We conducted a cross-sectional web-based survey aiming to investigate the perceptions and attitudes of potential users toward automated cardiac arrest diagnosis using smartwatches. We identified 2 groups of potential users: the first was consumers who already own a smartwatch (the consumer group), as these individuals could instantly make use of this technology as soon as it becomes available. The second was patients at increased risk of experiencing cardiac arrest (the patient group), as these individuals can potentially benefit the most from using this technology. A survey was deployed among both groups as described below, and participants had up to 4 weeks to respond. Data were collected between October 27, 2022, and March 17, 2023.

The study was assessed by the Medical Ethics Review Committee of VU University Medical Center (2022.0544), which declared on November 29, 2022, that the study was not subject to the Medical Research Involving Human Subjects Act (WMO), such that formal approval was not required. Participants received written information about the purpose of the research and consented to the use of their provided answers for research purposes. Participants in the consumer group were compensated with €0.30 to €0.50 (US $0.32 to $0.53) for their participation. Participants in the patient group did not receive any compensation. The data collected from both groups were completely anonymous.

An a priori sample size analysis revealed that a minimum sample size of 385 participants would be needed to attain a margin of error of no more than 5% at a 95% confidence level. To attain an even higher precision while still allowing for dropouts and subgroup analyses, we targeted approximately 1000 participants per user group.

The surveys were developed by WMFvdB and PS in collaboration with the research experts from Markteffect and the NPF. Both surveys underwent a comprehensive review and testing by members of our research group and Markteffect or the NPF, respectively, and the surveys can be found in Multimedia Appendix 1. Both surveys largely comprised the same introductory text and questions, but slight modifications were made to tailor the surveys to each group. The surveys included an introductory text explaining the prevalence of witnessed and unwitnessed OHCAs in the Netherlands, emphasizing the need for an alerting system to ensure prompt intervention in case of cardiac arrest. We explained that we are currently developing a novel technology that could potentially address this issue, as it is capable of automatically diagnosing cardiac arrest using smartwatches. We underscored our interest in obtaining their opinions and perspectives on this technology. If participants agreed to participate by continuing to the online survey, we would proceed to ask questions regarding automated cardiac arrest diagnosis using smartwatches. In addition, we asked questions about gender, age, income, and other demographic characteristics. The questions primarily consisted of closed-ended questions and included 5-point Likert-scale questions to assess their agreement with specific aspects of the technology. Limited open-ended questions were provided to enable supplementary insights. Markteffect hosted and distributed the online survey to the consumer group using Collecthor (Collecthor BV). The online survey sent to the patient group was created and hosted Castor Electronic Data Capture (Ciwit BV). The NPF distributed the survey to the patients who were at increased risk of cardiac arrest and had indicated interest to participate in the survey.

Markteffect cleaned the consumer data set, removing surveys with any of the following characteristics, according to their internal standard operating procedures: (1) speeders—surveys that were completed within an exceptionally implausibly fast timeframe; (2) double respondents—multiple surveys submitted by the same individual; (3) straight liners—participants who consistently selected the same answer choice (eg, always chose a neutral response or always chose the first option) or did not answer the questions; and (4) surveys with consistent unintelligible language. The cleaned data set was provided for further analysis. The data from at-risk patients were cleaned by removing completely empty surveys, speeders, and surveys in which less than 50% of the questions were completed.

Quantitative data were analyzed descriptively. Measures of central tendency, measures of variation, and measures of distribution were calculated [16]. The analyses were performed using R (version 4.2.1) and RStudio (version 2022.2.3.492; R Foundation for Statistical Computing).

A content analysis of the open-ended questions was conducted to provide deeper insights into the participants’ perspectives. WMFvdB created a coding framework consisting of inductive and deductive codes. The coding was reviewed by PS and any disagreements were resolved in consensus. The codes were categorized to identify emerging themes. The content analysis was performed using MAXQDA 2022 (VERBI Software).

In the consumer group and the patient group, 1135 and 1519 participants, respectively, met the inclusion criteria. In both groups, 88.5% (n=1005 and n=1344, respectively) of the participants were included for analysis (Figure 1).

