Excessive reassurance seeking and anxiety pathology: Tests of incremental associations and directionality

Abstract

Excessive reassurance-seeking (ERS) is hypothesized to play a key role in emotional disorders but has been studied mostly in relation to depression. Study 1 reports a new measure of reassurance seeking that assessed ERS related to general and evaluative threats in a non-clinical student sample, and its factor structure was further examined in Study 2. In Study 3, the scale, along with other symptom-related measures and an existing measure of depressive ERS, was administered to an undergraduate sample at two different time points, one month apart. Greater ERS was associated with greater symptoms of social anxiety, generalized anxiety disorder, and obsessive–compulsive disorder (OCD), even after controlling for trait anxiety, depression, and intolerance of uncertainty. Among OCD symptoms, only thoughts of harm were uniquely related to ERS, a finding consistent with emerging literature. ERS involving general threats also predicted changes in social anxiety and GAD symptoms one month later. Overall, the findings implicate an important role for ERS across anxiety disorders.

Introduction

Anxiety disorders constitute a significant public health burden. Recent epidemiological evidence suggests they are much more prevalent than traditionally thought, affecting up to half of the general population (Moffitt et al., 2010). They also generally maintain a chronic course when left untreated (Pine, Cohen, Gurley, Brook, & Ma, 1998) and result in substantial impairment across the lifespan (Ferdinand, van der Reijden, Verhulst, Nienhuis, & Giel, 1995). It is therefore noteworthy that cognitive-behavioral therapy (CBT) has demonstrated impressive effectiveness in the treatment of anxiety disorders, with rates of improvement ranging from 60 to 90% (Norton & Price, 2007). The rise of CBT in the treatment of anxiety disorders has led to the development of many disorder-specific treatment protocols. This development has been viewed as problematic by some who criticize it for being inefficient and as negatively impacting treatment dissemination (Barlow et al., 2004, Cougle, in press).

An approach to developing more efficient and parsimonious therapies is to focus on core processes occurring across anxiety disorders. One transdiagnostic process is safety behavior; safety behaviors are actions intended to avoid or cope with perceived threat (Salkovskis, Clark, & Gelder, 1996). They may take different forms depending on the anxiety disorder. For example, individuals with social anxiety disorder may grip objects tightly to avoid shaking while reading to a group or rehearse sentences in their mind and speak quickly to counter the fear of talking inappropriately (Wells et al., 1995). Individuals with panic disorder may carry safety aids such as a cell phone, water, or medication, or check for exits and restrooms (Rachman, 1984). There is evidence to suggest that safety behaviors are important maintaining factors in anxiety disorders (Helbig-Lang & Petermann, 2010). They may act to prevent threat disconfirmation through misattribution of safety to the safety behavior itself rather than the harmless nature of the stimulus or situation. Safety behaviors may also tax attentional resources that would be necessary to attend to and process disconfirming information (Sloan & Telch, 2002). Additionally, safety behaviors may actually exacerbate anxiety symptoms and lead to threat overestimation (Deacon & Maack, 2008), perhaps by focusing excessive attention on perceived threats.

One form of safety behavior that has received some research attention is excessive reassurance seeking (ERS). It has been discussed extensively in relation to health anxiety (e.g., seeking reassurance from health professionals) and plays a central role in cognitive models of this disorder (Salkovskis & Warwick, 1986). It has also been described in patients with generalized anxiety disorder (GAD) (Woody & Rachman, 1994). ERS related to general threats and social threats was also found in an interview-based study of obsessive–compulsive disorder (OCD) patients (Parrish & Radomsky, 2010). It was also more likely to be endorsed as a strategy used for coping with negative intrusions among individuals with OCD compared to depressed individuals and anxious and non-clinical controls (Morillo, Belloch, & García-Soriano, 2007). Rachman (2002) hypothesized that, for compulsive checkers, ERS is used to reduce anxiety, prevent harm, and decrease responsibility for harm.

