Flipbook

1 INTRODUCTION 13 SAFETY BEHAVIOR AND THE MAINTENANCE OF PATHOLOGICAL ANXIETY Clinical guidelines generally recommend to motivate patients to inhibit all safety behaviors during exposure therapy (e.g., Abramowitz, Deacon, & Whiteside, 2011; Craske et al., 2014; Keijsers, van Minnen, & Hoogduin, 2011). The theoretical justification for this guideline is that safety behaviors impede extinction. Salkovskis (1991) proposed that anxious individuals misattribute the nonoccurrence of the expected catastrophe to the safety behavior, which leaves the core threat belief unaffected (the ‘misattribution of safety’ hypothesis). For example, the patient with panic disorder may believe that he did not faint because he quickly sat down, and the patient with OCD may attribute the absence of illness to her careful hand washing. In addition to the misattribution of safety to the behavior, safety behavior prevents the occurrence of disconfirming learning experiences. By sitting down, the patient with panic disorder never experiences that he would not have fainted if he had remained standing, and washing her hands precludes the patient with OCD from learning that touching doorknobs without subsequent handwashing does not result in infection and illness either. Safety behavior prevents the disconfirmation of the CS – US association, and thus prevents inhibitory learning. As a result, the innocuous stimulus or situation (i.e., the CS) retains its subjective threat value. In other words, dizziness and doorknobs remain fearful to the patient with panic disorder and OCD, respectively. Finally, the extent to which inhibitory learning occurs depends on the discrepancy between the expected outcome (i.e., the feared catastrophe) and the actual outcome (i.e., the nonoccurrence of this catastrophe; Bouton, 2004). Hence, a larger violation of negative expectancies results in more inhibitory learning (Baker et al., 2010). When safety behavior reduces the perceived likelihood or severity of the expected catastrophe, it decreases this discrepancy, and thus hampers inhibitory learning (Blakey & Abramowitz, 2016; Craske et al., 2014).