The mean age in the consumer group was 48.6 (SD 13.2) years, and 48.7% (n=489) of the population was male. In the patient group, the mean age was 67.7 (SD 10.0) years, and 56.4% (n=750) of the population was male. Regarding level of education, smartwatch preference, and ethnic background, both groups followed a similar pattern. In the consumer and patient groups, 58.5% (n=588) and 56.1% (n=731), respectively, had a higher level of education. Apple and Samsung smartwatches were most prevalent, and the majority of the population had a Dutch background (Table 1). Both groups had a similar geographical distribution throughout the Netherlands as the general Dutch population (Figure 2).

Both groups expressed interest in the technology; 90.2% (n=906; 95% CI 88.1%-91.9%) of the consumer group and 89% (n=1196; 95% CI 87.3%-90.7%) of the patient group considered the technology as “interesting” or “very interesting” (Figure 3). Moreover, 75.2% (n=756; 95% CI 72.5%-77.9%) of the participants in the consumer group and 77.6% (n=1004; 95% CI 75.3%-79.8%) of the patient group indicated their willingness to use the technology (Figure 4). The most frequently cited reason for abstaining from or expressing uncertainty about adopting the technology in both groups was “not wanting to be resuscitated.” The second most common reason given by the consumers was cultural or religious objections, whereas none of the patients provided this as a reason for expressing uncertainty or abstaining from using the technology. In the patient group, 31.7% (n=92) of participants who had indicated abstaining from or expressed uncertainty about adopting the technology stated not wanting to use a smartwatch as a reason (Figure 5).

In both groups, around 75% (consumer group: n=760; patient group: n=1045) agreed that they would like to use a smartwatch that can be used to detect cardiac arrests, about 72% (consumer group: n=735; patient group: n=964) agreed they would recommend the technology to family and friends, and around 86% (consumer group: n=863; patient group: n=1116) agreed that it should be easy to sign up to use the technology. Both groups agreed (>93%) that the technology should be reliable (consumer group: n=936; patient group: n=1204), and agreed (around 92%) that the technology should be easy to operate (consumer group: n= 923; patient group: n=1223). The consumer group felt a little more strongly that data should be well protected; 91.6% (n=920) agreed compared to 86% (n=1071) in the patient group (Figure 6).

Participants were also surveyed regarding their willingness to make monthly payments for the technology, expressed in euros. The median amount the consumer group and patient group were willing to allocate, given they indicated that they would use the technology, was €5.0 (IQR €2.0‐€10.0) and €10.0 (IQR €5.0‐€20.0), respectively, per month (an exchange rate of €1 = US $1.06054 applied at the time of the study).

The open-ended questions were categorized into the most prevalent themes, as summarized below. Table 2 provides a quantitative overview of the content analysis.

We assessed attitudes and perceptions of potential users toward automated cardiac arrest detection using smartwatch sensor data. We found that the vast majority of the participants expressed their willingness to adopt this innovative technology. However, 1 of 20 individuals indicated that they did not want to adopt this technology, and a considerable number of individuals were yet undecided. Our findings revealed several barriers and concerns that warrant careful consideration.

Previous research assessed the acceptability of the use of wrist-worn wearables, mainly focusing on activity tracking and time spent using the device as a measurement of acceptance [20]. However, to our knowledge, our research is the first to focus on the perceptions of potential users and their willingness to accept automated cardiac arrest detection. Understanding users’ perspectives is imperative for research groups developing such technology, enabling them to address concerns early in the development process. Moreover, health care professionals involved in counseling patients at risk, as well as any stakeholders involved in the implementation, distribution, or marketing of the technology, should have a clear understanding of users’ perspectives. This is paramount in order to increase acceptability [21-23] and to ensure effective implementation [24].

In this context, the insights gained from this study have several important implications. One key factor contributing to reluctance in adopting the technology is that some individuals do not want to be resuscitated in the event of a cardiac arrest. Remarkably, this reluctance was not limited to older people or those with significant comorbidities in the patient group but also extended to the relatively young population in the consumer group. While there are legitimate reasons for refusing resuscitation attempts, it is likely that this reluctance partially stems from a lack of understanding about the prognosis of cardiac arrest, particularly when detected and treated early. Fear of being severely handicapped or incapacitated may have played a role. However, in the Netherlands, approximately 90%‐95% of cardiac arrest survivors are known to survive with a favorable neurologic outcome [25,26]. Education and awareness campaigns may play a pivotal role in addressing misconceptions, ensuring that potential users are well-informed about the life-saving potential of the technology [27].

