Does cyberchondria overlap with health anxiety and obsessive–compulsive symptoms? An examination of latent structure and scale interrelations
Introduction
Looking for health-related information online is a frequent behavior that is often accompanied by an initial sense of relief (White & Horvitz, 2009). Despite this fact, researchers have described a phenomenon named cyberchondria, in which Internet searches for medical information exacerbate distress and the searching behavior continues despite the related anxiety (Taylor & Asmundson, 2004). Factor analytic studies of the chief self-report measure of cyberchondria (i.e., the Cyberchondria Severity Scale; McElroy & Shevlin, 2014) suggest that this phenomenon consists of four components: repetition, exacerbation of distress, interruption of daily living, and generation of further reassurance seeking (Fergus, 2014; Norr, Allan, Boffa, Raines, & Schmidt, 2015).
To improve our understanding of cyberchondria, researchers have examined its interrelations with overlapping symptom types. For example, Starcevic and Berle (2013) proposed that cyberchondria is a behavior that typifies the wide range of worry individuals can have about their health, a concept referred to as health anxiety (Asmundson & Taylor, 2005). Extant studies have found that cyberchondria and health anxiety share moderate to strong correlations (.33–.62; Fergus, 2014, Fergus, 2015; Norr, Albanese, Oglesby, Allan, & Schmidt, 2015; Norr, Oglesby et al., 2015). Although informative, a limitation of these prior studies is that the researchers assessed health anxiety by using a total score. Norr, Allan et al. (2015) addressed this limitation, finding relations between cyberchondria and two components of health anxiety (thought intrusion and fear of illness) assessed using the Short Health Anxiety Inventory (Salkovskis, Rimes, Warwick, & Clark, 2002). As an extension of these findings, it would be informative to examine whether cyberchondria differentially relates to components of health anxiety represented within cognitive-behavioral models. Health anxiety is comprised of affective (worry about health), cognitive (dysfunctional beliefs about health), perceptual (hypervigilance to physical sensations), and behavioral (avoidance behavior, typically reassurance seeking) components within those models (Taylor & Asmundson, 2004). The Multidimensional Inventory of Hypochondriacal Traits (Longley, Watson, & Noyes, 2005) is a measure that allows for the assessment of each of these four components. Worry and reassurance seeking are central characteristics of repetitive, online searching behavior for medical information (McElroy and Shevlin, 2014, Starcevic and Berle, 2013). It is thus possible that cyberchondria may cluster with the affective and behavioral components of health anxiety.
In addition to health anxiety, Starcevic and Berle (2013) described the intersection between cyberchondria and obsessive–compulsive symptoms. In particular, they proposed that cyberchondria is marked by obsessional doubt and parallels compulsions that are evident in obsessive–compulsive disorder (OCD). Consistent with this proposal, Fergus (2014) found a moderate correlation (r = .49) between cyberchondria and obsessive–compulsive symptoms. A limitation of Fergus’s study was the assessment of obsessive–compulsive symptoms using a total score, as there are four core obsessive–compulsive symptoms dimensions within contemporary conceptualizations and self-report measures. These symptom dimensions include: contamination, responsibility for harm, unacceptable thoughts, and symmetry (Abramowitz et al., 2010, McKay et al., 2004). Norr, Oglesby et al. (2015) extended Fergus’s study and found that cyberchondria correlated with each of these four obsessive–compulsive symptom dimensions (rs ranging from .28 to .55). Based upon their findings, Norr, Oglesby et al., speculated that the repeated searches for medical information may function as a safety behavior to alleviate either contamination concerns or responsibility for harm.
The extant literature suggests that cyberchondria overlaps meaningfully with both health anxiety and obsessive–compulsive symptoms. However, it remains unclear whether that overlap is suggestive of potential redundancy among these constructs. For example, the recurrent internet searches for medical information that characterize cyberchondria may represent a behavioral component of health anxiety (e.g., reassurance seeking) or a neutralizing effort (e.g., checking) related to obsessive–compulsive symptomology. Confirmatory factor analyses (CFAs) are one way to examine whether constructs are distinct from one another (e.g., Wu, 2011). For example, CFAs can indicate whether intercorrelations among indicators are best represented by the indicators loading on the same or separate latent constructs. Finding indicators of cyberchondria, health anxiety, and obsessive–compulsive symptoms to be best modeled as loading on three separate latent constructs would suggest the respective measures appear to be assessing distinct constructs rather than simply different components of the same construct.
We completed the present study to further examine interrelations among cyberchondria, health anxiety, and obsessive–compulsive symptoms. We used CFAs to examine the distinguishability of cyberchondria from health anxiety and obsessive–compulsive symptoms at the latent level. Prior research has found that cyberchondria, health anxiety, and obsessive–compulsive symptoms are associated. Yet, their scale correlations suggest a substantial amount of non-overlapping variance. It was thus predicted that a measurement model depicting the indicators of cyberchondria as independent from, albeit related to, the indicators of health anxiety and obsessive–compulsive symptoms would provide the best fit to the data among competing measurement models. These predicted findings would support cyberchondria overlapping with, and yet being discernible from, both symptom types.
