The interaction between gender, health and smartwatch use on health anxiety

Abstract

Background

This study investigates the relationship between gender, smartwatch use, diagnosed health conditions, and health anxiety. Previous research has shown health anxiety is more common among women and those with chronic conditions, but little is known about how these factors interact with the use of wearable health technology.

Methods

This study employed a cross-sectional design with 259 participants, controlling for age and trait anxiety levels.

Results

A three-way ANCOVA revealed a significant interaction between gender, smartwatch use, diagnosed health conditions and health anxiety. Post-hoc analysis suggested that smartwatch use is potentially associated with higher health anxiety scores for men with health conditions and this pattern appears to reverse for men without health conditions. In contrast, smartwatch use is potentially associated with lower health anxiety scores for women with health conditions and this pattern appears to reverse for women without health conditions.

Conclusions

These findings suggest the need for personalised approaches to wearable health technology that consider gender differences and the potential psychological associations. Future research should explore how wearables impact health anxiety across other populations and examine whether different types of smartwatches have distinct effects on health anxiety.

Background

Health anxiety is a relatively new concept, evolving from what was traditionally known as hypochondria [1]. It is characterised by an excessive and often irrational fear of having or developing a serious illness [2]. This condition typically manifests in a cyclical pattern, beginning with checking the body for any signs of illness [3, 4]. In health anxious individuals, this behaviour is often followed by seeking reassurance from medical professionals or through the internet, which paradoxically tends to increase anxiety rather than alleviate it [5, 6]. This heightened anxiety often triggers more frequent and obsessive body checks [1], which, in turn, perpetuates the cycle of health anxiety [7].

Researchers have speculated that the COVID-19 pandemic has further amplified health anxiety [8]. The fear of infection and heightened awareness of health risks has led to an increase in health anxiety-related behaviours, such as excessive hygiene practices and extreme social isolation. While these behaviours have been protective, for some they have become maladaptive, leading to the development of health anxiety [9]. Consequently, the prevalence of clinical levels of health anxiety is rising, particularly among young children and middle-aged adults [10, 11].

Clinical levels of health anxiety can lead individuals to undergo unnecessary medical consultations, examinations and tests, which places undue strain on the healthcare system [12]. Additionally, health anxiety can significantly impact an individual’s mental and physical well-being [13]. If this necessitates treatment, it places further strain on healthcare resources [5].

Individuals with health conditions are more likely to experience health anxiety, particularly since the context of the COVID-19 pandemic. Heinen, Varghese [9] found that both physical and mental health conditions were significant predictors of higher scores on the Short Health Anxiety Inventory (SHAI) [14] when controlling for demographic variables such as gender. Jeffers, Benotsch [15] found that individuals who misuse prescription medications are more likely to have both a medical condition and health anxiety, indicating some relationship between health conditions and health anxiety. A systematic review by Storer, Holden [16] identified health anxiety as the most prevalent form of anxiety among gastroenterology and hepatology outpatients, surpassing generalised and social anxiety. Similarly, women with early-stage breast cancer have been shown to experience elevated levels of health anxiety following diagnosis and treatment [17]. In a study focusing specifically on COVID-related anxiety, King, McQuaid [18] found that while individuals with at-risk medical conditions did not report significantly higher levels of COVID-related anxiety, they did report poorer quality of life and more extreme protective behaviours such as staying home excessively and compulsively disinfecting. Taken together, the literature indicates a bidirectional and multifaceted relationship between health conditions and health anxiety, where the presence of health conditions may not always heighten anxiety levels directly, but does appear to shape how that anxiety is experienced and expressed, warranting further exploration.

A part of the broader category of consumer health wearables, smartwatches have become increasingly popular in recent years. These devices offer a range of health monitoring capabilities, including heart rate tracking, sleep analysis, and activity logging [19]. By providing real-time data, the technology has revolutionised health management as users can use it to make informed decisions, such as increasing physical activity or adjusting sleep patterns, to enhance overall well-being [20]. With the potential to lower healthcare costs, smartwatches are being investigated and trailed as replacements for traditional, in-person healthcare visits [21]. Despite the positive impact smartwatches may have on helping patients and healthcare professionals to monitor health and healthy behaviour, there is increasing concern that these devices may contribute to the development of health anxiety [22].

