Clinical Psychology Review 24 (2004) 529 – 555
An integrated cognitive model of panic disorder: The role of positive and negative cognitions Leanne M. Casey a,b,*, Tian P.S. Oei a, Peter A. Newcombe a b
a School of Psychology, University of Queensland, Brisbane, Queensland 4072, Australia Department of Psychology, The Prince Charles Hospital Health Service District, Brisbane, Queensland 4032, Australia
Received 18 November 2002; received in revised form 5 January 2004; accepted 12 January 2004
Abstract One reason for the neglect of the role of positive factors in cognitive – behavioural therapy (CBT) may relate to a failure to develop cognitive models that integrate positive and negative cognitions. Bandura [Psychol. Rev. 84 (1977) 191; Anxiety Res. 1 (1988) 77] proposed that self-efficacy beliefs mediate a range of emotional and behavioural outcomes. However, in panic disorder, cognitively based research to date has largely focused on catastrophic misinterpretation of bodily sensations. Although a number of studies support each of the predictions associated with the account of panic disorder that is based on the role of negative cognitions, a review of the literature indicated that a cognitively based explanation of the disorder may be considerably strengthened by inclusion of positive cognitions that emphasize control or coping. Evidence to support an Integrated Cognitive Model (ICM) of panic disorder was examined and the theoretical implications of this model were discussed in terms of both schema change and compensatory skills accounts of change processes in CBT. D 2004 Elsevier Ltd. All rights reserved. Keywords: Panic disorder; Positive cognitions; Catastrophic misinterpretation of bodily sensations; Panic self-efficacy
1. Introduction Although a number of reasons have been advanced to explain the relative neglect of positive cognitions in cognitive–behavioural therapy (CBT) (MacLeod & Moore, 2000), there has been little recognition that this problem may also relate to the continuing lack of integration between existing theoretical models within the cognitive framework. Bandura’s (1977) identification of self-efficacy * Corresponding author. School of Psychology, University of Queensland, Brisbane, Queensland 4072, Australia. Tel.: +617-3365-6230; fax: +61-7-3365-4466. E-mail address:
[email protected] (L.M. Casey). 0272-7358/$ - see front matter D 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.cpr.2004.01.005
530
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
beliefs as positive cognitions that mediate a range of emotional and behavioural outcomes generated influential support for cognitive models of human functioning (Maddux, 1995). To date, however, the primary emphasis in cognitive therapy has been upon identification and modification of negative cognitions. Integration of such models may provide both a more coherent account of change processes in CBT as well as establish the basis for further development in the field. Panic disorder provides a useful exemplar of a disorder in which development of an Integrated Cognitive Model (ICM) may provide important momentum to continuing research efforts. Cognitive theorists, such as Clark (1993b), have argued that changes in positive cognitions occur simply as a corollary to changes in negative, danger-related cognitions. In contrast, other theorists, such as Bandura (1988) and Barlow (1988), who have drawn attention to the importance of control or coping in panic disorder, have in turn relegated negative cognitions to a subordinate role. To date, much of the research has focused on negative cognitions (Cox, 1996) and there has been little attempt to integrate these positions. As the following review indicates, however, there are limitations to a cognitive account of panic disorder that focuses solely on negative cognitions, which in turn may constrain attempts to demonstrate cognitive mediation of treatment effects. In contrast, a cognitive model that integrates both negative and positive cognitions may not only assist treatment and research in panic disorder, but also provide a useful template for investigation of other psychological disorders.
2. Cognitive approaches to panic disorder 2.1. Clark’s explanation of panic disorder: catastrophic misinterpretation of bodily sensations Clark’s explanation of panic is illustrated in Fig. 1. According to Clark (1986), ‘‘panic attacks result from the catastrophic misinterpretation of certain bodily sensations’’ (p. 462). These bodily sensations, he suggested, were mostly those associated with normal anxiety responses, such as palpitations, breathlessness, and dizziness, but could also be caused by routine events (e.g., exercise) or even result from other, non-anxiety-related emotional states (e.g., excitement, anger, and happiness). Through the process of catastrophic misinterpretation, however, the individual perceives these essentially benign and normal sensations as evidence of imminent danger. Thus, palpitations are misinterpreted as a signal of a heart attack, dizziness as evidence of impending loss of control, etc. Noting that a broad range of stimuli (either external or internal) can produce panic attacks, Clark suggested that the perception of these stimuli as threatening results in ‘‘a mild state of apprehension,’’ which is accompanied by the bodily sensations associated with anxious arousal. Clark’s explanation of panic disorder thus postulates a vicious circle in which misinterpretation of these bodily sensations of anxiety results in a panic attack, although as noted, sensations associated with other causes may be involved in the initial stages. Initially, Clark (1988) argued that panic attacks always result from a catastrophic misinterpretation of internal or external events, although for some individuals, this process may occur out of awareness at the time. However, in a more recent paper, Clark (1996) has focused on the key importance of the repetitive nature of panic attacks. Drawing a distinction between the infrequent autonomic events reported as panic attacks in the normal population that may be caused by a number of factors, and the recurring panic attacks that typically culminate in a diagnosis of panic disorder, he has suggested that only the latter should be considered as linked to catastrophic misinterpretation of bodily sensations. Accordingly, a
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
531
Fig. 1. Cognitive account of panic attacks (Clark, 1986, p. 463).
