Based on recent estimates from 2023, there are 5.3 billion internet users and nearly 5 billion social media users worldwide [1], which equates to nearly two-thirds of the global population. This technology explosion is driving a paradigm shift in health care, facilitating patients’ ability to monitor their health from home, engage with fellow patients via social media, and have video consultations with care providers [2]. Such transitions have only been accelerated in the wake of the SARS-CoV-2 pandemic as health care providers seek novel solutions to enhance clinical care. Data collection via smartwatches and other personal devices offers a timely and relevant option.

Smartwatches are a widely used tool for tracking health metrics, enabling users to monitor health metrics, including heart rate, blood oxygen saturation, and conduct single-lead electrocardiograms [3]. Smartwatches contain a selection of sensors that hold potential for enhancing individual health care [4], collecting clinical trial data, and contributing to broader public health efforts. By 2020, one in 5 Americans had used a smartwatch [5], highlighting their growing acceptance and integration into daily routines.

The continuous use of smartwatches and users’ willingness to share the health data collected are essential to maximizing their benefits. Ongoing use ensures a comprehensive and continuous dataset, crucial for monitoring chronic conditions, detecting acute health changes, and supporting long-term health management, such as the development of large language models, machine learning, and artificial intelligence for both public and individual care services. Various factors influence whether users maintain regular engagement with their smartwatches, with enjoyment and habit formation being particularly impactful [6,7]. Further work has established that several factors, including utilitarian, emotional, and health-related reasons, play significant roles in the continued use of smartwatches [8]. Users tend to view their smartwatches as personal and enjoyable devices rather than strictly medical tools [9].

Digital health interventions have continued to show promise as effective and economical interventions [10,11]. Yet passively collected data from smartwatches remains largely unused and may offer rich data to clinicians, public health authorities, and private industry.

An equally critical aspect is a user’s level of comfort with sharing their health data with health care providers. The sensors embedded in these devices generate valuable and sensitive data [3,12,13] that have the potential to improve health care outcomes by providing practitioners with detailed, longitudinal, and up-to-date health information. The serial data contained within these datasets may hold significant value in terms of medical treatment and understanding population characteristics in the context of public health. To establish such datasets, the user must continuously use their device over time. Comfort levels in sharing this data may vary based on the sharing method, but this has remained largely unexplored. Concerns regarding privacy, data breaches, and trust in data-handling entities play crucial roles for end users in shaping their comfort levels [14,15].

Data sharing in health care has been investigated through several studies [16-19]. However, as technology continues to evolve and play a more dominant role in society, it is reasonable to expect attitudes to change. This highlights the importance of continued work in the field; not only have people become more accustomed to technology, but more data are being collected from personal devices than ever before [16,17]. A study from the mid-2000s examining the attitudes toward sharing of health data indicates that the recipient of the data, the level of anonymity, and the type of information are foundational factors [18].

Patients now hold vast amounts of their own health data stored on personal devices and online services. Historically, before the mainstream adoption of smart devices, most health data were kept by health services as medical records. Smart devices now hold an additional set of longitudinal health data outside of medical records. The impact of this on data sharing preferences is largely unknown.

With the widespread success of smartphones and mobile health apps in the 2010s, the general public began to appreciate the large amount of data they hold, and there has been heightened awareness surrounding data privacy [19]. A 2016 qualitative study on the topic highlighted that users saw the data on calories and exercise collected in apps to be private matters [20]. Among concerns was the potential exploitation of data by third parties (eg, insurance companies) [20].

The concern of data misuse is well-founded. Internet users face a near-continuous bombardment of tailored advertisements [21] and even automobiles that collect driving data, which can be sold to insurance companies [22]. To appreciate the potential benefit of the health data collected from smartwatches in health care, it is crucial to understand both the users’ concerns and the reasons behind them.

This study seeks to examine the key factors that determine continuous smartwatch use and users’ comfort levels in sharing their health data collected from a smartwatch with health care practitioners and public health authorities. By identifying these factors, strategies may be developed to enhance user engagement and data sharing, ultimately improving the integration of smartwatch sensor data into health care practices and public health authorities. As smartwatches are a rapidly evolving technology, it is crucial to re-establish the factors influencing continuous smartwatch use. This paper presents findings from an online survey of smartwatch users, offering insights into the key factors influencing their behaviors and preferences.

By identifying these factors, strategies may be developed to enhance user engagement and data sharing, ultimately improving the integration of smartwatch sensor data into health care practices and public health authorities. As smartwatches are a rapidly evolving technology, it is crucial to re-establish the factors influencing continuous smartwatch use. This paper presents findings from an online survey of smartwatch users, offering insights into the key factors influencing their behaviors and preferences.

This study is a cross-sectional online survey of smartwatch users. The survey was designed to assess the determinants influencing continuous use of smartwatch sensors and users’ comfort levels in sharing health data with health care practitioners and public health authorities. The paper was prepared in line with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines [23] (the STROBE checklist is provided in Checklist 1).

The survey was conducted via the internet using the Qualtrics (Qualtrics LLC) survey platform. Recruitment and data collection were conducted between July and November 2023.

