactment [46,47]. Accessible CPR training workshops, real-time guidance tools (smartphone apps and hotlines), and targeted public education on managing a range of patient scenarios can significantly alleviate practical barriers.

By separating normative influences from intention-behavior stages, this refined framework effectively identifies precise intervention points, thereby significantly improving the practical outcomes of bystander CPR interventions.

4.4. Practical implications: Translating theory into interventions

We propose a dual-path strategy to improve bystander CPR effectiveness by combining emotionally adaptive training with systemic empowerment. This strategy addresses both emotional barriers and contextual challenges, transforming bystanders from hesitant onlookers into confident responders.

The first key component of this strategy is virtual reality (VR)-based desensitization. By simulating OHCA scenarios in a virtual environment, trainees can practice CPR in a controlled, low-risk, and high-stress setting. This immersive experience helps to mitigate emotional barriers, such as death anxiety and guilt, that often impede real-life performance. Research has consistently demonstrated that VR technology not only enhances trainees’ willingness to perform CPR but also improves reaction times and self-efficacy [48]. Furthermore, it plays a critical role in reducing anxiety and fostering an overall enhancement of psychological resilience in rescuers [49].

Second, systemic empowerment strategies aim to simplify legal protections and improve access to resources, thereby further encouraging bystander involvement. One such strategy is the standardization of Good Samaritan Laws, which eliminates legal uncertainties and encourages more individuals to intervene during emergencies. Furthermore, the deployment of voice-guided AEDs and the provision of online maps of AED locations can lower the skill barrier for untrained bystanders, enabling more individuals to assist effectively. Empirical research has demonstrated that investments in CPR training and AED deployment significantly enhance survival rates. For instance, with a budget of S$1M, investments in CPR training and AEDs resulted in survival rates of 4.03% and 4.44%, respectively [50]. Moreover, platforms such as TikTok and WeChat promote prosocial behaviors by incentivizing participation through gamified rewards, further motivating bystanders to act during emergencies [51].

By aligning emotional readiness with systemic safeguards, our dual-path strategy bridges the intention-behavior gap, offering a culturally adaptable framework that significantly enhances the likelihood of bystander CPR in emergencies. This approach improves emotional preparedness, alleviates external pressures, and optimizes resources and legal protections, thereby contributing to a more responsive and globally effective emergency response system.