Interestingly, regulatory bodies, such as the American Food and Drug Administration (FDA), have recognised the potential risks associated with some functions of wearables [23]. For example, Dhruva, Shah [24] discussed the FDA’s clearance of Apple Watch features for detecting irregular heart rhythms and generating electrocardiograms (ECGs). The study identified several risks, including the potential for misinterpretation and over-reliance on the devices. These risks highlight the importance of considering the psychological impacts of smartwatch use as the potential for misinterpretation could cause or exacerbate health anxiety.

Demonstrating this risk, Rosman, Gehi and Lampert [25] conducted a case study following a patient diagnosed with atrial fibrillation, who purchased a commercially available smartwatch for cardiac monitoring. Certain notifications from the smartwatch, such as elevations in heart rate during exercise and inconclusive ECGs, were misinterpreted as indicators of the patient’s health condition worsening. This led to a significant increase in the number of ECGs performed over time, suggesting that the notifications were wrongly perceived as health threats. Rosman and colleagues noted that this misinterpretation fuelled the continuous cycle of excessive worry and compulsive body checking behaviours, ultimately leading to increased health anxiety.

As a further example of the potential for smartwatches to increase health anxiety, Filippaios, Tran [26] explored the psychological effects of receiving health alerts among poststroke adults. Their analysis revealed a significant reduction in self-rated physical health status among users who received alerts. Qualitative research has also found evidence of health anxiety that may have been exacerbated by wearable activity tracker use. Andersen, Langstrup and Lomborg [27], who explored wearable activity trackers in patients with chronic heart disease, found mixed experiences among participants. While some gained greater insight into their health and made positive health-related improvements, others experienced new health-related anxieties, such as doubts about their health and feelings of failure in monitoring it. It is important to acknowledge that the evidence for smartwatch-related increases in health anxiety is not equivocal. Echoing the positive experiences noted by Andersen, Langstrup and Lomborg [27], reported no significant difference in anxiety levels between the smartwatch and control groups. Although this study examined general anxiety, the health-focused context of their study allows its findings to be relevant and extendable to the realm of health anxiety. The evidence suggests that wearable health technologies, though helpful for some, may contribute to or worsen health anxiety in others, particularly when users lack medical literacy or support in interpreting the data they receive.

Beyond the influence of underlying health conditions, evidence points to gender as another key factor shaping experiences of health anxiety, with women consistently showing greater vulnerability. It is reported that women are more likely to develop anxiety disorders in general (McLean & Anderson, 2009), and this trend extends to health anxiety specifically. During the COVID-19 pandemic, women reported significantly higher health anxiety than men, with gender emerging as a significant predictor in regression analyses [28]. Similarly, Liu, Zhang [29] found that women experienced higher levels of post-traumatic stress symptoms during the pandemic, including hyperarousal and intrusive thoughts, features that closely align with cognitive and physiological aspects of health anxiety.

It is possible that women’s greater experiences of health anxiety stems from how they perceive and interpret bodily sensations. Grabauskaite, Baranauskas and Griskova-Bulanova [30] found that women report greater interoceptive awareness, including noticing and emotionally responding to internal sensations. However, they also tend to have lower trust in these sensations and are more likely to worry more about them in comparison to men, which are patterns closely aligned with the core features of health anxiety. This aligns with findings from Heinen, Varghese [9], who also noted greater sensitivity to bodily symptoms among women. Although O’Bryan and McLeish [31], did not identify gender as a direct predictor of health anxiety, they observed that women scored higher in anxiety sensitivity and negative affect, both of which are associated with elevated health anxiety through indirect pathways. Moreover, cultural and psychological factors, such as gender role socialisation and reinforcement of somatic concern in girls may further contribute to increased illness-related worry in women [32]. Taken together, the literature supports the notion that women are more likely to experience higher levels of health anxiety than men, and that this difference may be shaped by a complex interaction of biological, cognitive, and sociocultural factors.