Whereas the CFAs focused on relations among higher-order constructs, we next sought to identify whether cyberchondria differentially relates to the lower-order components of health anxiety and obsessive–compulsive symptoms. We predicted that cyberchondria would cluster with the affective and behavioral components of health anxiety because worry and reassurance seeking are important characteristics of recurring online searching behavior for medical information (McElroy and Shevlin, 2014, Starcevic and Berle, 2013). We also sought to replicate Norr, Oglesby et al. (2015) findings that cyberchondria clusters with the obsessive–compulsive symptom dimensions of contamination and responsibility for harm. These predictions were based upon raw correlations.
We then sought to replicate the pattern of raw correlation findings when examining the components of health anxiety and obsessive–compulsive symptoms that shared unique associations with cyberchondria. Examining unique associations is important, as cyberchondria may evidence a raw association with health anxiety or obsessive–compulsive symptoms as a result of shared variance with the other symptom type. Negative affect, which reflects the tendency to experience negative emotional states (Watson, Clark, & Tellegen, 1988), may additionally contribute to observed raw associations. Indeed, negative affect is considered a non-specific factor that contributes to all manifestations of anxiety (Clark and Watson, 1991). Norr, Oglesby et al. (2015) previously examined unique associations between obsessive–compulsive symptom dimensions and cyberchondria while accounting for health anxiety and negative affect. The present study extends Norr, Oglesby et al., work by being the first of its kind to examine unique associations between cyberchondria and the four components of health anxiety represented within cognitive-behavioral models. Finally, we completed exploratory analyses to examine whether the same components of health anxiety and obsessive–compulsive symptoms shared unique associations with each of the four components of cyberchondria.
The predicted findings were expected to be associated with important conceptual and therapeutic implications. For example, identifying specific components of health anxiety and obsessive–compulsive symptoms that cluster with cyberchondria may further our understanding of overlapping features. That knowledge could ultimately help advance treatment efforts for cyberchondria, as interventions shown to reduce the respective components of health anxiety or obsessive–compulsive symptoms may also be useful in decreasing cyberchondria. Potential intervention strategies for cyberchondria have yet to be examined (Starcevic & Berle, 2013).
Section snippets
Participants
The sample consisted of 375 adults located in the United States that were recruited through an online crowdsourcing website. The mean age was 31.6 years (SD = 10.2; ranging from 19 to 64) and a slight majority were male (52.7%). In terms of racial identification, 73.9% of the sample self-identified as White, 8.8% as Black, 6.7% as Asian, 6.4% as Latino, 3.2% as multi-racial, and 1.0% as “other.” A majority of the sample reported receiving a two-year college degree or higher (60.1%), as currently
CFAs
Upon our initial examination of the correlated three-factor model, we found its model fit to be equivocal [χ2(51) = 208.1, p < .001; CFI = .95; NNFI = .93; RMSEA (90% CI) = .09 (.08–.11); SRMR = .07]. While the CFI and SRMR met specified guidelines, the NNFI, RMSEA, and RMSEA 90% CI did not meet these same regulations. When examining the parameter estimates, the perceptual indicator of health anxiety demonstrated a low factor loading (.27) on the health anxiety construct. This finding is consistent with
Discussion
Results from CFAs support the presence of meaningful relations between cyberchondria and both health anxiety and obsessive–compulsive symptoms, as a correlated measurement model provided a better fit to the data than did an uncorrelated measurement model. These results are consistent with proposals (Starcevic & Berle, 2013) and preliminary data (Fergus, 2014, Fergus, 2015, Norr et al., 2015, Norr, Allan et al., 2015, Norr, Oglesby et al., 2015) linking cyberchondria to both symptom types.
References (37)
The Cyberchondria Severity Scale (CSS): an examination of structure and relations with health anxiety in a community sample
Journal of Anxiety Disorders
(2014)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
(2015)- et al.
A taxometric study of hypochondriasis symptoms
Behavior Therapy
(2010) - et al.
The development and initial validation of the cyberchondria severity scale (CSS)
Journal of Anxiety Disorders
(2014) - et al.
A critical evaluation of obsessive–compulsive disorder subtypes: symptoms versus mechanisms
Clinical Psychology Review
(2004) - et al.
Anxiety sensitivity and intolerance of uncertainty as potential risk factors for cyberchondria
Journal of Affective Disorders
(2015) - et al.
Validation of the Cyberchondria Severity Scale (CSS): Replication and extension with bifactor modeling
Journal of Anxiety Disorders
(2015) - et al.
Relationships between cyberchondria and obsessive–compulsive symptoms
Psychiatry Research
(2015) - et al.
Psychological treatments of health anxiety & hypochondriasis: a biopsychosocial approach
(2008) - et al.
Assessment of obsessive–compulsive symptom dimensions: development and evaluation of the dimensional obsessive–compulsive scale
Psychological Assessment
(2010)