Www.research.manchester.ac.uk



Online Health Research and Health Anxiety: A Systematic Review and Conceptual Integration Richard J. Brown a,b*, Niamh Skellyc and Carolyn A. Chew-Grahamda Centre for New Treatments and Understanding in Mental Health (CeNTrUM), School of Health Sciences, University of Manchesterb Complex Trauma and Resilience Research Unit (CTRU), Greater Manchester Mental Health NHS Foundation Trust, Manchester Academic Health Sciences Centrec Paediatric Psychology Service (CAMHS), NHS Lanarkshire, University Hospital Wishaw, 50 Netherton Street, Wishaw ML2 0DPd Research Institute, Primary Care and Health Sciences, Keele University, Keele, Staffordshire ST5 5BG, United KingdomAccepted for publication in Clinical Psychology: Science and Practice on 19th June 2019* corresponding author: richard.j.brown@manchester.ac.uk; Dr Richard J Brown, Centre for New Treatments and Understanding in Mental Health (CeNTrUM), School of Health Sciences, University of Manchester, 2nd Floor Zochonis Building, Brunswick Street, Manchester, M13 9PL, United KingdomRunning Head: Online Health Research and Health AnxietyAbstract Using the Internet to obtain health information (“online health research”, OHR) is commonplace. This paper provides a systematic narrative review of evidence concerning the relationship between OHR and health anxiety. We conclude that health anxiety is associated with more frequent self-reported OHR, heightened distress after OHR, and increased doctor visits post-OHR. Evidence suggests that OHR often has a reassurance seeking function and can relieve anxiety, but that it can also cause alarm and become a distressing, compulsive behaviour. We present a novel model that integrates these perspectives and existing research within a single explanatory framework that distinguishes between problematic OHR and compulsive OHR and describes the role of positive and negative metacognitions in their respective development. Keywords: cyberchondria; health anxiety; hypochondriasis; Internet; reassurance; online health researchPublic Health StatementsUsing the Internet to obtain health information is commonplace. Such Internet usage involves risks, including the risk of maintaining or exacerbating health anxiety. Health professionals should explore Internet usage with health anxious patients, and consider referral for psychological input where Internet usage appears compulsive or associated with high levels of distress.INTRODUCTIONGlobally, almost 60% of households have Internet access, and 90% of the world’s population live in reach of a mobile broadband network (International Telecommunication Union, 2018). There is an abundance of health information available on the Internet, and using the Internet to obtain health information is commonplace (Fox & Duggan, 2013; Office for National Statistics, 2018). The most visited health-focused website, WebMD, receives an estimated 80,000,000 unique monthly visitors (eBiz MBA, 2019).Many advantages of searching online for health information have been proposed, including anonymity, economy, convenience, interactivity, and speed (Bull, 2011; Starcevic & Berle, 2013; Tanis, 2008). The Internet can also empower patients/health consumers, affording them greater control over their health, and greater input into health care decisions (e.g., Tan & Goonawardene, 2017). In discussions of disadvantages, two issues predominate. Firstly, as Internet access expanded in the 1990s, concerns were repeatedly expressed about the quality of online health information (Cline & Haynes, 2001; Coiera, 1996; Silberg, Lundberg, & Musacchio, 1997). Since then, the majority of evaluation studies have concluded that the quality of online health information is problematic (Zhang, Sun, & Xie, 2015). Secondly, in both the academic (e.g., Harding, Skritskaya, Doherty, & Fallon, 2008; Loos, 2013) and popular press (e.g., Davis, 2018; Fisher, 2006; Mahdawi, 2015; Shanahan, 2016), online health information has been identified as a possible source of anxiety for users. Health anxiety refers to excessive, unwarranted fear, provoked by a perceived health threat (Abramowitz & Braddock, 2008). Bodily symptoms and sensations, and other forms of health information such as test results, may be interpreted as indicators of serious physical illness, leading to distress. Taxometric evidence suggests that health anxiety is ubiquitous, occurring on a continuum (Ferguson, 2009; Longley et al., 2010) that ranges from mild anxiety associated with health-promoting behaviours and appropriate care-seeking to pathological anxiety associated with maladaptive behaviour, impaired functioning, and sustained distress (Asmundson, Abramowitz, Richter, & Whedon 2010). Although health anxiety is not a diagnostic category, it is closely related to historic and current diagnostic entities such as hypochondriasis and illness anxiety disorder (Bailer et al., 2015; Starcevic, 2013). Most empirical studies of health anxiety are based on the cognitive-behavioural model (Warwick & Salkovskis, 1990), which places particular emphasis on dysfunctional beliefs about health and illness. These beliefs are maintained by various cognitive and behavioural factors, including biases in memory (Brown, Kosslyn, Delamater, Fama, & Barsky, 1999) and attention (Owens, Asmundson, Hadjistavropoulos, & Owens, 2004), a ruminative cognitive style (Marcus, Hughes, & Arnau, 2008), and a range of behaviours intended to prevent a feared outcome and/or to reduce subjectively unpleasant anxiety symptoms (Abramowitz & Moore, 2007). Such “safety” behaviours include checking (i.e., examination of the body for signs of illness or symptoms); avoidance of activities that are perceived as risky (e.g., exercise) or that provoke anxiety despite posing no objective health threat (e.g., medical television programmes); and reassurance seeking, which involves seeking confirmation of good health (and/or the absence of health threat) from external sources, such as medical professionals and family members. Experimental findings indicate that health-related safety behaviours are negatively reinforced (Abramowitz & Moore, 2007) and lead to an increase in dysfunctional health beliefs (Olatunji, Etzel, Tomarken, Ciesielski, & Deacon, 2011), thereby maintaining health anxiety. TerminologyThe term health-related Internet use refers to any form of purposeful online activity that is health-related. Although there are different categories of health-related Internet use (HRIU; Eysenbach, 2003), the literature on HRIU and health anxiety is almost exclusively focused on online searches, generally conducted via search engines, for information on symptoms and illnesses. Thus, we employ the term online health research (OHR) in this review. We do not employ the more specific term online illness research because a small number of studies in this area use measures that do not distinguish between searching for symptoms/illness and health searching more generally. We use OHR to mean searching that pertains to one’s own health. We acknowledge the single non-peer reviewed study on ‘cyberchondria by proxy’, or undertaking anxiety-provoking online health searches for another person (Aiken & Kirwan, 2014), as well as the nascent literature on health anxiety by proxy (Thorgaard et al., 2018). However, space considerations mean this review does not extend to such phenomena. Although the term cyberchondriac has sometimes been applied to anyone one who engages in OHR (e.g., Taylor, 2010), cyberchondria is generally considered to be distinct from OHR. Cyberchondria has been defined in myriad ways, including as the emotional distress or exacerbated anxiety resulting from OHR (e.g., Aiken & Kirwan, 2014; Fergus, 2014; Fergus & Dolan, 2014; McElroy & Shevlin, 2014; White & Horvitz, 2013) and as an escalating cycle of distress and engagement in OHR (e.g., Barke, Bleichhardt, Rief, & Doering, 2016; Norr, Albanese, Oglesby, Allan, & Schmidt, 2015; Norr, Oglesby et al., 2015; te Poel, Baumgartner, Hartmann, & Tanis, 2016). McElroy and Shevlin (2014) defined cyberchondria as a multi-dimensional construct encompassing searching that is distressing, repetitive, interrupts other activities, and leads to doctor consultation. The confusion surrounding the term cyberchondria is a barrier to the development of a coherent, comprehensive account of the relationship between OHR and health anxiety. For that reason, we avoid referring to “cyberchondria”, except when discussing studies that have employed the term. Theoretical perspectives on the relationship between health anxiety and OHRThere is a paucity of theoretically rooted studies of OHR (Marton & Choo, 2011), and, to date, the relationship between health anxiety and OHR has received minimal theoretical attention (Starcevic & Berle, 2015). OHR has been described as a form of reassurance seeking (Baumgartner & Hartmann, 2011; Eastin & Guinsler, 2006; Muse, McManus, Leung, Meghreblian, & Williams, 2012; Singh & Brown, 2016; Tanis, Hartmann, & te Poel, 2016). From this cognitive-behavioural perspective, health anxious individuals are motivated to engage in OHR as it is intermittently reinforced by a short-term reduction in anxiety and distress. Like other safety behaviours, such as visiting a doctor to discuss symptoms, OHR has the potential to maintain, and perhaps even increase, health anxiety. Distressing HRIU has also been understood as being related to, but separate from, health anxiety (Fergus, 2014; Fergus & Russell, 2016; Mathes, Norr, Allan, Albanese, & Schmidt, 2018; McElroy & Shevlin, 2014; Norr, Allan, Boffa, Raines, & Schmidt, 2015; Norr, Oglesby, et al., 2015). This literature, which often employs the term cyberchondria, emphasises OHR that has an excessive, and often escalatory nature to it. In this case, OHR may be better understood as a compulsive neutralising behavior, similar to OCD symptomatology, than as a type of reassurance-seeking similar to health anxiety symptomatology (Fergus & Russell, 2016). A related perspective is that OHR might be understood as a behavioural addiction, and possibly as a form of compulsive or problematic Internet use more generally (Fergus & Dolan, 2014; Fergus & Spada, 2017; Fineberg, et al., 2018; Singh & Brown, 2014). Starcevic and colleagues (Starcevic & Berle, 2013, 2015; Starcevic & Aboujaoude, 2015) present a hybrid model in which they suggest that the emotional outcome of searching is a key determinant of whether OHR constitutes reassurance seeking or a distinct and problematic pattern of behaviour in its own right (Figure 1). They characterise Internet searching that results in decreased distress/anxiety as “classical” reassurance seeking. In contrast, searches that result in increased distress/anxiety lead to either avoidance of Internet searching, or repeated excessive, distressing searching, which they term cyberchondria. Starcevic and Berle (2013) suggest that the challenge in understanding cyberchondria is why some individuals continue searching the Internet despite not being reassured by the information they find there. They suggest that particular features of the online environment, namely