The advent of wearable activity trackers and smartwatches has seen them employed by researchers and clinicians to help motivate and monitor patients’ physical activity. However, there has been very little research investigating the maladaptive effects of smartwatch use, particularly in clinical populations. The primary aim of this research is to investigate whether smartwatch use can increase health anxiety when comparing both clinical and non-clinical populations. Furthermore, as there are known gender differences in health anxiety between males and females, this study will also investigate gender differences in health anxiety. Given the aims of the present research, the following predictions were made:

• H1

Previous research has documented that real-time health data is sometimes inaccurate and can exacerbate anxiety by inflicting worry and perceptions of ill-health. It is therefore predicted that smartwatch use will be positively associated with health anxiety scores, such that users report higher scores than non-users.

• H2

The cognitive behavioural model of health anxiety posits that individuals with chronic illnesses are more likely to focus on potential health threats. Therefore, it is predicted that a relationship exists between health status and anxiety, where participants with diagnosed conditions report higher anxiety scores than those without.

• H3

Previous research has demonstrated consistent evidence showing that women are generally more vigilant about health-related risks and are more prone to anxiety than men. It is therefore predicted that female participants will report higher health anxiety scores compared to male participants.

• H4

It is expected that the combination of these factors will reveal nuanced differences in anxiety levels, reflecting the complex interplay highlighted in the literature. It is predicted that there will be a significant interaction between smartwatch use, health condition, and gender in relation to health anxiety scores. Due to the exploratory nature of this prediction, we did not predict the anticipated direction of these effects.

Method

Design

The study employed a correlational, between-participant cross-sectional design, analysed using a 2 × 2 × 2 ANCOVA. The design involved three independent variables and one dependent variable. The first independent variable, ‘smartwatch use’, had two levels: ‘smartwatch user’ and ‘non-user’. The second independent variable, ‘health condition’, had two levels: ‘diagnosed’ and ‘no condition’. The third independent variable, ‘gender’, had two levels: ‘male’ and ‘female’. The dependent variable measured was health anxiety. Age and trait anxiety were included as covariates to control for their potential confounding effects. Participants self-selected into the conditions based on their current use of smartwatches, if they had any diagnosed health conditions and whether they identified as male or female.

Participants

Criteria required participants to be aged 18 years or older, identify as either male or female and be capable of understanding and completing the study materials in English. Participants were recruited using convenience and volunteer sampling methods through the SONA system, social media platforms, and word-of-mouth referrals. All participants engaged in the study online.

A total of 278 people responded to the survey. Before analysis, data were screened for quality (e.g., duplicates, straightlining) and completeness. This resulted in the removal of 15 participants who failed to complete all the primary measures. An additional two responses were removed as they were identified as duplicates. Finally, two participants who did not identify as either male or female were excluded. This resulted in a final sample of 259 participants. The mean age of the participants was 35.86 years (SD = 15.56), with a mode of 23 years. To treat age as a continuous variable, midpoints of the age categories were assigned to each participant. The gender distribution consisted of 70 males and 188 females. Participants were predominantly of white ethnic background (n = 194 participants, 77.6%), with the remaining being Asian (n = 26, 10%), Black (n = 5, 1.9%), Mixed (n = 8, 3.1%) and Other (n = 14, 5.5), with 14 participants preferring not to say. This distribution reflects a similar level of ethnic diversity as the 2021 UK census (Office for National Statistics, 2022) and can therefore be seen as representative.

Materials

The study utilised an online survey administered via Qualtrics, which comprised both a researcher-designed section and standardised psychometric scales to assess various constructs.

Participant and demographic information

Demographic information was collected from participants, including age and gender. At the end of the survey, participants were asked if they owned a smartwatch and if they had a diagnosed health condition. These questions were asked at the end of the survey to reduce the potential for response bias. Participants reported cardiovascular conditions (n = 23), respiratory conditions (n = 20), metabolic conditions (n = 10), musculoskeletal conditions (n = 22), neurological conditions (n = 16), mental health conditions (n = 37), gastrointestinal conditions (n = 27), autoimmune conditions (n = 6), endocrine disorders (n = 23), skin conditions (n = 16), chronic pain (n = 3 ) and cancer (n = 9).

An a priori power analysis was conducted using GPower 3.1 [33] for a three-way ANCOVA with two covariates, assuming a small-to-medium effect size (η² = 0.05), α = 0.05, and power (1–β) = 0.80. The analysis indicated that a sample size of n = 250 would be required. The obtained sample (n = 259) was therefore adequate in meeting this requirement.