Most psychopathology-related research on ERS has been conducted in relation to depression (e.g., Joiner & Metalsky, 2001). According to Coyne's (1976) interpersonal theory, the individual with depression has a tendency to seek reassurance to alleviate doubts that other people truly care about them. Others may provide reassurance, but the depressed individual doubts their sincerity. Reassurance seeking continues, assurance is given but doubted, and a negative spiral results. Those who are close to the depressed individual become increasingly likely to reject him or her, which further disrupts the depressed person's environment and maintains or exacerbates his or her symptoms.

The type of ERS described above has been found to predict future depressive symptoms (Joiner & Metalsky, 2001) and depressive, but not anxious, reactions to acute stress (Joiner & Schmidt, 1998). It has been given little consideration in relation to symptoms of different anxiety disorders. Further, ERS investigated across these studies does not consider the threat-related nature of reassurance-seeking that is often reported by individuals with anxiety disorders. When assessing the role of ERS in anxiety disorders, additional motivations and contexts that may drive such behavior are important to consider, especially threat-related motivations. Thus, in order to better understand ERS and anxiety pathology, our first study focuses on assessing ERS in a manner that includes threat-related motivations.

Several explanations could account for relations between ERS and anxiety pathology. First, ERS might be a consequence of anxiety pathology. According to this view, the anxious individual seeks the assurance of others that danger is not imminent and situations or stimuli are safe. Such behavior may be carried out to reduce anxiety and risk of harm. Second, ERS and anxiety may be related because of their joint association with depressive symptoms (Joiner & Schmidt, 1998). Third, anxious individuals may also engage in ERS to reduce unbearable feelings of uncertainty; thus, ERS may simply be a consequence of intolerance of uncertainty. Intolerance of uncertainty has been implicated as an important construct in anxiety pathology and figures prominently in cognitive conceptualizations of GAD (Dugas, Gosselin, & Ladouceur, 2001). Greater intolerance of uncertainty has also been linked to OC checking (Tolin, Abramowitz, Brigidi, & Foa, 2003) and social anxiety (Boelen & Reijntjes, 2009). Finally, it is possible that ERS plays a maintaining or causal role in anxiety pathology. It may function similarly to other safety behaviors in that it prevents threat disconfirmation (Salkovskis, 1991). ERS may also contribute to anxious symptoms by focusing increased attention on perceived threat and decreasing self-confidence to cope with perceived threats.

It is difficult to choose among the possible explanations described above because of important gaps in the literature regarding the role of ERS in anxiety pathology. To date, ERS has been described among different clinical groups, but it has been assessed quantitatively primarily in the context of health-related behaviors specific to hypochondriasis (Salkovskis, Rimes, Warwick, & Clark, 2002). Its unique relations with other forms of anxiety pathology have not been explored. Rival constructs that may explain the proposed relations between ERS and anxiety pathology, including depression and intolerance of uncertainty, have been given little consideration. In addition, the directionality of the ERS and anxiety association and the potential role of ERS in predicting future anxiety pathology have not been tested.

Based on clinical reports and existing theory and research (e.g., Parrish and Radomsky, 2010, Rachman, 2002, Salkovskis and Warwick, 1986), we propose that two additional forms of ERS may be related to anxiety pathology. The first type of ERS, related to general threats, is carried out to receive assurance from others that negative outcomes will not occur. The second type is more self-focused and evaluative in nature and is carried out so that the individual is assured that others do not think negatively of him or her. We hypothesize that the first type is prevalent across anxiety disorders and the second type is more central to social anxiety disorder.

The current research sought to clarify the role of ERS in anxiety pathology. Specifically, it had four separate aims: (1) to develop a means of assessing reassurance-seeking related to general threats and threats of negative evaluation; (2) to examine the associations of multiple reassurance-seeking dimensions, including depressive reassurance seeking, with self-report measures of GAD, social anxiety, and OCD; (3) to examine whether depressive symptoms, trait anxiety, and intolerance for uncertainty account for the relations between ERS and anxiety disorder symptoms; and (4) to test whether ERS tendencies predict anxiety disorder symptoms prospectively and vice versa. We also conducted exploratory analyses of the role of gender in moderating the reassurance seeking and anxiety pathology relationship.