the quantity, ambiguity, and uncertain trustworthiness of available information, contribute to further searching, alongside individual differences in perfectionism, intolerance of uncertainty, and obsessive-compulsive traits. Their model is silent on whether it represents possible outcomes for different individuals, or for the same individual at different points in time. The latter would imply that cyberchondria is a transient state that may arise in any health anxious individual; the former seems to suggest that cyberchondria is a more enduring tendency that only arises in some health anxious individuals..Regardless, Starcevic and Berle (2013) seem to understand cyberchondria as a pattern of behavior exhibited by health anxious individuals, and occurring within the context of health anxiety rather than as a distinct entity.Theory in this area generally assumes that health anxiety precedes OHR. Scant attention has been paid to the notion that OHR can lead to clinical health anxiety. The primacy of health anxiety has been deemed “more plausible” (Starcevic & Berle, 2015, p. 107). As will be seen, this assumption has not been well tested due to a lack of longitudinal studies. It is possible that the strength of this assumption has been a factor in the heavy reliance on cross-sectional studies, and also in strong assertions about the direction of causality based solely on cross-sectional data. Te Poel et al. (2016) speculate that individuals might engage in OHR, not motivated by health anxiety, but by reasons such as having diet-related questions or seeking information regarding a medication. However, the nature of the information they encounter online may then induce health anxiety. Detailed theoretical and empirical analysis of this causal pathway is notably absent from the literature. ---INSERT FIGURE 1 ABOUT HERE---There have been three previous reviews of the literature in this area (Aiken, Kirwan, Berry & O’Boyle, 2012; McMullan, Berle, Arnáez & Starcevic, 2019; Starcevic & Berle, 2013). Neither Aiken et al. nor Starcevic and Berle employed systematic search strategies, and over twenty relevant papers have been published in the short time since these reviews. McMullan et al. undertook two meta-analyses, the results of which are discussed further below. Our review is more theoretical in its focus, and considers a much broader range of findings. More specifically, this review aims to (a) describe what is known about how the extent, quality, emotional impact and purpose of online health research vary according to individual differences in health anxiety; (b) integrate previous research and theory on this topic within an overarching conceptual framework that suggests new hypotheses for further study; and (c) outline priorities for future work on online health research and health anxiety. METHODSSearch strategyWe sought to identify all empirical studies pertaining to the relationship between online health research and health anxiety, regardless of design or research question, as the basis for a narrative appraisal and synthesis of the available literature. To that end, the databases PubMed, PsychInfo, Science Direct, Web of Science and Google Scholar were searched using the following terms: (“health anxiety” OR “health anxious” OR hypochondria*) AND (online OR Internet OR web OR net)) OR cyberchondria*). The terms were kept deliberately broad to ensure that all of the relevant literature was captured. Searches were restricted to study title, abstract and keywords, or title and abstract, dependent on the functionality of the specific database. Where the advanced search option did not support wildcards (Science Direct), “hypochondriasis” and “hypochondriac”, and “cyberchondria” and “cyberchondriac” were added to the search terms. Where the advanced search option placed a limit on number of Boolean Operators per search (Science Direct), two searches were performed within that database, one for (“health anxiety” OR “health anxious” OR hypochondria OR hypochondriac or hypochondriasis) AND (online OR Internet OR web OR net)), and one for (cyberchondria OR cyberchondriasis OR cyberchondriac). A second PubMed search was undertaken using the medical subject heading (MeSH) system, for studies indexed under both “hypochondriasis” and “Internet”. Reference lists of included articles were also searched to identify relevant articles. No time limits were imposed on these searches as widespread Internet use, and engagement in health-related Internet use more specifically, are relatively recent phenomenaTitles and abstracts of articles resulting from these searches were screened. Full texts of potentially relevant articles were then examined. Quantitative studies were retained if they measured or observed HRIU, and either claimed to measure health anxiety, measured a variable closely related to health anxiety, or made inferences about health anxiety based on behavioural data. Qualitative studies were retained if they contained discussion of the relationship between HRIU and health anxiety. Articles were excluded if they: were not written in English; were not empirical; or were not published in a peer-reviewed journal. Full papers published in conference proceedings were retained, while abstracts of conference presentations and posters were excluded. The final searches were undertaken in June 2018. The search yielded a total of 441 results, of which 237 were duplicates (Figure 2). Application of exclusion criteria yielded a total of 38 papers for review, spanning from 2006 to 2018. One further paper (McMullan et al., 2019), was published after search completion, and subsequently incorporated into our review due to its relevance. --- INSERT FIGURE 2 ABOUT HERE---In the sections that follow we group these studies thematically, evaluating what they tell us about the extent, nature, purpose, impact and outcome of HRIU in health anxious individuals. REVIEW OF EMPIRICAL LITERATURE Approaches to measurement Various approaches to measurement are adopted in the literature on OHR, health anxiety, and the relationship between these variables. These approaches are outlined below to help contextualise the relevant study findings. Many studies use single-item, self-report scales to measure the frequency or duration of OHR (Baumgartner & Hartmann, 2011; Doherty-Torstrick, Walton, & Fallon, 2016; Eastin & Guinsler, 2006; Fergus & Dolan, 2013; Muse et al., 2012; Tanis et al., 2016), the impact of OHR on health anxiety and/or distress (Doherty-Torstrick et al., 2016; Fergus, 2013; Fergus & Dolan, 2014; Muse et al., 2012), and the occurrence of specific emotional responses to OHR (Baumgartner & Hartmann, 2011). There is great deal of variation in how these single-item measures are constructed. For example, studies have asked about frequency/duration of OHR in general, of recent OHR, and of the last episode of OHR. Use has also been made of unvalidated composite measures of OHR (Lagoe & Atkin, 2015; Singh & Brown, 2014; te Poel et al., 2016). McElroy and Shevlin (2014) devised a scale purporting to measure cyberchondria (the Cyberchondria Severity Scale; CSS), which they defined as “anxiety about one’s own health status, as a result of excessive reviews of online health information” (p. 259). They provided only brief detail on the item generation process, stating that they were developed based on a review of existing literature. In addition to the original English version, German (Barke et al., 2016), Turkish (Selvi, G?k?e Turan, Asena Sayin, Boysan, & Kandeger, 2018) and Brazilian Portuguese (da Silva, Andrade, Silva & Cardoso, 2016) translations have been published, but fewer psychometric data are available for these. Excellent full-scale internal consistency has been reported for the CSS (ranging from .91 reported by Selvi et al., 2018 to .96 reported by Norr, Albanese et al., 2015 and Mathes et al., 2018). The original, 33-item CSS was validated with a student sample (N = 190), and exploratory factor analysis indicated a five-factor structure. Norr, Allan et al.’s (2015) modelling indicated that the CSS can be conceptualised as measuring a general cyberchondria factor, with four orthogonal subfactors (excessiveness, distress, compulsion, and reassurance seeking). The mistrust subscale, pertaining to placing greater trust in online health information than information provided by medical professionals, has been excluded from revised versions of the CSS on the basis of Norr, Allan, et al.’s findings. The internal consistency of the remaining subscales ranges from good-to-excellent for the English-language CSS (ranging from .81 reported by Norr, Allan, et al., 2015 to .96 reported by Mathes et al., 2018). No study to date has reported on test-retest reliability of the CSS. With regards convergent validity, the CSS shows weak correlations with depression (Barke et al., 2016; McElroy & Shevlin, 2014) and a moderate correlation with generalised anxiety (McElroy & Shevlin, 2014). Moreover, total score on the CSS is consistently associated with health anxiety, with most studies reporting strong to moderate correlations (Barke et al., 2016; Fergus, 2014; Fergus, 2015; Norr, Albanese, et al., 2015) although weak correlations have also been reported (Fergus & Russell, 2016; Selvi et al., 2018). In a meta-analysis, McMullan et al. (2019) found that total CSS score and health anxiety were strongly correlated, but with a high degree of heterogeneity of effect sizes between studies (r = .62, p <.0001, Q = 76.49, p < .0001, I2 = 90.13%). There is variation in how the individual CSS subscales relate to health anxiety, as will be discussed in later sections. Extent of OHR OHR and health anxiety evidently co-vary. Several cross-sectional studies have found a significant positive correlation between the extent (i.e., frequency and/or duration) of OHR and a validated measure of health anxiety (Baumgartner & Hartmann, 2011; Doherty-Torstrick et al., 2016; Fergus, 2013; Muse et al., 2012; Singh & Brown, 2014; Tanis et al., 2016; see Table 1). Effect sizes are weak-to-moderate. However, the only study to control for generalised anxiety and depression found no relationship between health anxiety and extent of OHR (Singh & Brown, 2014). The meta-analysis undertaken by McMullan et al. (2019) examined the relationship between health anxiety and OHR, finding a moderate positive relationship, but also a high degree of heterogeneity (r = .34, p < .0001, Q = 133.93, p < .0001, I2 = 92.42%). Studies to date are limited by their reliance on self-reported data, often from single items, concerning the frequency and duration of OHR. It is also important to note that studies to date have provided very limited information on the actual amount of time participants spend engaging in OHR. More generally, although there are data on the prevalence of OHR (Fox & Duggan, 2013; Office for National Statistics, 2018), no study has provided information on patterns of OHR (frequency, duration) in the general population, or a representative sample of clinically health anxious individuals. --- INSERT TABLE 1 ABOUT HERE---Other studies have found associations