Behavioural inhibition system scale

The Behavioural Inhibition System (BIS; Carver and White [34], was used to assess trait anxiety levels. The BIS is a measure of anxiety that assesses an individual’s sensitivity to potential punishment cues, specifically designed to measure anxiety related to sensitivity to potential punishment, which is a distinct approach compared to other general anxiety measures. Therefore, it was chosen for this study as it facilitates the exploration of anxiety in the context of fear of negative outcomes. Participants respond to items on a 4-point Likert scale that ranges from 1 (strongly agree) to 4 (strongly disagree). Participants are required to rate statements such as, ‘I worry about making mistakes’. Lower total scores indicate greater sensitivity to punishment cues, which correlates with higher levels of anxiety. This scale has been shown to have good reliability (α = 0.84) [35] and high construct reliability (α > 0.80) [36], suggesting that the scale reliably measures anxiety.

Short health anxiety inventory

Health anxiety was measured using the SHAI [14], an abbreviated version of the Health Anxiety Inventory [37], that screens for general health anxiety in the past six months. Participants respond to each item by indicating which of four statements best describes them (e.g., 0 = ‘I do not worry about my health’ to 3 = ‘I spend most of my time worrying about my health’). The scale’s instructions specify that participants can select more than one statement if they identify with more than one. If participants select more than one statement, the highest selected was be taken as their response. Scores are summed for a total composite score ranging between 0 and 42, with higher scores indicating a higher level of health anxiety. It is common for studies to define prevalence of clinical health anxiety by scoring greater than or equal to 18 [4, 9]. This cut off has been shown to reliably identify individuals with clinically significant levels of health anxiety [37]. Low levels of health anxiety are defined as scoring 10 or below [38]. The scale has been shown to have good internal reliability in both non-clinical samples (α = 0.89) [14] and for patients with existing medical conditions (α = 0.84) [39]. Diamond, Dysch and Daniels [4] found this same level of internal reliability (α = 0.87).

Procedure

The study was conducted online, meaning the participants could complete the survey at their convenience, while maintaining consistent methodological standards. Upon expressing interest, potential participants were directed to a secure online platform hosted by Qualtrics. The first page of the survey provided detailed information about the study, including its purpose, the nature of the tasks, potential risks, benefits of participating, and confidentiality measures. Participants were required to provide electronic informed consent before proceeding with the survey. They were informed that participation was voluntary, they could withdraw at any time without penalty, and all responses would be anonymised to protect their privacy.

Once consent was obtained, participants completed the survey, which included the demographic information section and the BIS and SHAI scales. Instructions were provided at the start of each section to ensure participants understood how to respond to each question. The survey was designed to be intuitive and user-friendly to encourage complete and accurate responses. Upon completing the survey, participants were directed to a debriefing page which provided additional resources about mental health. Participants were thanked for their contribution and provided with contact information should they have any questions about the study, wish to withdraw or to receive information about the results of the study.

This study was performed in line with the principles of the Declaration of Helsinki. The study protocol was approved by Nottingham Trent University’s ethics committee prior to initiation. The study protocol was approved by Nottingham Trent University’s ethics committee prior to initiation. All data were collected and stored in accordance with the university’s guidelines to ensure confidentiality and security of participant information. The dataset supporting the conclusions of this article is available in the Figshare repository, [10.6084/m9.figshare.30196978].

Results

Preliminary analysis and assumptions

The internal consistency of the scales used was assessed using Cronbach’s alpha. The BIS scale had a Cronbach’s alpha of 0.734 (M = 14.14, SD = 3.14) and the SHAI had a Cronbach’s alpha of 0.867 (M = 13.47, SD = 5.86). These values confirm the internal consistency of these measures in this study.

The assumption of normality for the residuals of the dependent variable (SHAI scores) was evaluated. Skewness and kurtosis were normal across groups when examining both the descriptive statistics and visual examination of the normal distribution. The assumption of homogeneity of variance was evaluated using Levene’s test. The results indicated that the assumption was met, as evidenced by a non-significant result (F(7, 251) = 2.02, p = .053). This suggests that the variances of SHAI scores are equal across the groups, fulfilling the requirement for conducting ANCOVA. The assumption of linearity was evaluated to ensure a linear relationship between the dependent variable (SHAI scores) and the covariates (age, and trait anxiety (BIS scores)). The visual spread in a boxplot indicate that the linearity assumption was adequately met, meeting the assumptions for ANCOVA.