The inclusion criteria required participants to be aged 18 years or older and current or past users of smartwatches. Participation was voluntary, and no incentives were offered. Participants were recruited via convenience sampling through various online platforms, including social media (eg, LinkedIn [LinkedIn Corp]), university networks (University College Cork), and community forums (eg, Reddit [Reddit Inc]).

The survey consisted of 2 main sections. The first section focused on factors influencing continuous use of smartwatches, based on the Expectation-Confirmation Model (ECM), as shown in Figure 1 [6,24]. The ECM is a widely accepted and validated model that has been used in similar studies, focused on continuous intention to use consumer electronics [6,25]. The first section included 2 items measuring confirmation; 3 items for enjoyment and habit formation; 4 items for perceived usefulness, satisfaction, and continuous intention; and 8 items measuring perceived usability. These were measured on a 7-point Likert scale. For example, perceived enjoyment was assessed with 3 questions: (1) “I have fun interacting with the smartwatch,” (2) “Using the smartwatch provides me with a lot of enjoyment,” and (3) “I enjoy using the smartwatch.” For each question, respondents were asked to choose from a 7-point Likert scale ranging from “entirely disagree” to “entirely agree.” All the questions and response options are included in Multimedia Appendix 1. Each question is coded to the construct it is evaluating, corresponding to Tables 1 and 2. Thus, the first section of the survey is a validated questionnaire that has been used in previous studies [6,25]. Continuous intention is defined as the intention of the user to engage in regular, habitual use of the device for the foreseeable future. Perceived usefulness is defined as the user’s perception of the utility and value derived from the possession and use of the device. Generally, constructs refer to user-endorsed feelings, experience, and perceptions regarding habit formation, satisfaction with the device, perceived enjoyment of using the device, perceived usability of the device and its features, and confirmation of these perceptions.

The second section addressed comfort levels in sharing health data, concentrating on different data-sharing methods (eg, internet vs noninternet) and included items assessing preferences for various methods and concerns about privacy and trust. This section measured responses on a 5-point Likert scale ranging from “extremely comfortable or confident” to “extremely uncomfortable or unconfident.” We use the term “comfort level” to refer to the respondent’s self-reported comfort in sharing smartwatch health data as outlined in each particular scenario presented in the survey. Refer to Multimedia Appendix 1 to view the survey in its entirety.

Data were collected using the Qualtrics online survey platform, ensuring anonymity and confidentiality of responses. The survey was open for responses for a period of 3 months, aiming to capture a diverse sample of smartwatch users. An open link was used to distribute the survey. To prevent false responses, tools offered by the survey platform, such as “Bot-Detection” and “reCAPTCHA,” were enabled.

The survey aimed to recruit approximately 200 participants, as recommended in the literature, to provide sufficient power for the model to yield a meaningful analysis to reduce the impact of random error in structural equation modeling (SEM) [26].

Data were analyzed using SmartPLS (version 4.1.1.1; SmartPLS GmbH) and Python (version 3.10; Python Software Foundation). Descriptive statistics, including means, SDs, and frequencies, were used to summarize participant demographics and overall response patterns. Inferential statistical tests were used to examine relationships between variables relevant to continuous smartwatch use.

For the analysis of factors influencing continuous smartwatch use, SEM was conducted using SmartPLS version 4.1.1.1. The SEM analysis was used to assess the reliability and validity of the ECM and test the relationships among the constructs, including habit formation, perceived enjoyment, satisfaction, perceived usefulness, confirmation, perceived usability, and continuance intention. A bootstrapping procedure was applied to assess the significance of path coefficients.

For the analysis of the sharing health data section, Python was used with key packages including “pandas” for data manipulation, “numpy” for numerical operations, “scipy” for statistical testing, and “matplotlib” for data visualization. Wilcoxon signed-rank tests were used to compare paired data, reflecting respondents’ comfort under different sharing scenarios. The first comparison assessed comfort with sharing data with a physician via noninternet methods (eg, wired connection, Bluetooth, and near-field communication [NFC]) versus internet-based methods (eg, web-based cloud services) without anonymization. The second comparison evaluated comfort levels in sharing fully anonymized data with public health authorities versus partially anonymized data intended for public safety purposes. These nonparametric tests were chosen due to the ordinal nature of the survey data and the non-normal distribution of differences between paired responses. A chi-square test of independence was used to evaluate for differences in data sharing comfort between national health authorities, international health authorities, and private companies.

The study protocol [27] underwent peer review and was published in an academic journal. The study involves human participants and received ethical approval from the Social Research Ethics Committee, University College Cork, Ireland (SREC/SOM/21062023/2). Participants provided informed consent to participate in the study, participation was voluntary, and the participants could withdraw at any time. The data collected were anonymized, and no personal identifiers were collected. No compensation or incentives were provided for participation.

A total of 273 survey responses were collected, 198 responses were completed to the end of the continuous use section, and 169 responses were completed in their entirety. Participants identified as 52.7% (144/273) male, 45.8% (125/273) female, and 1.5% (4/273) other. The age of participants ranged from 18 to 65 years, with a mean age of 35.6 (SD 11.7) years.