between health anxiety and the excessiveness subscale of the CSS (Barke et al., 2016; Fergus, 2014, 2015; Fergus & Russell, 2016; Mathes, et al., 2018; Norr, Allan, et al., 2015; Norr, Oglesby et al., 2015; Selvi et al., 2018), which is said to reflect “an unnecessary amount of time spent researching the same symptoms and health conditions” (McElroy & Shevlin, 2014, p. 261). However, only three items in this seven-item subscale have obvious face validity pertaining to repetitive searching, casting doubt on whether these studies really speak to the relationship between health anxiety and extent of OHR. It is also unclear whether positive scores on this subscale reflect repeated, normatively high levels of OHR, or use that is perceived by the individual as excessive (and hard to control). This is important theoretically: the former may be consistent with a reassurance seeking perspective, and the latter more indicative of a compulsive behaviour.A preponderance of cross-sectional studies means that the direction of causality between health anxiety and OHR is unclear. To date, only one longitudinal study has investigated whether health anxiety predicts the extent of subsequent OHR. Te Poel et al. (2016) undertook a four-wave longitudinal study over a six-month period with a large, representative community sample, utilising a random intercept cross-lagged panel modelling strategy (Hamaker, Kuiper & Grasman, 2015). Similar to previous, cross-sectional studies, individuals who were higher in health anxiety reported engaging in more OHR. A novel finding was that, for those low in health anxiety at baseline, a reciprocal relationship was observed over time between health anxiety and OHR. No such reciprocal relationship over time was observed for those in the clinical range for health anxiety at baseline. This study provides some support for the hypothesis that OHR can drive health anxiety upwards over time. It remains unstudied whether OHR in low health anxious individuals can contribute to a shift towards clinically significant anxiety, and greater consideration needs to be given to the mechanisms underpinning any such shift. Emotional Consequences of OHRThe proportion of people who report becoming anxious about their health following OHR varies considerably. Studies using convenience samples and single, self-report items tend to report relatively high rates (72.7%, Berezovska, Buchinger, & Matsyuk, 2010; 31.4%; Fergus & Dolan, 2014; 38.5%, White & Horvitz, 2009b), whereas the one study with a representative sample (N = 4906) reported a much lower rate (15%; Andreassen et al., 2007). Health anxiety is significantly related to greater self-reported anxiety and distress post-OHR (Doherty-Torstrick et al., 2016; Baumgartner & Hartmann, 2011; Muse et al., 2012; Singh & Brown, 2014; Table 2). There is also a consistent relationship between the distress subscale of the CSS and health anxiety (Barke et al., 2016; Fergus, 2014; Fergus & Russell, 2016; Mathes el., 2018; Norr, Allan, et al., 2015; Norr, Oglesby et al., 2015; Selvi et al., 2018; Table 2). The distress subscale is a unique predictor of score on the Short Health Anxiety Inventory (SHAI; Salkovskis, Rimes, Warwick, & Clark, 2002), independent of the other CSS subscales (Fergus, 2014). Similarly, in their structural equation modelling of latent dimensions of the CSS and SHAI, Norr, Allan, et al. (2105) found that distress was significantly related to both SHAI subscales. Thus, evidence to date suggests that distress post-OHR is more pronounced in health anxious individuals. However, none of these cross-sectional studies attempted to quantify the level or duration of post-OHR distress experienced, and there is a lack of naturalistic, ecologically valid studies of emotional responses to OHR.--- INSERT TABLE 2 ABOUT HERE---Greater frequency of OHR in health anxious people could be a factor in the increased distress experienced by this group as a result of OHR. An alternative (or perhaps complementary) interpretation is that OHR exposes users to information that is particularly threatening to people who are high in health anxiety. Consistent with this, qualitative studies with health anxious participants suggest that exposure to threatening information is one of the perceived downsides of OHR (McManus, Leung, Muse & Williams, 2015; Singh, Fox & Brown, 2016). Lauckner and Hsieh (2013) presented participants with mock results of searches for physical symptoms and found that feeling frightened by the search results was related to the perceived severity of the condition (β = .34, p < .001) and weakly related to the participant’s perceived susceptibility to it (β = .06, p < .05). Although no measure of baseline health anxiety was included in this study, it can be hypothesised that health anxious individuals are more inclined to perceive health conditions encountered online as serious, and to perceive themselves as susceptible to illness. There is some evidence that the relationship between health anxiety and emotional responses to OHR is dependent on the nature of the material encountered online. Baumgartner and Hartmann (2011) presented participants with an online text about a fictitious disease and manipulated the apparent trustworthiness of the information presented. Baseline health anxiety was associated with increased negative responses, including worry about the disease (b = 0.99, p < .01), but only when online health information was attributed to a trustworthy source. Health anxious individuals appeared to demonstrate the ability to discount information from a dubious online source. How health anxious individuals behave online might also be a factor in their greater levels of post-OHR distress. Singh and Brown (2016) observed high and low health anxious students engaging in 15 minutes of OHR for a symptom identified as personally relevant and threatening. Participants then watched a screenshot recording of their search and were asked to rate their anxiety during the search. Significant increases in anxiety were found immediately after participants accessed pages pertaining to serious illnesses. Such “query escalation” (White & Horvitz, 2009a,b) was more common in the health anxious participants. This study suggests that everyone is susceptible to feeling more anxious when they search for, and identify, information that poses a potential health threat to them; however, health anxious individuals may be more likely to engage in maladaptive search strategies that increase the likelihood of exposure to health threats. Online health threats might produce similar levels of anxiety in everyone, but the subsequent threat value of that anxiety may be greater in those who are particularly anxious about their health. Health anxiety is known to correlate with personality characteristics such as anxiety sensitivity (Otto, Demopulos, McLean, Pollack & Fava, 1998; Reiss, 1987), and intolerance of uncertainty (Fergus & Valentiner, 2011). The fear generated by encountering threatening online health information could lead to more intense, and enduring, distress in health anxious individuals who are very sensitive to physical anxiety symptoms, and intolerant of uncertainty regarding the meaning of these symptoms; this could then give rise to an increase in reassurance-seeking, setting up a vicious cycle. However, in the absence of longitudinal studies that include measures of anxiety sensitivity and intolerance of uncertainty, this remains untested. Cross-sectional studies of anxiety sensitivity and intolerance of uncertainty indicate that they do have some importance in understanding emotional responses to OHR.Fergus (2013) found a positive relationship between frequency of OHR and negative, anxious response to OHR for those high in intolerance of uncertainty, suggesting this individual difference may confer vulnerability for at least short-term negative emotional responses to online health information. In an experimental study, Norr, Capron and Schmidt (2014) found that exposure to medical symptom websites resulted in a greater increase in anxiety sensitivity than general health and wellness websites (β = .14, p < .05) for those high in intolerance of uncertainty (β = .30, t = 3.28, p = .002), suggesting a mechanism by which individuals high in this trait might become more distressed by OHR (i.e., via a spike in anxiety sensitivity triggered by OHR). Norr, Albanese et al. (2015) found significant positive relationships between total CSS scores and both anxiety sensitivity (β = .36, p < .001) and inhibitory intolerance of uncertainty (β = .22, p < .001); these were both independent of general health anxiety. Similar findings were reported by Fergus (2015). Norr, Albanese et al. suggest that these individual differences might be risk factors for experiencing ‘cyberchondria’, as measured by total CSS score, but the cross-sectional design does not allow for conclusions regarding such temporal relationships. Nonetheless, the finding of a relationship between these variables and the CSS independent of health anxiety is intriguing. It supports their understanding of ‘cyberchondria’, as measured by the CSS, as a phenomenon distinct from health anxiety that cannot be wholly understood as reassurance seeking. Why do health anxious people engage in more OHR?If OHR is more threatening and distressing for people with higher levels of health anxiety, why do these individuals spend the most time searching the Internet for health information? The reassurance seeking model suggests that health anxious people engage in more OHR because they have a greater need to restore emotional equilibrium, perhaps due to higher levels of anxiety in general, greater anxiety sensitivity or greater intolerance of uncertainty. Consistent with this, Bell, Hu, Orrange and Kravitz (2011) found that engaging in OHR after a medical consultation was predicted by the level of worry about the problem in question, and by feeling more worried post-consultation. A qualitative study by Singh et al. (2016) also found that a key motive for OHR in health anxious students was a desire for reassurance regarding a perceived health threat, and that one of the main factors informing the decision to terminate a search was whether reassurance had been achieved. Similarly, a qualitative study by McManus et al. (2015) found that some health anxious individuals experienced the search process as comforting. Strong evidence that health anxious people are more likely to experience relief after searching is lacking, however. Although Singh and Brown (2014) found that relief post-OHR was a significant unique predictor of SHAI score (β = .16, p < .01), no such relationship was found by Baumgartner and Hartmann (2011). Singh et al. (2016) found that participants could experience either anxiety or relief depending on whether the search resulted in the perceived resolution of a given health threat, potentially supporting the Starcevic and Berle (2013) model described earlier. A variation on the reassurance model is that health anxious people engage in repeated OHR because they consider it necessary to manage potential