Panic Disorder 305
matic physical threat appraisals and more realistic reinterpretations would be a fruitful
area of research.
Hypothesis 6. Safety Seeking
Avoidance and maladaptive safety- seeking behavior will intensify anxiety and panic
symptoms in those with panic disorder relative to nonpanic controls.
Since safety seeking is any cognitive or behavioral strategy that is intended to pre-
vent or minimize a feared outcome, it includes escape as well as all forms of avoidance
(Salkovskis, 1988, 1996b). Any cognitive or behavioral strategy (e.g., controlled breath-
ing, relaxation, sitting, being accompanied by a friend, distraction) that subverts access
to information that would disconfirm the catastrophic belief is considered maladaptive
and will contribute to the persistence of panic symptoms (D. M. Clark, 1999; Salkovs-
kis, 1988). Lohr et al. (2007) argue that safety signals may reduce the immediate experi-
ence of fear but ultimately contribute to the maintenance of pathological fear.
As noted previously, there is an extensive research literature showing that safety-
seeking behavior and avoidance contribute to the persistence of anxiety (see Chapter 3,
Hypotheses 2 and 10). Schmidt et al. (2006), for example, found that the provision of
safety cues undermined the effectiveness of safety information in reducing anxiety to
a CO 2 inhalation challenge, whereas other studies found a strong link between safety
behaviors and catastrophic misinterpretations (e.g., Salkovskis et al., 1996). Lundh et
al. (1998) found that a recognition bias for safe faces correlated with avoidance of fear
situations, which suggests a strong link between pursuit of safety and avoidance. In a
naturalistic self- monitoring study, Radomsky, Rachman, and Hammond (2002) found
that individuals with panic disorder used a variety of safety- seeking strategies to has-
ten the end of a panic attack, the most common being an effort to calm down. Indi-
viduals believed that these strategies helped terminate panic as indicated by a reduction
in bodily sensations and fearful cognitions. Even though individuals believed that the
safety- seeking behaviors made them feel somewhat safe and reduced the likelihood of
another immediate panic attack, Radomsky and colleagues employed a repeated hyper-
ventilation challenge test to show there was, in fact, no panic-safe refractory period.
Thus individuals may believe that safety seeking helps terminate a panic attack and
reduce the likelihood of an immediate recurrence even though the prophylactic effect of
safety seeking is highly unlikely.
Furthermore, there is evidence that a reduction in safety seeking can have positive
therapeutic effects on anxiety and panic symptoms (see Salkovskis et al., 1999; Salk-
ovskis et al., 2006). And yet Rachman found that the provision of safety- signal training
increased predictions of safety, reduced expectations of fear, and inhibited panic when
individuals were exposed to their fear situations (Rachman & Levitt, 1985; Rachman,
Levitt, & Lopatka, 1988b). Similarly Milosevic and Radomsky (2008) found that snake-
fearful individuals had significant reductions in subjective anxiety and fearful cognitions
as well as increased approach behavior with a single 45-minute exposure session whether
or not they were allowed to rely on safety behavior during the exposure session.
In summary, research on safety seeking indicates that a distinction must be made
between safety- seeking behavior and feelings of safety. Producing a sense of safety