The assumption of homogeneity of regression slopes was tested by including the interaction terms between each covariate (age and BIS scores) and each independent variable (gender, smartwatch ownership, diagnosed health condition) in the ANCOVA model. The interaction terms were non-significant, all p > .05 (e.g., Gender × Age2: F(1, 247) = 0.33, p = .567), indicating that the relationship between the covariates and SHAI scores did not differ significantly across groups. Thus, the assumption was met, supporting the appropriateness of conducting ANCOVA.

Descriptive statistics

Table 1 presents the demographic distribution of participants across the eight experimental conditions as well as the mean SHAI score, where higher scores indicate higher levels of health anxiety. It is important to note that due to the larger proportion of female participants in the study, the groups differed in size.

Table 1.

Descriptive statistics of participants by group

Group	n	Age (M)	SHAI (M, SD)	Min	Max	Range
1: Smartwatch user, Diagnosed, Male	13	39	15.46(6.90)	2	27	25
2: Smartwatch user, Diagnosed, Female	44	37.20	15.43(5.54)	5	31	26
3: Smartwatch user, No condition, Male	21	35.43	8.62(4.43)	0	18	18
4: Smartwatch user, No condition, Female	53	32.08	14.85(7.74)	1	36	35
5: Non-user, Diagnosed, Male	21	44.19	11.95(5.23)	2	23	21
6: Non-user, Diagnosed, Female	37	39.24	16.03(6.36)	2	30	28
7: Non-user, No condition, Male	16	39.13	10.31(3.09)	5	16	11
8: Non-user, No condition, Female	54	31.37	12.50(5.97)	0	29	29

n  number of participants, SD Standard deviation,min minimum score,max maximum score

Main analysis

An ANCOVA was conducted to examine differences in health anxiety (SHAI scores) based on smartwatch ownership, diagnosed health condition, and gender, while controlling for age and trait anxiety (BIS scores).

Age was a significant covariate (F(1, 249) = 12.567, p < .001, η2 = 0.048), indicating a significant relationship with health anxiety scores. Descriptive statistics showed that younger participants reported higher health anxiety compared to older participants. Trait anxiety was also significant (F(1, 249) = 34.168, p < .001, η2 = 0.121). A Pearson correlation confirmed a significant negative association between trait anxiety (BIS) and health anxiety (SHAI) (r(259) = − 0.429, p < .001). This indicates that higher trait anxiety (represented by lower BIS scores) co-occurred with higher health anxiety (higher SHAI scores). Therefore, trait anxiety was also retained as a covariate.

• H1

The first hypothesis predicted that smartwatch use would be positively associated with health anxiety scores. However, no significant relationship was found between smartwatch users and non-users (F(1, 249) = 0.891, p = .346, η² = .004).

• H2

The second hypothesis predicted a relationship between health status and anxiety, with those having diagnosed conditions reporting higher anxiety scores. Supporting this prediction, a significant main effect of health condition was observed (F(1, 249) = 19.951, p < .001, η² = .074), indicating that individuals with a diagnosed health condition reported significantly higher levels of health anxiety than those without.

• H3

The third hypothesis predicted that female participants would report higher health anxiety scores than male participants. However, the difference between gender groups was non-significant (F(1, 249) = 2.131, p = .146, η² = .008).

Interaction effects

The final hypothesis predicted a significant interaction between smartwatch use, health condition, and gender in relation to health anxiety scores. In support of this, a significant three-way interaction was observed (F(1, 249) = 6.396, p = .012, η² = 0.025), indicating that the relationship between these factors and health anxiety varied significantly across groups. The interaction between smartwatch use and gender was non-significant (F(1, 249) = 0.000, p = .983, η² = 0.000). Similarly, the interaction between gender and health condition was non-significant (F(1, 249) = 1.970, p = .162, η² = 0.008). The interaction between smartwatch use and health condition also did not reach significance (F(1, 249) = 0.056, p = .813, η² = 0.000).