health threats. From this perspective, a successful search might be seen as one where useful information is gathered, even if that information is seen as negative or anxiety-provoking (a “better safe than sorry” strategy). Consistent with this, health anxious participants in the study on query escalation by Singh and Brown (2016) identified the main motivation for query escalations as the need to determine whether the escalated-to condition was a possible explanation for their symptoms, suggesting a desire to manage the health threat. From a cognitive behavioural perspective, this OHR is likely to be driven by the belief that knowing the threat level increases safety. However, this behaviour is perhaps better understood from a metacognitive perspective. The metacognitive model proposes that psychopathology is maintained by maladaptive positive and negative beliefs about thinking (Wells, 2009). In this instance, a positive metacognition about the protective value of thinking about risk might be at play. If the perceived trustworthiness of online health information influences how anxiety-provoking it is (Baumgartner & Hartmann, 2011; see above), then it may also influence whether someone expects to be reassured by what they encounter. If so, the correlation between health anxiety and the extent of OHR if health anxiety might be driven by a tendency to perceive the Internet as a source of credible information in the most health anxious individuals. However, neither Muse et al. (2012; t[165] = 0.51, p = .61) nor Singh and Brown (2014; r = .12, p > .05) found a relationship between health anxiety and the perceived reliability of online health information.Another possibility is that health anxious individuals are more likely to perceive their doctor as relatively incapable of managing their health threat effectively, prompting reassurance seeking from other sources. Both Singh et al. (2016) and McManus et al. (2015) found that some health anxious people reported engaging in OHR because they found doctors inadequately reassuring or had had negative experiences with health professionals. Singh and Brown (2014) found a significant correlation between health anxiety and negative attitudes towards doctors (β = .14, p = .01). Fergus (2014) reported a significant, but weak, relationship between health anxiety and the mistrust subscale of the original CSS (partial r = .15, p < .01). However, most studies that considered the relationship between OHR and the perceived trustworthiness of information provided by doctors versus the Internet have reported non-significant findings (r = .13, p > .05, r = .18, p >.05, Barke et al., 2016; r = -.02, p > .05, Norr, Albanese, et al., 2015; r = .01, p > .05, r = .-08, p > .05, Norr, Allan, et al., 2015; β = .10, p > .05, Singh & Brown, 2014). It is also possible, but little studied, that excessive OHR is damaging to the doctor-patient relationship for health anxious individuals, and negatively affects the perception of specific aspects of medical care, which could in turn fuel further OHR. Tanis et al. (2016) found that the extent of prior OHR by health anxious individuals was negatively correlated with their satisfaction with the length of a medical consultation (b = -.18, t = -3.38, p < .001). However, their study also reported several non-significant relationships between OHR and care satisfaction.Intrinsic to the reassurance seeking concept is the idea that OHR confers some benefit to the individual (at least on some occasions), and that this explains repeated searching in the face of other negative effects. The compulsion concept, in contrast, suggests that health anxious individuals are more likely to engage in OHR because they struggle to limit or control their use, even though they want to. Evidence for this was obtained in the qualitative study by Singh et al. (2016), who found that some health anxious individuals described difficulties controlling their search behaviour despite its potential to increase anxiety. Similarly, the compulsion subscale of the CSS is a consistent correlate of health anxiety (r = .42, r =.39, p < .001, Barke et al., 2016,; r = .51, p < .001, Fergus, 2015; r = .41, p < .01, Fergus & Russell, 2016,; r = .37, r = .41, p < .05, Mathes et al., 2018; r = .49, p > .001, Norr, Oglesby, et al., 2015; r = .23, p < .01, Selvi et al., 2018). However, it was not a unique predictor of SHAI scores when controlling for other CSS subscales (partial r = .05, p > .05) in a study by Fergus (2014). Moreover, there was no significant pathway between compulsion and either SHAI subscale in the structural equation model developed by Norr, Allan, et al. (2015). The CSS uses the term “compulsion” in a very specific way, with the corresponding subscale items only pertaining to the interruption of other activities by OHR (McElroy & Shevlin, 2014). There is no reference to whether the OHR in question is ego dystonic or feels outside the individual's control, which are arguably more central aspects of compulsion. Singh and Brown (2014) found that health anxiety was associated with searching for longer than intended, increasing OHR over time, and unsuccessful attempts to limit and reduce OHR, as well as a negative impact on one’s social, work and/or academic life. These features, coupled with feelings of restlessness and irritability when unable to engage in OHR, and the belief that life would be worse without it, point to compulsive or ‘problematic’ Internet use that may amount to addiction in the worst cases (Young, 1998; Aboujaoude, 2010). It may be that this is part of a broader problem with limiting Internet use that is not specific to health information. Indeed, Fergus and Dolan (2014) found that people who reported feeling more anxious following medical searches were more likely to report problematic internet use more generally than those who reported no change (Cohen’s d = .33, p < .01) or a reduction in their anxiety post-OHR (d = .61, p < .01); this relationship held even when controlling for frequency of OHR. Baseline health anxiety was not measured, however. Ivanova (2013) reported similar findings using an unvalidated measure of anxiety arising from OHR. More recently, Fergus and Spada (2018) found a strong, positive association between total scores on the CSS and a general measure of problematic Internet use (r = .59, p < .001), with the latter accounting for 13% of the variance in CSS scores after controlling for age, gender, physical health status, negative affect and health anxiety. According to this study, metacognitive beliefs about the uncontrollability of health-related thoughts were a unique predictor of CSS scores, even when taking health anxiety and other cognitive factors into account (β = .19, p < .008); this provides further support for a compulsion-based understanding of OHR, and further highlights the potentially important but neglected role of metacognitions in distressing, repetitive OHR. Behavioural consequences of OHR If health anxious people are more likely to engage in OHR and more likely to have an emotional response to it, we might expect there to be a relationship between health anxiety and what people do after searching online. Research has to date focused on reassurance seeking and healthcare utilisation post-OHR. The most widely used measure is the reassurance subscale of the CSS, which consists of items regarding doctor utilisation and the tendency to raise information found online with doctors. Consistent correlations with health anxiety have been found (Table 3), although the subscale ceases to be uniquely associated with health anxiety when other CSS factors are taken into account (Fergus, 2014; Fergus & Russell, 2016; Norr, Allan et al., 2015). The extent and emotional impact of OHR appear to be the main sources of the relationship between health anxiety and post-search reassurance seeking.A study by Eastin and Guinsler (2006) suggests that associations between frequency of OHR and post-OHR health service utilisation depend on participants’ health anxiety, with significant correlations only being found at high (β = .24, t = 5.36, p < .05) but not low levels of health anxiety (β = - .03, t = -0.53, p > .05). This seems to suggest that the overall amount of OHR has less of a bearing on the health care utilisation of people who tend not to worry about their health, perhaps suggesting that they are better able to judge the level of threat indicated by online information and therefore whether a doctor visit is necessary. Additionally, the most health anxious individuals may be more likely to encounter health threats whilst searching, because of search behaviours like query escalation, thereby increasing the perceived need to seek medical input. These findings should be treated with caution, however, as the 12-month recall period used in this study may not provide a reliable estimate of HSU, and self-reported HSU can be confounded in multiple ways (Bhandari & Wagner, 2006; Short et al., 2009). No study to date has employed an objective measure of HSU, such as electronic patient records.Mathes et al. (2018) is the only study to have looked separately at physical and mental health care utilisation, using an online sample and a self-report measure pertaining to the preceding 60 days. Unsurprisingly, the reassurance subscale of the CSS shared a positive, unique association with physical health care utilization, independent of health anxiety. Reassurance was also uniquely, positively associated with mental health care utilization, suggesting that individuals who present more frequently to medical professionals seeking reassurance, and perhaps specifically wishing to discuss information encountered online, may be more likely to be referred to, or encouraged to engage with, mental health services. --- INSERT TABLE 3 ABOUT HERE---There is very little research on what drives health utilisation and other reassurance seeking behaviour after someone has engaged in OHR. Qualitative data from Singh et al. (2016) suggest that non-resolution of perceived health threats is a key driver of this process in health anxious individuals. This may arise in situations where the person perceives that a serious threat is present that requires medical input, or where the search itself is inconclusive because of inadequate or conflicting information. It is unknown whether this is particularly characteristic of people who are anxious about their health, however. It is also unknown whether OHR drives other behaviours associated with health anxiety, such as body checking, avoidance of feared activities and other safety/precautionary actions, which might be expected. SUMMARY, DISCUSSION AND CONCEPTUAL INTEGRATION The number of studies on health anxiety and OHR has grown considerably in recent years and some aspects of the relationship between OHR and health anxiety are beginning to be understood: (i) extent (i.e., frequency and/or duration) of OHR correlates consistently with health anxiety; (ii) a significant proportion of people find OHR anxiety-provoking, and this is more common in those who are health anxious more generally; and (iii) higher