In the consumer group, cultural or religious objections to adopting the smartwatch-based cardiac arrest technology were frequently reported. Although our survey did not delve into the specific nature of these objections, the evident heterogeneity among potential users highlights the need for customized product development and implementation strategies. Acknowledging and understanding the cultural and religious dynamics that influence technology adoption decisions is critical and should be investigated in more depth in future studies.

In the patient group, a significant barrier to adopting smartwatch-based cardiac arrest detection technology was the reluctance to use a smartwatch. This may partly be attributed to a lack of digital literacy. In particular, older individuals may be less accustomed to digital technology and may perceive a high complexity of operating such devices. Emphasizing simplicity and intuitive operation in the development process is therefore crucial.

Potential users also expressed concerns regarding the potentially high cost of the technology and that financial inaccessibility may exacerbate disparities in health equity [28]. It is worth noting that low socioeconomic status is associated with a higher incidence of cardiac arrest [29], suggesting that this demographic may potentially benefit the most from this technology. Therefore, future research on automatic cardiac arrest detection should include a sufficient number of participants from lower socioeconomic backgrounds. Given that low socioeconomic status is associated with a higher incidence of cardiac arrest, it is crucial to consider cost-effectiveness and affordability early in the development phase. Proactive collaboration with health insurance providers, public social welfare systems, nonprofit health care foundations, and government organizations is essential to ensure financial accessibility.

The analysis also highlighted concerns about privacy, data protection, and data use, which are common concerns when introducing novel medical technologies [30]. Governmental and regulatory bodies have been developing guidelines for software used as a medical device to ensure privacy and data protection [31-34]. For developers and researchers involved in creating such technologies, strict adherence to these regulations plays a vital role in building user trust by ensuring the robust protection of their privacy.

Another concern identified was the reliability of automated cardiac arrest detection systems. Developers are thus challenged to achieve exceptionally high levels of sensitivity and specificity [35]. The goal is to create a system capable of accurately identifying cardiac arrest events while simultaneously minimizing false alarms, which are detrimental to both user trust and system efficacy.

This study features a considerable sample size encompassing a diverse range of participants who might use this technology. The questionnaire was carefully designed using the joint expertise of a leading professional market research organization, a patient federation experienced in researching patient perceptions, and a medical research group with documented experience in survey methodology [36-38].

Our research not only focuses on patients at increased risk of cardiac arrest, but also includes consumers who already own a smartwatch, which positions them to be early adopters as soon as the technology is implemented. It is noteworthy that approximately 50% of individuals experiencing a cardiac arrest have no prior history of cardiac symptoms or events, and OHCA also frequently affects middle-aged adults [39]. This underscores the applicability of this technology across a broad demographic, including those perceived as “healthy.” By encompassing both high-risk patients and regular smartwatch users, our study captures a wide spectrum of perspectives, enhancing the relevance and applicability of our findings.

We acknowledge some limitations. First, a few participants (n=30) seemed to conflate the terms cardiac arrest and heart attack, suggesting potential misunderstandings about what the technology monitors and detects. This confusion may have influenced their responses. However, both heart attack and cardiac arrest are serious medical conditions that benefit from early detection, and the rationale for continuous monitoring should logically extend from one condition to the other. This confusion also underscores the necessity for enhanced public education to improve understanding of these distinct medical events [27].

Second, the research was conducted with patients and consumers in the Dutch population, where local culture and health care systems may shape attitudes toward technology adoption [40]. This aspect has to be considered when extrapolating our results to other regions, with different health care systems or cultures.

The vast majority of potential users expressed a positive attitude toward automated cardiac arrest detection using smartwatch technology. The primary concerns raised by participants included privacy, data protection, reliability, and accessibility of the technology. Despite such concerns, the vast majority indicated that they would be willing to use the technology. This indicates a strong potential user base but also underscores the importance of addressing the identified concerns to optimize acceptance and effectiveness of the technology.