As the two-way interactions did not elucidate what was driving the three-way interaction, further analyses were conducted to determine if conditional two-way effects were present within levels of the third variable, health condition. This relationship is illustrated in Figs. 1 and 2. While these did not reach the threshold for statistical significance, they provided preliminary insights into the group patterns. For participants with a diagnosed health condition (Fig. 1), the smartwatch use × gender interaction trended towards, but did not reach, statistical significance (F(1, 109) = 2.91, p = .091, η² = 0.026). However, descriptive patterns suggested that within this group, males using smartwatches and females not using smartwatches reported higher health anxiety scores. For participants without a diagnosed health condition (Fig. 2), the smartwatch use × gender interaction also trended towards, but was not, statistically significant (F(1, 138) = 3.47, p = .065, η² = 0.025). In this group, the observed trend was reversed, with males using smartwatches reporting lower health anxiety scores, and females using smartwatches reporting higher health anxiety scores. These differing trends across the groups help to explain the significant three-way interaction observed in the overall model.

Fig. 1.

The interaction between gender and smartwatch use for individuals without a health condition

Fig. 2.

The interaction between gender and smartwatch use for individuals with a health condition

Discussion

The aim of the present research was to examine associations between health condition, gender, smartwatch use and health anxiety. In line with current knowledge, this study found that individuals with diagnosed health conditions reported significantly higher levels of health anxiety compared to those without such conditions [1, 11, 40]. The hypothesis that gender and smartwatch use would be independently associated with health anxiety was not supported. However, post hoc analysis suggested an interaction between smartwatch use and gender on health anxiety that is dependent on health condition.

The prediction that health anxiety scores would be higher in individuals with health conditions was supported. This finding is consistent with a growing body of research demonstrating a clear link between health conditions and elevated health anxiety [9, 15, 16]. Notably, this result challenges the findings of King, McQuaid [18], who reported no significant difference in COVID-related anxiety between individuals with and without at-risk health conditions. One possible explanation for this discrepancy is that King et al. focused specifically on anxiety related to COVID-19, which may be influenced by additional factors such as perceived risk or public health messaging, rather than generalised health anxiety. In contrast, the present study measured health anxiety more broadly, offering a more comprehensive understanding of the relationship between living with a health condition and health-related fears. Therefore, while King et al. highlighted differences in behavioural outcomes and quality of life, the current findings suggest that the presence of a health condition is more closely associated with heightened health anxiety than their results indicated. When considered together, the findings from both the current study and King et al. suggest that health conditions are linked to higher health anxiety and may be related to how it is expressed, with the nature of these depending on the individual’s context and the nature of the health threat.

A novel finding revealed that there was a significant three-way interaction between health condition, smartwatch use and gender. To explore this relationship, post-hoc tests were conducted to examine which interactions were driving this effect. The absence of significant two-way interactions suggests that no consistent pairwise relationships were present across the full sample. Instead, the significant three-way interaction indicates that these relationships are conditional and only emerge when considering smartwatch use, gender, and health condition simultaneously. The absence of significant two-way interactions can be considered consistent with the current literature, which implies that the relationships between smartwatch use, health condition, and health anxiety are not straightforward. For example, previous research has shown mixed effects of smartwatch use on anxiety, with some studies reporting increases and others finding no significant differences (25–27, Paul et al., 2023). Similarly, while health conditions are often associated with greater health anxiety, this relationship is not consistently observed [9, 15–18]. Together, this variability supports the present finding that no single pairwise relationship was sufficient to explain health anxiety outcomes.

Examining whether conditional two-way effects (between smartwatch use and gender) were present within levels of the third variable (health condition) was non-significant, however it did reach the level of a trend. The relationship suggests a potential framework for future research. Specifically, the data point towards a hypothesis that within the health-condition group, men using smartwatches tend to report higher health anxiety scores than women. In contrast, among those not using smartwatches, women tended to report higher health anxiety scores than men. These observed tendencies reflect the mixed results Andersen [27], who found that while some individuals utilise smartwatches without reporting associated health anxiety, others display heightened anxiety scores in relation to device use. To interpret these distinct patterns, we draw on Lupton et al.’s [41] concept of ‘data sense,’ which posits that engaging with sensor data is not merely a cognitive task, but a deeply embodied process. Data sense requires users to actively negotiate between digital metrics and their own physical, sensory, and emotional experiences. When dealing with discrepancies between digital data and bodily sensations, users must work hard at this sense-making process, which can trigger strong affective responses such as anxiety or reassurance.