baseline health anxiety is associated with a tendency to report more doctor visits following OHR. A degree of caution is necessary when drawing conclusions from this literature, however. Many studies have significant limitations, including the use of unvalidated self-report measures or single-item measures with questionable psychometric properties. Potentially unrepresentative samples, such as students or people recruited entirely online, are commonplace, and the potentially confounding effect of actual physical health status is also typically overlooked, as is co-morbid psychopathology. Moreover, most of the research describes cross-sectional correlations that say little about the direction of causality. In general, there is a lack of evidence regarding what kind of self-report items concerning OHR give rise to the most reliable data, and minimal consideration of issues such as reporting timescales, and the utility of general questions that fail to specify a timescale.Another limitation of the research in this area is the absence of studies that are firmly grounded in a theoretical model of the relationships between the constructs in question, which has resulted in a lack of clarity about the phenomenon of interest, be it OHR per se or online health behaviour of a particular kind, such as searching that is perceived by the individual as excessive, or searching that is normatively excessive. Nevertheless, our review found some empirical support for the reassurance and compulsion concepts, suggesting that a hybrid model may be necessary to capture all of the relevant phenomena and their inter-relationships. The model outlined by Starcevic and Berle (2013) is one candidate in this respect, although it seems to rests on the assumption that cyberchondria only develops in individuals who have never experienced OHR as reassuring. Although it is certainly true that many individuals score high on scales measuring their tendency to feel distressed as a result of searching, qualitative studies suggest that the same individuals may also feel reassured by OHR at other times. Regardless, it is apparent that even those individuals who find OHR distressing are motivated by a desire for reassurance (whether it is forthcoming or not), which needs to be accommodated by any comprehensive account of this phenomenon. The Starcevic and Berle model is also relatively circumscribed, and says little about the various factors that appear to contribute to the decision to engage in or terminate OHR. Modelling the relationship between health anxiety and pathological online health research In this section, we present a novel model of the relationship between health anxiety, OHR and cyberchondria (i.e., pathological OHR) that seeks to integrate existing research and theory about these concepts whilst addressing some of the shortcomings of existing accounts. The model accommodates both the reassurance seeking and compulsion perspectives by drawing on the cognitive behavioural model of health anxiety (Warwick & Salkovskis, 1990), the metacognitive model of generalized anxiety disorder (GAD; Wells, 1997), and the model of cyberchondria by Starcevic and Berle (2013), creating a hybrid account that distinguishes between normal and pathological OHR. According to this approach (Figure 3), OHR is a common response to health threat and is typically motivated by a need to evaluate and hopefully eliminate the threat. This may be true regardless of a person’s pre-existing health anxiety, although we assume it is more common in health anxious individuals because they are more likely to perceive potential health threats. In cases where the outcome of the search suggests that the threat is minimal, further online behaviour can be terminated and the emotional outcome will be reassurance/relief; this negatively reinforces OHR, whilst fostering the belief that it is an effective way of managing health threat and thereby reducing worry (a positive meta-cognitive belief; Wells, 2000). In contrast, when the outcome of the search confirms the possible significance of the threat (or, at least, is ambiguous in this respect), the emotional outcome is further worry and anxiety, motivating the individual to engage in additional OHR to manage the threat and the accompanying distress. As health anxious individuals have a tendency to misinterpret illness information as potentially threatening (Marcus et al., 2007), they are more likely to experience the search process as anxiety provoking (as are those with heightened anxiety sensitivity), with a corresponding increase in their OHR. Depending on the nature of the search and the resulting outcome, there may be a significant increase in anxiety at this point, particularly in cases where query escalation exposes the individual to information about potentially serious health threats. We conceptualise the query escalation process as an attempt to evaluate the nature and extent of any possible health threats, with a view to establishing some certainty in this respect (i.e., a “better safe than sorry” strategy). As such, we might expect query escalation to be more common in people who are anxious about their health, particularly those who are intolerant of uncertainty. ---INSERT FIGURE 3 ABOUT HERE---We hypothesise that one consequence of obtaining worrying search results is the development of negative beliefs about the impact of searching on the individual’s wellbeing and thereby the potential threat value of OHR itself (e.g., “I’m making myself sick with OHR”). Over time, the unpredictable outcome of OHR is likely to set up a conflict between the impulse to manage health threat by engaging in OHR and the need to protect one’s mental state by not doing so. If this conflict (which is likely to be intrinsically stressful) cannot be resolved satisfactorily, then the individual will become stuck engaging in a behaviour that they perceive as necessary but ultimately undesirable. At this point, their OHR takes on a compulsive quality, compounded by the perception that they are at the mercy of their conflict and have little choice or control over their behaviour; by now, their progression to excessive, distressing and compulsive OHR (what is often called cyberchondria) is complete.The advantage of this model is that it makes a clear distinction between “normal” and “abnormal” OHR whilst explaining the progression from one to the other. There are three key differences between our account and that of Starcevic and Berle (2013), which are worth highlighting at this stage. First, we suggest that OHR in general should be regarded primarily as a form of threat evaluation and reassurance seeking (i.e., a safety behaviour), and that this is true regardless of whether the outcome is threatening or reassuring. Second, we suggest that search behaviour with a worrying or distressing outcome should not be regarded as problematic OHR in its own right, since some anxiety may be an appropriate response to genuine threat and motivate adaptive illness behaviour. Third, we assume that there are two different types of pathological OHR, which are problematic for different reasons. More specifically, we distinguish between (i) OHR that functions as a safety behaviour within the context of health anxiety more generally (as in traditional cognitive behavioural models; e.g., Warwick & Salkovskis, 1990); we call this problematic OHR because it maintains preoccupation with health threat and the likelihood of having or developing a serious illness; and (ii) OHR that the individual perceives as distressing, compulsive and out of control, where the focus of the threat is the Internet use itself and its impact and implications for the person’s mental and physical state; we call this compulsive OHR. Although these states clearly overlap and may co-occur, we suggest that they are separable and may have distinct clinical implications (see below). The inclusion of positive and negative metacognitive beliefs about OHR as a worry management strategy that becomes threatening in its own right (a notion that is central to the metacognitive model of GAD; Wells, 2000), is one of the most novel features of our account. At this stage, the inclusion of this element is largely on theoretical grounds (although see Singh et al., 2016, and Fergus & Spada, 2018, for some supportive evidence) and evaluating this component of the model should be an important focus for future research. In the next section, we consider other empirical implications of this model and further priorities for future work in this area.287543113007Ten priorities for the future Establish conventions regarding relevant concepts, terms and definitions. For the field to advance, there needs to be greater clarity and consistency about the most important constructs. Our perspective is that there is an important distinction between online health research (i.e., using online resources to obtain information about health matters, which can be qualified if necessary by terms denoting its frequency, focus and emotional outcome e.g., frequent OHR; illness-related OHR; distressing OHR; reassuring OHR), problematic online health research (i.e., OHR that functions as a safety behaviour in the context of health anxiety) and compulsive online health research (i.e., OHR that the individual perceives as distressing, undesirable and beyond their control). OHR is a subset of health related internet use more generally. Our model suggests that different forms of HRIU are only likely to relate to health anxiety if they are motivated by the desire to ameliorate health threat. We prefer not to use the term cyberchondria, which we believe lacks face validity and is applied inconsistently. The fact that many patients do not like the root term hypochondriasis, and its discontinuation in the Diagnostic and Statistical Manual (DSM-5; APA, 2013), are also pertinent considerations. Develop better ways of measuring core concepts. At present, it is impossible to say with any certainty how much OHR health anxious individuals engage in and how that compares to those who are less anxious about their health. Similarly, it is unclear exactly what individuals, at different levels of health anxiety, are searching for online, what their searches yield, how often the results are interpreted as reassuring or threatening, and what the emotional and behavioural impact of this is. A key part of the problem has been a reliance on unvalidated, often single-item self-report measures, typically obtained retrospectively, where the individual is asked to report about their OHR in general. To date, no measure of OHR has been validated against actual online search behaviour. Develop tools that distinguish between OHR in general, problematic OHR and compulsive OHR. At present, the only established measure in this area is the CSS. Although this scale has good psychometric properties and is a consistent correlate of health anxiety, questions remain over what it measures and whether it taps the most relevant dimensions. According to our model, simply measuring the extent to which a person engages in repeated Internet searching (excessiveness