The effects of safety-seeking behavior and guided threat reappraisal on fear reduction during exposure: an experimental investigation

Abstract

We examined the effects of safety-seeking behavior and guided threat focus and reappraisal on fear reduction during exposure. Participants (N=46) displaying marked claustrophobic fear were randomized to one of three 30-min exposure conditions: (a) guided threat focus and reappraisal; (b) safety-behavior utilization; or (c) exposure only control. Tripartite outcome assessments during a behavioral approach test, along with measures of suffocation and restriction fears were obtained at pre- and post-treatment, and at a 2-week follow-up. Treatment process measures were collected throughout treatment and consisted of indices of fear activation; within and between-trial fear habituation; and suffocation and entrapment expectancies. Measures of safety behavior utilization and attentional focus were also collected to assess the integrity of the experimental manipulations. Consistent with prediction, those encouraged to utilize safety-behaviors during exposure showed significantly more fear at post-treatment and follow-up relative to those encouraged to focus and reevaluate their core threat(s) during exposure. Moreover, growth curve analyses of treatment process data analyses revealed that safety-behavior utilization exerted a detrimental effect on between-trial habituation; whereas guided threat reappraisal enhanced between-trial habituation.

Introduction

The concept of safety has played an increasingly important role in our understanding of pathological fear and its modification. The ubiquitous nature of safety-seeking behavior in anxiety disorders has been well documented, although specific patterns may vary. People with social phobia, for example, will mentally rehearse sentences to counter a fear of talking funny; eat small amounts of food to counter a fear of vomiting; or avoid holding an object to counter a fear of shaking uncontrollably (Wells et al., 1995). Similarly, patients with panic disorder and agoraphobia will engage in a variety of safety behaviors such as check for the presence of hospitals, bathrooms, and exits; carry safety aids such as water, medications, or cellular phones; and avoid activities that elicit strong physical sensations, for example, caffeine, exercise, or alcohol (Kamphuis & Telch, 2000; Rachman, 1983, Rachman, 1984). Those with generalized anxiety disorder will repeatedly seek reassurance from others, insist on regular and frequent contact with family, avoid risks, and engage in checking and overprotective behavior (Woody & Rachman, 1994).

Recognition of the linkage between safety behaviors and anxiety disorders has led to theorizing on the role of safety behaviors in the maintenance and modification of pathological fear. As early as 1974, Bandura, Jeffery and Wright emphasized the importance of fading patients' response aids during exposure to enhance their one's sense of personal mastery. Rachman (1984) proposed a safety signal theory based on Gray's (1971) elaboration of Mower's (1960) two-stage theory. Rachman proposed that the pairing of safety cues with feared stimuli could be used therapeutically to enhance motivation for regular exposure practice thus facilitating long-term reductions in fear and avoidance. Included in the theory are specific factors that enhance or reduce a sense of safety.

Rachman's safety signal perspective has been outlined in the analysis and treatment of both agoraphobia (Rachman, 1983) and generalized anxiety disorder (Woody & Rachman, 1994). In treating agoraphobia, patients would be encouraged to travel towards rather than away from their safety signals (e.g., home). Although acknowledging that this may initially strengthen the reliance on the safety signal, Rachman (1983) believes argued that the opportunity to establish new safety signals might outweighs any adverse effects of increased reliance on the original safety signal.

In support of the immediate fear reducing effects of safety cues, it has been found that they can aid in the reduction of initial anxiety while the safety cues are available. For instance, during CO2 provocation, the presence of a safe person decreases panic patients' subjective anxiety, whereas the absence of a safe person increases anxiety (Carter, Hollon, Carson, & Shelton, 1995). Moreover, those who are provided safety information or safety cues are less likely to experience heightened fear in response to biological challenges (Rapee, Telfer, & Barlow, 1991, Telch, Silverman, & Schmidt, 1996, Schmidt & Telch, 1994).

Although the provision of safety cues may reduce the immediate experience of fear, the use of safety-seeking behaviors in response to phobic situations that pose no real threat might ultimately contribute to fear maintenance (Salkovskis, 1991). How might the utilization of safety-seeking behaviors contribute to the maintenance of pathological fear? One possibility is that safety-seeking behaviors prevent or weaken threat disconfirmation through a misattribution of safety (Salkovskis, 1991). Anxious patients erroneously attribute their failure to be harmed (safety) to their judicious use of one or more safety-seeking behaviors, thus leaving the original faulty threat perception unscathed (Salkovskis, 1991, Telch, 1991).

In order for safety-seeking behavior to affect threat perception, one would expect a linkage between safety-seeking behaviors and threat-relevant cognitions. Such was found to be the case in a recent study showing that the safety-seeking behaviors of 147 panic disorder patients were meaningfully related to their perceived threats (Salkovskis, Clark, & Gelder, 1996). These findings are consistent with the hypothesis that safety behaviors function to help patients avert a perceived threat.

Preliminary support for the deleterious effects of safety-behavior utilization (SBU) on fear reduction during exposure comes from two studies (Wells et al., 1995; Salkovskis, Clark, Hackman, Wells, & Gelder, 1999). In the first study (Wells et al., 1995), significantly greater anxiety reduction and cognitive change were observed among eight social phobics instructed to refrain from using safety behaviors during exposure.

Salkovskis et al. (1999) randomly assigned 18 patients with panic disorder with agoraphobia to either 15 min of situational exposure with safety-behavior fading plus a disconfirmation rationale (experimental condition), or situational exposure without safety-behavior fading (exposure control). Compared to patients receiving 15 min of exposure, patients who were instructed to withdraw their safety-behaviors during exposure reported significantly greater reduction in subjective fear. These data support the hypothesis that exposure can be more effective when in-situation safety-seeking behaviors are identified and eliminated.

Several limitations deserve mention. First, neither study included manipulation checks to determine patients' actual use of safety behaviors during treatment. Second, therapeutic rationale was confounded with safety behavior fading, namely, a disconfirmation rationale was provided to patients in the safety-behavior fading conditions, whereas an extinction rationale was provided to patients in the exposure control conditions. Third, since patients in the safety-behavior fading condition were instructed to focus on their perceived threat in addition to refraining from using safety behaviors, it is unclear whether focusing on fears or fading of safety behaviors accounted for the greater fear reduction observed in the safety-behavior fading condition.

The present study sought to clarify the effects of safety-seeking behavior on fear reduction during exposure. Students displaying marked claustrophobia were randomly assigned to one of three exposure conditions: (a) exposure with guided threat focus and reappraisal (GTR); (b) exposure with SBU; and (c) exposure only control (CRL). All three groups received 30 min of self-guided exposure. Participants in the GTR condition were instructed to focus on their fear-relevant threats during each treatment trial and to test the validity of their threat perceptions. This treatment element has been previously shown to enhance fear reduction relative to an CRL condition (Kamphuis & Telch, 2001). Those in the SBU condition were provided several specific safety strategies for managing their fear during exposure (i.e., checking the door latch to insure it was unlocked, opening an air portal to let fresh air into the chamber, speaking to the experimenter through an intercom). Those in the CRL condition received the same amount of self-guided exposure but without access to safety strategies and without specific instructions to focus and reappraise their perceived threat. To assess the integrity of the experimental manipulations, data were collected on the pattern and frequency of SBU as well as attentional focus during exposure. We hypothesized that relative to those who had no access to these safety strategies, participants in the safety-behavior condition would display less fear reduction pre- to post-treatment and greater return of fear at follow-up. Moreover, we hypothesized that compared to exposure without threat focusing and reappraisal; exposure with threat focusing and reappraisal would result in greater fear reduction and less return of fear.