These observed patterns suggest a hypothesis that for men with health conditions, there may be a unique association between smartwatch use and health anxiety. Given that research has often found that that men are generally less aware of bodily sensations than women [9], they may lack a strong internal sensory baseline, making the ‘data sense’ negotiation inherently more challenging. It could be that men rely more heavily on smartwatch data, which possibly conflicts with or lacks physical context. When such data is interpreted as being linked to a worsening health condition, it could be associated with higher reported health anxiety, particularly as users attempt to reconcile digital metrics with physical sensations. This relationship may be further complicated in instances where device data is inaccurate or unclear [23].

Among females, the observed pattern differed: women with a health condition who did not use a smartwatch reported higher health anxiety, whereas those without a health condition who used a smartwatch displayed higher scores. Overall, these descriptive trends suggest Rosman’s [25] case report may represent an isolated case; the current data indicate that for women with health conditions smartwatch use is not necessarily associated with higher health anxiety scores. Andersen [27] support this notion, finding that some patients used their smartwatches for reassurance about heart function, thus avoiding unnecessary anxiety. In the context of ‘data sense’, this represents a successful integration of digital metrics and embodied knowledge. Because the device validates their existing fleshly senses rather than conflicting with them, the sense-making process is easily resolved, resulting in a reassuring affective response rather than an anxious one.

The findings of Lomborg and Frandsen [42] provide insight into why females without health conditions could have higher health anxiety scores when using smartwatches compared to non-users. Lomborg and Frandsen emphasise the communicative nature of self-tracking, where users often engage with their health data in a social context. This engagement may be associated with higher anxiety through social comparison or the pressure to meet perceived health standards. As research suggests, women are generally more prone to social comparison and have a stronger desire for social acceptance compared to men [43]. As a result, the constant monitoring and sharing of health data through smartwatches could be linked to higher health anxiety in this group, particularly if there is a perceived need to conform to health norms or expectations. In contrast, the relationship between these social dynamics and health anxiety may be less pronounced in men, which could explain why their health anxiety scores may be lower when using smartwatches in the absence of a health condition.

Implications

The study extends current knowledge by surveying the general population, rather than focusing solely on clinical populations. This broader application enhances our understanding of health anxiety in a wider context and makes the findings more applicable to everyday smartwatch users. While the sub-group patterns observed in this study are exploratory, they highlight important avenues for future research regarding the design and implementation of wearable health technology. For example, for men with health conditions, technology providing contextual information to accompany raw data could perhaps be associated with a lower likelihood of misinterpretation. For women, particularly those without health conditions, designers might consider how the data presentation interacts with social pressures. The potential for misinterpretation of health data highlights the need for better education around the use of smartwatches. Users may need guidance on accurate interpretation and measures to prevent over-reliance on the technology. However, there is a need to consider how such personalisation could influence health disparities, particularly if certain groups (e.g., older adults, lower-income individuals) are less able to access or effectively use these tailored technologies. Furthermore, healthcare providers should be aware of the interaction between gender and wearable health technologies in relation to health anxiety. This awareness can guide more personalised recommendations about the appropriateness of smartwatches for individuals and ensuring they are given appropriate instruction regarding the interpretation of health-related data.

Strengths and limitations

The study included 259 participants, which is a robust sample size for the analyses conducted. The power sensitivity analysis confirms that the study had sufficient power to detect significant effects, strengthening the findings. The study also controlled for important covariates (age and trait anxiety), which helps to isolate the effects of the primary independent variables on health anxiety.

Since participants self-reported their health conditions and smartwatch use, there is a possibility of response bias, particularly social desirability bias. The researcher placed the questions most likely to provoke response bias at the end, so although this particular confound is unlikely, it cannot be entirely ruled out. Attention checks during questionnaire completion (e.g., instructed response items) were not used, and this is also a limitation. While the sample was predominantly white, the ethnic composition of our participants is similar to the current UK national average. Consequently, while the findings may be considered representative of and generalisable to the UK population, the generalisability to more globally diverse or non-Western populations is limited. Future research should aim to include broader international participant pools to ensure findings are applicable across a wider range of cultural and socioeconomic contexts. Regarding gender, the sample was skewed toward female participants. While this reflects higher rates of health data engagement often observed in this demographic (Carroll et al., 2017), the imbalance may have limited the statistical power to detect more subtle gender-based associations. Future studies should utilise targeted recruitment to achieve a more equitable gender distribution.