subscale), how distressing they find it (distress subscale), the extent to which it interrupts other activities (compulsion subscale) and whether they visit the doctor afterwards (reassurance subscale) may or may not tell us whether the person’s OHR is pathological and, if so, in what way. In particular, the locus of threat in problematic OHR may be different to that in compulsive OHR. As such, we might expect problematic OHR to correlate mainly with measures of health anxiety and perceived health threat, whereas compulsive OHR may also correlate with measures tapping the threat value of online behaviour itself. With that in mind, there is a pressing need for a measure that address the individual’s perceptions of their online behaviour, particularly how worried they are by the amount that they are searching, the extent to which they would rather not do it and their beliefs regarding their ability to control their OHR. Our account predicts that these measures will be correlated with one another but that there will be individuals who engage in problematic OHR but who would not be considered to have compulsive OHR as such, which is an important test of the model. Target more representative populations. Many of the studies considered in this review utilised arguably unrepresentative convenience samples. As far as possible, future studies should seek to recruit clinical participants and/or community samples via other mechanisms to ensure generalisability. Determine the scale of the problem. Having established measures of the relevant constructs, we would be in a much better position to establish what constitutes clinically significant levels of problematic and compulsive OHR, and then to answer basic epidemiological questions about their prevalence and incidence, as well as the difficulties that these people face and their healthcare needs. There is also the broader question of how damaging problematic OHR is in terms of maintaining, exacerbating or indeed provoking health anxiety. Characterise problematic and compulsive OHR in more detail. Being able to identify problematic and compulsive OHR more accurately would also enable us to study their characteristics in much greater depth. There are numerous questions to address in this respect. For example, which individuals are most prone to developing problematic and compulsive OHR? What are the differences between health anxious individuals with versus without problematic or compulsive OHR? What are their relationships like with healthcare providers? How do they interact with doctors in relation to their OHR? What do health providers think about these people and what effect does it have on their clinical practice with them?Elucidate aspects of the theoretical model and test key predictions. The model described above is an attempt to organise existing concepts and data within a unifying theoretical framework, with a view to identifying important areas and new hypotheses for further study. It is far from a definitive account, however, and is likely to be modified considerably as the field moves forward. An important step is to establish the importance or otherwise of positive and negative meta-beliefs in shaping online behaviour and the different types of pathological OHR. Longitudinal studies could be useful for testing whether negative meta-beliefs concerning the dangers and controllability of online behaviour are a pre-cursor for compulsive OHR, perhaps in individuals who are already high, but not compulsive, users. Experimental studies looking to change negative meta-beliefs could also help establish their importance in compulsive OHR and point towards possible clinical strategies with this group. The model also suggests that an individual will continue to engage in OHR until they are reassured by the outcome. Whether this is really the case needs to be established empirically, and will require detailed study of individual search episodes. It is also an open question about what constitutes a worrying versus a reassuring outcome, which is likely to vary considerably between individuals and contexts. It may be that reassurance is as much about having some certainty regarding the level of threat (whether it is low or not) and what one should do about it (e.g., visit a doctor) as the threat level itself.Identify, study and tackle the most damaging online behaviours and web designs. The few studies on query escalation suggest that this may be a particularly toxic online behaviour. Confirming the extent to which query escalation occurs, and what its effects are over time would help establish its importance, and whether strategies specifically targeting it are warranted. It is possible that online health websites (e.g., WebMD) could be configured to recognise when users are engaging in common query escalations (e.g., escalating from headache to brain tumour) and to deploy simple interventions aimed at minimising the negative impact of this (Starcevic & Berle, 2013). There is also evidence that common and typically benign symptoms (e.g., headache, abdominal pain) will be perceived as more severe if the output of searches pertaining to them includes more mentions of serious illness and if these appear nearer the top of the page (Lauckner & Hsieh, 2013; see also White & Horvitz 2009a). If so, it should be possible to design health search engines in a way that minimises how anxiety provoking they are, whilst ensuring that they solicit appropriate illness behaviour where necessary.Establish what appropriate OHR looks like and for whom. In research on the consulting behaviour of health anxious individuals there is a risk that any increase in doctor visits, such as that following OHR, will be seen as unnecessary reassurance seeking. Conversely, it is possible that any reduction in healthcare utilisation associated with OHR in low anxious individuals (see e.g., Eastin & Guinsler, 2006) will be regarded in a positive light, with online resources being seen as a useful way of reducing medical expenditure for this group. Although this may be true, these are empirical questions that require detailed study. It is possible, for example, that the “better safe than sorry” strategy employed by some health anxious individuals does confer some health benefit, or that the self-diagnoses made by low anxious individuals are dangerously complacent. In an age where online medical information and patient empowerment are proliferating, and health services are increasingly stretched, it is crucial that we establish what kind and level of OHR is appropriate for different people, and when OHR should prompt or inhibit subsequent help-seeking. Develop appropriate intervention strategies. It may be that treatment strategies based on a conventional CBT model of health anxiety are not the only or most appropriate way of tackling pathological OHR; indeed, our model suggests that practitioners might benefit from considering metacognitive concepts and interventions (e.g., Wells, 2000), particularly in cases where OHR has a compulsive quality to it.ConclusionsIn the digital age, people using the Internet for health purposes is only likely to become more common. Our review explores the relationship between health-related Internet use and health anxiety, and demonstrates the risks of OHR for certain individuals. As the field moves forward, there is a pressing need for more theoretically-driven research on this topic, with greater clarity and consistency about relevant terms and concepts, more nuanced approaches to measurement, and more emphasis on characterising the phenomena of interest. We hope that the model and research agenda presented here, and the review on which it is based, provide a useful impetus for this work and a platform for the development of more focused treatments for problematic and compulsive health-related Internet use.REFERENCESAboujaoude, E. (2010). Problematic Internet use: An overview. World Psychiatry, 9, 85–90. doi:10.1002/j.2051-5545.2010.tb00278.xAbramowitz, J. S., & Braddock, A. E. (2008). Psychological treatment of health anxiety and hypochondriasis: A biopsychosocial approach. G?ttingen: Hogrefe & Huber.Abramowitz, J. S., & Moore, E. L. (2007). An experimental analysis of hypochondriasis. Behaviour Research and Therapy, 45, 413–424. doi:10.1016/j.brat.2006.04.005Aiken, M., & Kirwan, G. (2014). The psychology of cyberchondria and ‘cyberchondria by proxy. In A. Power and G. Kirwan (Eds.), Cyberpsychology and new media: A thematic reader (pp. 158 – 169). Hove: Psychology Press. Aiken, M., Kirwan, G., Berry, M., & O’Boyle, C. (2012). The age of cyberchondria. Royal College of Surgeons in Ireland Student Medical Journal, 5, 71–74.American Psychiatric Association (2013). Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition. Arlington, VA: American Psychiatric Association.Andreassen, H. K., Bujnowska-Fedak, M. M., Chronaki, C. E., Dumitru, R. C., Pudule, I., Santana, S., … Wynn, R. (2007). European citizens’ use of E-health services: A study of seven countries. BMC Public Health, 7. doi:10.1186/1471-2458-7-53Asmundson, G. J. G., Abramowitz, J. S., Richter, A. A., & Whedon, M. (2010). Health anxiety: Current perspectives and future directions. Current Psychiatry Reports, 12, 306–312. doi:10.1007/s11920-010-0123-9Bailer, J., Kerstner, T., Witth?ft, M., Diener, C., Mier, D., & Rist, F. (2015). Health anxiety and hypochondriasis in the light of DSM-5. Anxiety, Stress, & Coping, 29, 219–239. doi:10.1080/10615806.2015.1036243Barke, A., Bleichhardt, G., Rief, W., & Doering, B. K. (2016). The Cyberchondria Severity Scale (CSS): German validation and development of a short form. International Journal of Behavioral Medicine. 