Table 1 shows that the construct reliability of the measurement model was satisfactory, as composite reliability and Cronbach α coefficients are greater than 0.7 (except for confirmation). Bootstrapping the results showed P<.001 for all the indices in Table 1. Similarly, the model exhibited strong reliability, with all the loadings (in bold) greater than 0.7 except for USAB 1 (usability) and USAB 2. However, USAB 1 and USAB 2 were kept in the model because the loadings were still greater than 0.4.

The average variance extracted for all constructs was greater than 0.5, indicating good convergent validity (Table 1). The discriminant validity of the model was satisfactory, as the square roots of average variance extracted (in bold) are greater than the correlations (Table 1). The higher loadings of the indicators (in bold) compared with the cross-loadings also support the discriminant validity of the model (Table 2).

The path diagram (Figure 2) shows the R² values, which, translated to percentages, show that the model explained 15%, 34%, 50%, 51%, and 62% of the variation in perceived usability, perceived enjoyment, continuance intention, perceived usefulness, and satisfaction, respectively.

The path diagram indicates that 50% of the variation in continuance intention to use a smartwatch was explained through habit (β=.35; P<.001) and satisfaction (β=.38; P<.001). Although other constructs did not show a significant direct influence on continuance intention to use smartwatch, 69% of the variation in satisfaction is explained through perceived usefulness (β=.38; P<.001), perceived enjoyment (β=.32; P<.001), confirmation (β=.24; P<.001), and perceived usability (β=.10; P=.03). Thus, the result indicates that continuous use of smartwatches is explained by habit and satisfaction, which are in turn explained by perceived usefulness, perceived enjoyment, confirmation, and perceived usability. Refer to the questions corresponding to each of these constructs in Multimedia Appendix 1.

The analysis of data-sharing preferences indicated significant differences in comfort levels depending on the method of data transmission and the level of data anonymization. These are subcategorized in detail in the following subsections.

Given the growing body of literature examining smartwatches and their use in health care, we aimed to expand this base knowledge to better understand the determinants of continuous intention and data sharing from the smartwatch user’s perspective. The findings of this study align with previous literature emphasizing the importance of habit formation, satisfaction, enjoyment, and perceived usefulness in driving technology adoption and sustained engagement [6-8]. Previous studies have consistently highlighted that habit formation plays a crucial role in the long-term use of health technologies, reinforcing our results that daily integration and routine use significantly boost engagement. Novel findings in our study include understanding that the use of internet-based data sharing methods can lead to users feeling less comfortable sharing data. This, in addition to the other findings related to data sharing with different levels of organizations, helps fill in gaps in the literature.

This study has limitations. First, the reliance on self-reported data introduces potential biases, including social desirability bias, where participants may overstate their engagement or comfort with data-sharing practices. Second, the convenience sampling approach may not fully capture the broader diversity of smartwatch users, skewing results toward more tech-savvy or health-conscious individuals through self-selection bias. Future studies could use alternative sampling models, such as stratified sampling, to account for this.

The cross-sectional design captures attitudes at a single point in time, limiting insights into how preferences and behaviors might change over time, particularly as technology and privacy practices evolve. Future work could include longitudinal studies to address this.

Further, to the knowledge of the authors, no existing, validated questionnaire collects the specific outcomes deemed pertinent to the data sharing subsection of this study. Development of new questionnaires is a time-intensive and rigorous process and typically requires multiple evaluations to ensure validity and reliability. Drawing on the broad expertise of the research team, the current questionnaire was developed in line with the primary aims of this study. This questionnaire did undergo peer review [27], giving credence to its face validity. However, further psychometric evaluations would be recommended to test the inherent properties of this instrument before its broad uptake.

This study did not differentiate between smartwatch brands, operating systems, or regional data legislation. This may be addressed in future research with a larger sample size or more restrictive inclusion criteria. In addition, as features continue to emerge, particularly artificial intelligence–based features, evaluating their impact on continuous intention and data-sharing preferences is warranted.

The study did not differentiate between various data-sharing scenarios, such as differences in security features or types of health data being shared, which could influence user comfort levels. Future research should further explore these nuances to better understand what drives trust in data sharing through qualitative methodologies such as focus groups or interviews. Of note, there may be geographic variations in preference due to factors such as private health care; these could be investigated in future research. Finally, while the study provides valuable insights into user perceptions, further research is needed to establish the actual clinical impact of smartwatch data integration on health care outcomes.

The relationship between the individual constructs of the ECM and continuous intention and comfort with data sharing was not addressed in our survey. This could be evaluated in future work.

To successfully integrate smartwatch data into clinical care, designing features that support habit-building, enhance user satisfaction, and create enjoyable experiences are necessary for sustained engagement. Furthermore, the preference for noninternet data-sharing methods and fully anonymized data reflects ongoing user concerns about privacy and security, indicating a need for robust and transparent data-sharing practices. Users were more inclined to share data with physicians and health organizations than with private companies, while generally expressing confidence in the potential of smartwatch data to improve health care. By aligning device features and data-sharing protocols with user preferences, manufacturers, health care providers, and policy makers can enhance user engagement and maximize the potential of smartwatches to support individual health management and public health initiatives.