The treatment of health condition as a monolith in this study also limits the nuances that exist within this category. A wide range of self-reported health conditions were self-reported in this study. For the purposes of analysing the data, these were treated as a single category. However, given the range of conditions included in this study, it does limit the degree to which we can understand the impact of a specific health condition. Moreover, the study lacks insight into how individuals may perceive their condition. For example, two patients with the same type of cancer may have very different beliefs about how it is impacting their life and their future. These beliefs are an important mechanism in determining an individual’s health anxiety. Taken together, it is important that the findings of the present research are interpreted with caution.

There are also aspects of the measurement of smartwatch use and how users may choose to interact with the device that warrant consideration. First, we had no control over the type of devices participants used. It may be that some users used fairly simple devices that have limited health data, while others used devices capable of producing more advanced health-related data. As Andersen, Langstrup and Lomborg [27] noted, there is a distinction between clinically validated self-monitoring technologies and the lower-cost sensors available in the consumer market. The current study did not control for or collect data on the specific type of smartwatch used. Consequently, the varying reliability and accuracy of these devices could contribute to different levels of health anxiety. For example, users may place excessive trust in data from less reliable wearables. This highlights the potential for technology type to significantly influence the psychological outcomes associated with health monitoring devices. Therefore, we cannot make claims about how smartwatches may most influence health anxiety.

Finally, it is important to acknowledge the cross-sectional, correlational nature of this study. Because the independent variables (smartwatch use, health condition, and gender) were not manipulated, definitive causal inferences cannot be drawn. While the study identifies a significant association between health condition and health anxiety scores and describes non-significant interaction patterns, it cannot determine causation. Future longitudinal or experimental research is needed to establish the directionality of these relationships.

Future research

The findings underscore the complexity of the relationship between gender, health technology use, and health anxiety, highlighting the need for further research to explore these dynamics in greater depth. Future studies could examine how different types of wearables, such as higher-quality devices that may meet clinical standards versus lower-quality consumer-grade products, influence health anxiety across these diverse groups. It would also be useful to understand if it is specific health metrics recorded by smartwatches that induced health anxiety, or whether this effect is related to having a wide range of health information via the smartwatch. Future research could also utilise qualitative methods to explore the underlying factors contributing to this discrepancy, offering deeper insights into how individuals with and without health conditions engage with smartwatches and how these interactions influence their health anxiety.

Conclusion

This study offers insights into the complex interactions between smartwatch use, health conditions, gender, and health anxiety. Consistent with previous research, individuals with diagnosed health conditions were found to experience significantly higher levels of health anxiety compared to those without such conditions. Notably, smartwatches may alleviate and exacerbate health anxiety, depending on the user’s characteristics and health status.

These findings underscore the importance of considering individual differences when developing and recommending the use of health technologies. The psychological effects of smartwatches are not uniform; they may be influenced by factors such as gender, health status, and the way individuals interact with and interpret health data. This suggests a need for personalised approaches to the use of wearable health technologies, particularly for populations at risk of heightened anxiety.

Abbreviations

FDA

American Food and Drug Administration

ECGs

Electrocardiograms

BIS

The Behavioural Inhibition System (scale)

SHAI

Short Health Anxiety Inventory

Authors’ contributions

A.P wrote the main manuscript. The methods were devised by A.P and R.S, and both authors contributed to the data analysis. Both authors reviewed the manuscript.

Funding

Not applicable.

Data availability

The datasets generated during and/or analysed during the current study are available in the Figshare repository: Steel, Richard (2025). Perry_Steel_HealthAnxiety_Smartwatch. figshare. Dataset. https://doi.org/10.6084/m9.figshare.30196978.v1.

Declarations

Ethics approval and consent to participate

The study protocol was approved by Nottingham Trent University’s ethics committee prior to initiation. Informed consent was obtained from all individual participants included in the study. All data were collected and stored in accordance the university’s guidelines to ensure confidentiality and security of participant information.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.