23(5), 595-605. doi:10.1007/s12529-016-9549-8Baumgartner, S. E., & Hartmann, T. (2011). The role of health anxiety in online health information search. Cyberpsychology, Behavior and Social Networking, 14, 613–618. doi:10.1089/cyber.2010.0425Bell, R. A., Hu, X., Orrange, S. E., & Kravitz, R. L. (2011). Lingering questions and doubts: Online information-seeking of support forum members following their medical visits. Patient Education and Counseling, 85, 525–528. doi:10.1016/j.pec.2011.01.015Berezovska, I., Buchinger, K., & Matsyuk, O. (2010). Evolving facets of cyberchondria: Primum non nocere “First, do no harm.” In M. Kalogiannakis, D. Stavrou, & P. Michaelidis (Eds.), Proceedings of the 7th International Conference on Hands-on Science. Retrieved from , A., & Wagner, T. (2006). Self-reported utilization of health care services: Improving measurement and accuracy. Medical Care Research and Review, 63, 217–235. doi:10.1177/1077558705285298Brown, H. D., Kosslyn, S. M., Delamater, B., Fama, J., & Barsky, A. J. (1999). Perceptual and memory biases for health-related information in hypochondriacal individuals. Journal of Psychosomatic Research, 47, 67–78. doi:10.1016/S0022-3999(99)00011-2Bull, S. (2011). Technology-based health promotion. London: Sage.Cline, R. J. W., & Haynes, K. M. (2001). Consumer health information seeking on the Internet: The state of the art. Health Education Research, 16, 671–692. doi:10.1093/her/16.6.671Coiera, E. (1996). The Internet’s challenge to health care provision. British Medical Journal, 312, 3–4. doi:10.1136/bmj.312.7022.3da Silva, F., Andrade, R., Silva, I., & Cardoso, A. (2016). Cross-cultural adaptation of the Cyberchondria Severity Scale for Brazilian Portuguese. Trends in Psychiatry and Psychotherapy, 38(2), 90-95. doi:10.1590/2237-6089-2015-0063Davis, N. (2018, October 9). Cyberchondria and cyberhoarding: Is Internet fuelling new conditions? The Guardian. Retrieved from , R. A. (2001). Cognitive-behavioral model of pathological Internet use. Computers in Human Behavior, 17, 187–195. doi:10.1016/S0747-5632(00)00041-8Doherty-Torstrick, E. R., Walton, K. E., & Fallon, B. A. (2016). Cyberchondria: Parsing health anxiety from online behavior. Psychosomatics. Advance online publication. doi:10.1016/j.psym.2016.02.002Eastin, M. S., & Guinsler, N. M. (2006). Worried and wired: Effects of health anxiety on information-seeking and health care utilization behaviors. CyberPsychology & Behavior, 9, 494–498. doi:10.1089/cpb.2006.9.494eBiz MBA. (2019). Top 15 most popular health websites: January 2019. Retrieved from , G. (2003). The impact of the Internet on cancer outcomes. CA: A Cancer Journal for Clinicians, 53, 356–371. doi:10.3322/canjclin.53.6.356Fergus, T. A. (2013). Cyberchondria and intolerance of uncertainty: Examining when individuals experience health anxiety in response to Internet searches for medical information. Cyberpsychology, Behavior and Social Networking, 16, 735–739. doi:10.1089/cyber.2012.0671Fergus, T. A. (2014). The Cyberchondria Severity Scale (CSS): An examination of structure and relations with health anxiety in a community sample. Journal of Anxiety Disorders, 28, 504–510. doi:10.1016/j.janxdis.2014.05.006Fergus, T. A. (2015). Anxiety sensitivity and intolerance of uncertainty as potential risk factors for cyberchondria: A replication and extension examining dimensions of each construct. Journal of Affective Disorders, 184, 305–309. doi:10.1016/j.jad.2015.06.017Fergus, T. A., & Dolan, S. L. (2014). Problematic Internet use and Internet searches for medical information: The role of health anxiety. Cyberpsychology, Behavior, and Social Networking, 17, 761–765. doi:10.1089/cyber.2014.0169Fergus, T. A., & Russell, L. H. (2016). Does cyberchondria overlap with health anxiety and obsessive–compulsive symptoms? An examination of latent structure and scale interrelations. Journal of Anxiety Disorders, 38, 88–94. doi:10.1016/j.janxdis.2016.01.009Fergus, T.A., & Spada, M.M. (2017). Cyberchondria: Examining relations with problematic Internet use and metacognitive beliefs. Clinical Psychology and Psychotherapy, 24, 1322-1330. doi: 10.1002/cpp.2102Fergus, T. A., & Valentiner, D. P. (2011). Intolerance of uncertainty moderates the relationship between catastrophic health appraisals and health anxiety. Cognitive Therapy and Research, 35, 560-565. 011-9392-9Ferguson, E. (2009). A taxometric analysis of health anxiety. Psychological Medicine, 39, 277–285. doi:10.1017/S0033291708003322Fineberg, N.A., Demetrovics, Z., Stein, D.J., Ionnidis, K., Potenza, M.N., Grunblatt, E.,…Chamberlain, S.R. (2018). Manifesto for a European research network into Problematic Usage of the Internet. European Neuropsychopharmacology, 28, 11, 1232-1246. doi:10.1016/j.euroneuro.2018.08.004Fisher, R. (2006). Just can’t get e-nough. New Scientist, 192(2583), 34–37. doi:10.1016/S0262-4079(06)61454-1Fox, S., & Duggan, M. (2013). Health online 2013. Washington DC: Pew Research Center’s Internet & American Life Project. Retrieved from , E.L., Kuiper, R.M., & Grasman, R.P. (2015). A critique of the cross-lagged panel model. Psychological Methods, 20(1), 102-116. doi: 10.1037/a0038889Harding, K. J., Skritskaya, N., Doherty, E., & Fallon, B. A. (2008). Advances in understanding illness anxiety. Current Psychiatry Reports, 10, 311–317. doi:10.1126/stke.3542006pe36International Telecommunication Union. (2018). Measuring the information society report – volume 1. Geneva: Author. Ivanova, E. (2013). Internet addiction and cyberchondria: Their relationship with well-being. Journal of Education, Culture and Society, 1, 57–69. doi:10.15503/jecs20131-57-70Lagoe, C., & Atkin, D. (2015). Health anxiety in the digital age: An exploration of psychological determinants of online health information seeking. Computers in Human Behavior, 52, 484–491. doi:10.1016/j.chb.2015.06.003Lee, S.Y.?& Hawkins, R.P. (2016).?Worry as an uncertainty-associated emotion: exploring the role of worry in health information seeking. Health Communication, 31,?926-933,10.1080/10410236.2015.1018701Lauckner, C., & Hsieh, G. (2013). The presentation of health-related search results and its impact on negative emotional outcomes. In W. E. Mackay, S. Brewster, & S. B?dker (Eds.), Proceedings of CHI 2013 (pp. 333–342). doi:10.1145/2470654.2470702Longley, S. L., Broman-Fulks, J. J., Calamari, J. E., Noyes, R., Wade, M., & Orlando, C. M. (2010). A taxometric study of hypochondriasis symptoms. Behavior Therapy, 41, 505–514. doi:10.1016/j.beth.2010.02.002Loos, A. (2013). Cyberchondria: Too much information for the health anxious patient? Journal of Consumer Health on the Internet, 17, 439–445. doi:10.1080/15398285.2013.833452Mahdawi, A. (2015, February 11). Googled your cough? Is death now upon you? Then read this article. The Guardian. Retrieved from , D. K., Gurley, J. R., Marchi, M. M., & Bauer, C. (2007). Cognitive and perceptual variables in hypochondriasis and health anxiety: A systematic review. Clinical Psychology Review, 27, 127–139. doi:10.1016/j.cpr.2006.09.003Marcus, D. K., Hughes, K. T., & Arnau, R. C. (2008). Health anxiety, rumination, and negative affect: A mediational analysis. Journal of Psychosomatic Research, 64, 495–501. doi:10.1016/j.jpsychores.2008.02.004Marton, C., & Choo, C. W. (2012). A review of theoretical models of health information seeking on the web. Journal of Documentation, 68 (3), 330-352. doi: 10.1108/00220411211225575 Mathes, B.M., Norr, A.M., Allan, N.P., Albanese, B.J., & Schmidt, N.B. (2018). Cyberchondria: Overlap with health anxiety and unique relations with impairment, quality of life, and service utilization. Psychiatry Research, 261, 204-211. doi: 10.1016/j.psychres.2018.01.002McElroy, E., & Shevlin, M. (2014). The development and initial validation of the cyberchondria severity scale (CSS). Journal of Anxiety Disorders, 28(2), 259–265. doi:10.1016/j.janxdis.2013.12.007McManus, F., Leung, C., Muse, K., & Williams, J. M. G. (2015). Understanding “cyberchondria”: An interpretive phenomenological analysis of the purpose, methods and impact of seeking health information online for those with health anxiety. The Cognitive Behaviour Therapist, 7(e21), 1–13. doi:10.1017/S1754470X14000270McMullan, R.D., Berle, D., Arnaez, S., & Starcevic, V. (2019). The relationship between health anxiety, online health information seeking, and cyberchondria: Systematic review and meta-analysis. Journal of Affective Disorders, 245, 270-278. doi: 10.1016/j.jad.2018.11.037Muse, K., McManus, F., Leung, C., Meghreblian, B., & Williams, J. M. G. (2012). Cyberchondriasis: Fact or fiction? A preliminary examination of the relationship between health anxiety and searching for health information on the Internet. Journal of Anxiety Disorders, 26, 189–196. doi:10.1016/j.janxdis.2011.11.005Myin-Germeys, I., Oorschot, M., Collip, D., Lataster, J., Delespaul, P., & van Os, J. (2009). Experience sampling research in psychopathology: Opening the black box of daily life. Psychological Medicine, 39, 1533–1547. doi:10.1017/S0033291708004947Norr, A. M., Albanese, B. J., Oglesby, M. E., Allan, N. P., & Schmidt, N. B. (2015). Anxiety sensitivity and intolerance of uncertainty as potential risk factors for cyberchondria. Journal of Affective Disorders, 174, 64–69. doi:10.1016/j.jad.2014.11.023Norr, A. M., Allan, N. P., Boffa, J. W., Raines, A. M., & Schmidt, N. B. (2015). Validation of the Cyberchondria Severity Scale (CSS): Replication and extension with bifactor modeling. Journal of Anxiety Disorders, 31, 58–64. doi:10.1016/j.janxdis.2015.02.001Norr, A. M., Capron, D. W., & Schmidt, N. B. (2014). Medical information seeking: Impact on risk for anxiety psychopathology. Journal of Behavior Therapy and Experimental Psychiatry, 45, 402–407. doi:10.1016/j.jbtep.2014.04.003Norr, A. M., Oglesby, M. E., Raines, A. M., Macatee, R. J., Allan, N. P., & Schmidt, N. B. (2015). Relationships between cyberchondria and obsessive-compulsive symptom dimensions. Psychiatry Research, 230, 441–446. doi:10.1016/j.psychres.2015.09.034Office for National Statistics. (2018). Internet access - households and individuals, Great Britain: 2018. London: Author.Olatunji, B. O., Etzel, E. N., Tomarken, A. J., Ciesielski, B. G., & Deacon, B. (2011). The effects of safety behaviors on health anxiety: An experimental investigation. Behaviour Research and Therapy, 49, 719–728. doi:10.1016/j.brat.2011.07.008Otto, M. W., Demopulos, C. M., McLean, N. E., Pollack, M. H., & Fava, M. (1998). Additional findings on the association between anxiety sensitivity and hypochondriacal concerns: examination of patients with major depression. Journal of Anxiety Disorders, 12, 225-232. (98) 00011-5.Owens, K.M.B., Asmundson, G.J.G., Hadjistavropoulos,T., & Owens,T.J. (2004). Attentional bias toward illness threat in individuals with elevated health anxiety. Cognitive Therapy and Research,28,57–66. doi:10.1023/B:COTR.0000016930.85884.29Reiss, S. (1987). Theoretical perspectives on the fear of anxiety. Clinical Psychology Review, 7, 585–596. doi:10.1016/0272-7358(87)90007-9Salkovskis, P. M., Rimes, K. A., Warwick, H. M. C., & Clark, D. M. (2002). The Health Anxiety Inventory: Development and validation of scales for the measurement of health anxiety and hypochondriasis. Psychological Medicine, 32, 843–853. doi:10.1017/S0033291702005822Selvi, Y., G?k?e Turan, S., Asena Sayin, A., Boysan, M., & Kandeger, A. (2018). The Cyberchondria Severity Scale (CSS): Validity and reliability study of the Turkish version. Sleep and Hypnosis, 20(4), 241-246. doi: 10.5350/Sleep.Hypn.2018.20.0157 SleepShanahan, C. (2016, April 3). Checking symptoms online leading to growing levels of “cyberchondria.” The Irish Examiner. Retrieved from , M. E., Goetzel, R. Z., Pei, X., Tabrizi, M. J., Ozminkowski, R. J., Gibson, T. B., … Wilson, M. G. (2009). How accurate are self-reports? Analysis of self-reported health care utilization and absence when compared with administrative data. Journal of Occupational and Environmental Medicine, 51, 786–796. doi:10.1097/JOM.0b013e3181a86671Silberg, W. M., Lundberg, G. D., & Musacchio, R. A. (1997). Assessing, controlling, and assuring the quality of medical information on the Internet: Caveant lector et viewor - let the reader and viewer beware. JAMA, 277, 1244–1245. doi:10.1001/jama.1997.03540390074039Singh, K., & Brown, R. J. (2014). Health-related Internet habits and health anxiety in university students. Anxiety, Stress, & Coping, 27, 542–545. doi:10.1080/10615806.2014.888061Singh, K., & Brown, R. J. (2016). From headache to tumour: An examination of health anxiety, health-related Internet use and “query escalation.” Journal of Health Psychology, 21(19), 2008-2020. doi:10.1177/1359105315569620Singh, K., Fox, J.R.E., & Brown, R.J. (2016). Health anxiety and Internet use: A thematic analysis. Cyberpsychology: Journal of Psychosocial Research on Cyberspace, 10(2). doi: 10.5817/CP2016-2-4Starcevic, V. (2013). Hypochondriasis and health anxiety: Conceptual challenges. British Journal of Psychiatry, 202, 7–8. doi:10.1192/bjp.bp.112.115402Starcevic, V., & Aboujaoude, E. (2015). Cyberchondria, cyberbullying, cybersuicide, cybersex: "new" psychopathologies for the 21st century? World psychiatry, 14(1), 97-100. doi:10.1002/wps.20195Starcevic, V., & Berle, D. (2013). Cyberchondria: Towards a better understanding of excessive health-related Internet use. Expert Review of Neurotherapeutics, 13, 205–213. doi:10.1586/ern.12.162Starcevic, V., & Berle, D. (2015). Cyberchondria: An old phenomenon in a new guise? In E. Aboujaoude and V. Starcevic (Eds.), Mental health in the digital age: Grave dangers, great promise (pp. 106 – 117). Oxford: Oxford University Press. Tan, S. S.L., & Goonawardene, N. (2017). Internet health information seeking and the patient-physician relationship: A systematic review. Journal of Medical Internet Research, 19(1), e9. doi:10.2196/jmir.5729Tanis, M. (2008). Health-related on-line forums: What’s the big attraction? Journal of Health Communication, 13, 698–714. doi:10.1080/10810730802415316Tanis, M., Hartmann, T., & te Poel, F. (2016). Online health anxiety and consultation satisfaction: A quantitative exploratory study on their relations. Patient Education and Counseling, 99(7), 1227-1232. doi:10.1016/j.pec.2016.01.021Taylor, H. (2010). The Harris Poll #95: “Cyberchondriacs” on the rise? Retrieved from Poel, F.. Baumgartner, S.E., Hartmann, T., & Tanis, M. (2016). The curious case of cyberchondria: A longitudinal study on the reciprocal relationship between health anxiety and online health information seeking. Journal of Anxiety Disorders, 43, 32-40. doi: 10.1016/j.janxdis.2016.07.009Thorgaard, M.V., Frostholm, L., Walker, L., Jensen, J.S., Morina, B., Lindegaard, H.,...Rask, C.U. (2018). Health anxiety by proxy in women with severe health anxiety: A case control study. Journal of Anxiety Disorders, 52, 8-14. doi: 10.1016/j.janxdis.2017.09.001Warwick, H. M. C., & Salkovskis, P. M. (1990). Hypochondriasis. Behaviour Research and Therapy, 28, 105–117. doi:10.1016/0005-7967(90)90023-CWells, A. (1997). Cognitive thearpy for anxiety disorders: A practice manual and conceptual guide. Chichester: John Wiley & Sons. Wells, A. (2000). Emotional disorders and metacognition: Innovative cognitive therapy. Chichester: John Wiley & Sons. Wells, A.(2009).Metacognitive therapy for anxiety and depression.New York:Guilford Press.White, R. W., & Horvitz, E. (2009a). Cyberchondria: Studies of the escalation of medical concerns in web search. ACM Transactions on Information Systems, 27(4). doi:10.1145/1629096.1629101White, R. W., & Horvitz, E. (2009b). Experiences with web search on medical concerns and self diagnosis. In AMIA Annual Symposium Proceedings (pp. 696–700). Retrieved from , R. W., & Horvitz, E. (2013). Captions and biases in diagnostic search. ACM Transactions on the Web, 7(4). doi:10.1145/2486040Young, K. S. (1998). Internet addiction: The emergence of a new clinical disorder. CyberPsychology & Behavior, 1(3), 237–244. doi:doi:10.1089/cpb.1998.1.237Zhang, Y., Sun,Y., & Xie, B.(2015).Quality of health information for consumers on the Web: A systematic review of indicators, criteria, tools, and evaluation results. Journal of the Association for Information Science and Technology, 66, 2071–2084. doi:10.1002/asi.23311Table 1: Relationship Between Health Anxiety and Extent (Frequency/Duration) of OHRStudySampleAnalysisHealth anxiety measureOHR variable(s)Effect sizeBaumgartner & Hartmann (2011)N = 104. Students at Dutch university and their family/ friendsMultiple regressions (health anxiety = DV) WIFrequency of active online posting of health questions/answers Frequency of online searching for health information β = .45** β = .46 **Doherty-Torstrick et al. (2016)N = 720. American community and student samplet tests comparing OHR in high (WI > 30) and low (WI < 30) health anxiety groups WI Longest episode of OHR in past monthFrequency of OHRd = 1.10** d = 1.21** Fergus (2013)N = 512. Online sample recruited via MTurkCorrelationSHAI (14 item)Frequency of OHRr = .48**Muse et al. (2012)N = 219. Students at British university and previous participants in RCT for health anxiety treatmentMultiple regressions with OHR variables as outcomeSHAI (18 item)Frequency of OHRUsual duration of OHRβ = .06**β = 0.43*Singh & Brown (2014)N = 255. Students at British universityMultiple regressions (DV = health anxiety)SHAI (18 item)Frequency of OHRFrequency of illness-focused searching (composite)Average Duration of OHRFrequency of wellness-focused searching (composite)β = .06β = .32**β = .07β = .13*Tanis et al. (2016)N = 239. Attenders at Dutch medical practiceCorrelationSHAI (18 item)Total duration of searching regarding specific health difficulty r = .15*te Poel et al. (2016)N = 5322. Representative Dutch sampleCross-lagged panel analysisDSHAI Frequency of OHR in past two months (composite)For clinical group (DSHAI > 18)β = .26** For non-clinical group (DSHAI < 18)β = .29*** p < .05, ** p < .01DSHAI = Dutch Short Health Anxiety Inventory; DV = Dependent variable; MTurk = Amazon’s Mechanical Turk; SHAI = Short Health Anxiety Inventory; WI = Whitely IndexTable 2: Relationship between Health Anxiety and Negative Emotional Responses to OHRStudySampleAnalysisEmotion variable(s)Health anxiety measureEffect sizeBarke et al. (2016)N = 368 (Sample A), N = 132 (Sample B)Online samplesCorrelations CSS distress subscale (German translation)mSHAIr = .65**r = .61**,Baumgartner & Hartmann (2011)N = 104Dutch university students and their family/ friendsLogistic regressions (emotion variable = DV) Feeling frightened/overwhelmed/confused/frustrated after most recent episode of OHR. WIOR = 12.52** (DV = frightened)OR = 6.24** (DV = overwhelmed)OR = 4.04** (DV = confused)OR = 3.13* (DV = frustrated)Doherty-Torstrick et al. (2016)N = 720. American community and student sampleMultiple regressions (emotion variable = DV) Anxiety during and post OHRWIβ = .34** (DV = anxiety during OHR)β = .41** (DV = anxiety post-OHR)Fergus (2014)N = 539. Online sample recruited via MTurkCorrelationCSS distress subscaleSHAI (14 item)r = .61**Fergus & Russell, 2016N = 375. Online sample recruited via MTurkMultiple regressionCSS distress subscale (DV)MIHT subscales (IVs)pr = .38** (Affective subscale)pr = .09, p = .15 (Cognitive subscale)pr = -.07, p = .17 (Perceptual subscale)pr = .01, p = .94 (Behavioural subscale)Mathes et al., 2018N = 462. Online sample recruited via MTurkCorrelationCSS distress subscaleSHAI (14 item)r = .49* (Thought intrusion subscale), r = .52* (Fear of illness subscale)Muse et al. (2012)N = 219. British university students and previous participants in health anxiety intervention studyt-tests comparing emotional impact in high (SHAI > 17) and low (SHAI < 17) health anxiety groupsDistress post-OHR; Anxiety about health post-OHRSHAI (18 item)t = -5.25** (Distress post OHR = DV),t = -2.23* (Anxiety about health post OHR = DV)Norr, Allan, et al. (2015)N = 526Online sample recruited via MTurkStructural equation modelCSS distress subscaleSHAI (13 item)β =.37** (association with Thought Intrusion subscale) β =.25** (association with Fear of Illness subscale)Norr, Oglesby et al., 2015N = 468. Online sample recruited via MTurkStructural equation modelCSS distress subscaleSHAI (18 items)β = .52*Selvi et al. (2018)N = 337 Turkish university studentsCorrelationsCSS distress subscale (Turkish translation)HAI (Turkish translation)r = .32**Singh & Brown (2014)N = 255 British university studentsMultiple regression with health anxiety as outcomeTension post-OHRSHAI (18 item)β = .27*** p < .05; ** p < .01CSS = Cyberchondria Severity Scale; DV = Dependent variable; HAI = Health Anxiety Inventory; IV = Independent variable; mSHAI = Modified version of SHAI, translated into German; OR = odds ratio; SHAI = Short Health Anxiety Inventory; WI = Whitely IndexTable 3: Studies of Reassurance-Seeking Behaviour Associated with or Following OHRStudySampleVariablesAnalysisResultBarke et al., (2016)N = 500. Online sampleCSS reassurance subscale (German translation)mSHAICorrelationr = .28**Fergus (2014)N = 539. Online sample recruited via MTurkCSS reassurance subscale SHAI (14 item)Correlationr = .33*Fergus (2015)N = 578. Online sample recruited via MTurkCSS reassurance subscale WICorrelationr = .38**Fergus & Russell (2016)N = 375. Online sample recruited via MTurkCSS reassurance subscale (DV)MIHT subscales (IVs)Multiple regressionpr = .08, p = .15 (Affective subscale)pr = .05, p = .36 (Cognitive subscale)pr = .03, p = .59 (Perceptual subscale)pr = .09, p = .13 (Behavioural subscale)Norr, Oglesby, et al. (2015)N = 468Online sample recruited via MTurkCSS reassurance subscale SHAI (18 item)Structural equation modelβ = .41**Singh & Brown (2014)N = 255. Students at British universityLikelihood of post-OHR doctor utilisation (single item; DV)SHAI (18 item; IV)Linear regressionβ = .15, p = .006White & Horvitz (2009b)N = 515. Microsoft EmployeesExperiencing increased anxiety as a result of OHR (yes/no)Seeking professional medical attention on the basis of information found online (yes/no)Chi squareThose who experienced increased anxiety significantly more likely to report seeking medical attention on the basis of online information (p < .01)* p < .05, ** p < .01DV = Dependent variable; CSS = Cyberchondria Severity Scale; IV = Independent variable; mSHAI = modified version of SHAI, translated into German; MTurk = Amazon’s Mechanical Turk; SHAI = Short Health Anxiety Inventory; WI = Whitely Index-1714503556000Figure 1. Hybrid model of reassurance seeking and compulsive health-related Internet use (cyberchondria). Suggested by, and adapted from, Starcevic and Berle (2013).-26670021653500Figure 2. Overview of systematic search strategy and results42729805638800412115-127000Figure 3. Proposed processes driving the decision to engage in online health research (OHR) and how this can develop into compulsive OHR when worrying outcomes lead to the development of negative beliefs about the dangers of searching online and the person’s ability to control their use. The model applies regardless of whether someone already has a tendency to be health anxious, although pre-existing health anxiety is assumed to moderate several of the processes described here, including the tendency to perceive health threat in the first place, the perceived need to engage in self-regulation, the likelihood of interpreting online information as worrying, the tendency to escalate queries, and the presence of positive and negative meta-beliefs about OHR. We assume that compulsive OHR can be triggered by the perception of health threat (link not shown for clarity) although it can also occur in its absence.432261854864000 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download