central tenet of cognitively driven treatment applications is that symptom change in panic disorder occurs through the reduction of catastrophic misinterpretation of bodily sensations (Clark et al., 1994; Rachman, 1994). 2.2. Bandura’s theory of self-efficacy In conceptualizing the role of self-efficacy in anxiety disorders, Bandura (1988) argued that the sense of threat, which is the hallmark of these disorders, should not be regarded as a fixed property. Rather, it is derived both from the individual’s appraisal of perceived coping abilities and the assessment of the perceived dangers of the environment. However, Bandura explicitly ascribed a subordinate role to the role of perceived danger. In panic disorder, this perceived danger is defined in Beck, Emery, and Greenberg’s (1985) terms as thoughts of ‘‘physical or psychosocial harm, such as dying, going crazy or humiliating oneself’’ (Williams, 1995). In other words, perceived danger closely approximates the construct of catastrophic misinterpretation of bodily sensations. In Bandura’s model, therefore, it is assumed that perceived danger is an effect of low self-efficacy and that changes in self-efficacy will lead to changes in perceived danger. Thus, the self-efficacy model would predict that improvement in CBT for panic disorder is mediated by changes in self-efficacy, rather than by changes in catastrophic misinterpretation of bodily sensations. 2.3. Beck’s cognitive approach Ironically, a potential resolution of these two foregoing positions was presaged in Beck’s seminal work on cognitive therapy for anxiety disorders, but has been largely overlooked. Although Clark’s construct has been identified as directly relating to Beck’s theory (Beck et al., 1985) concerning danger-
532
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
laden schemas in anxiety disorders (e.g., Beck et al., 1985; Salkovskis, 1998), a comparison of the construct of catastrophic misinterpretation of bodily sensations and Beck’s original conceptualization of anxiety disorders suggests a striking and important difference. The broader cognitive theory from which the construct of catastrophic misinterpretation of bodily sensations is derived suggests that while anxiety occurs because of the appraisal of events or situations as threatening, this appraisal is derived from four constituent elements, of which the ‘‘perceived cost or awfulness of danger’’ is only one element. The other elements are perceptions of the likelihood of danger, coping ability, and rescue factors. For Beck, the central problem of anxiety disorders was a sense of vulnerability, defined as a person’s perception of himself as subject to internal or external dangers over which his control is lacking or is insufficient to afford him a sense of safety. (Beck et al., 1985, p. 67) In Beck’s original work, then, it was the interaction between negative cognitions concerning the presence of danger, and positive cognitions regarding control or coping that contributes to anxiety, a topic that will be discussed in more detail in the following sections. In Clark’s explanation of panic, however, catastrophic misinterpretation of bodily sensations provides the key focus and explanation of clinical symptoms, and it is the role of this catastrophic misinterpretation of bodily sensations that has dominated cognitive mediational accounts of panic disorder to date.
3. Empirical support for catastrophic misinterpretation of bodily sensations: Clark’s predictions Clark (1986, 1988, 1996) suggested that there were a number of predictions that should be investigated in order to support the construct of catastrophic misinterpretation of bodily sensations in panic disorder. Evidence to support Clark’s predictions has steadily accumulated over the last decade or so, and has been presented in detail in a number of extensive reviews (Clark, 1996, 1999; Khawaja & Oei, 1998). Therefore, this body of evidence is only briefly reviewed in this section, with an emphasis on studies that suggest there are limitations to an exclusive focus on the role of catastrophic misinterpretation of bodily sensations in cognitive accounts of panic disorder. 3.1. Cognitive specificity of catastrophic misinterpretation of bodily sensations in panic disorder A number of studies indicate that catastrophic misinterpretation of bodily sensations appears to be more prevalent in patients in panic disorder in comparison to both the normal population and other clinical disorders (for review, see Khawaja & Oei, 1998). Consistent with much evidence, a comparison between patients with panic disorder, social phobia, and nonanxious controls on response times to a range of emotionally positive and threatening words found that patients with panic disorder had longer response times to panic-specific threat words (Maidenberg, Chen, Craske, Bohn, & Bystritsky, 1996). However, these patients also displayed longer response times to threat words associated with social and general concerns, suggesting that patients with panic disorder may also be characterized by a more generalized attentional bias to threat. Using a modified lexical decision task, Schniering and Rapee (1997) failed to replicate differences between patients with panic disorder and nonclinical controls in the association of somatic sensations and threat. Although they noted a number of methodological issues that may have accounted for this discrepant finding, importantly, they suggested that other cognitive
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
533
factors may distinguish individuals with panic disorder from nonpanickers and that these factors should be considered in addition to threat-related cognitions. Specifically, Schniering and Rapee suggested that people with panic disorder may be more clearly distinguished by their belief that they have little capacity to alter or influence threat-related situations, a conclusion which may be consistent with the more generalized threat bias detected by Maidenberg et al. However, the possibility that patients with panic disorder may be characterized by both high levels of catastrophic misinterpretation of bodily sensations and low levels of panic self-efficacy emerges more clearly in studies investigating experimental provocation and manipulation of panic attacks. 3.2. Cognitive mediation of panic attacks in biological challenge tests There is considerable experimental evidence to support cognitive mediation of panic attacks induced by biological challenge tests (Clark, 1993b; Khawaja & Oei, 1998; Rapee, 1995). However, the most relevant question regarding the experimental evidence in the area of biological challenge tests concerns exactly which cognitive variables are responsible for the effects demonstrated in the majority of these experiments. Reviewing this area of research, Rapee (1995) suggested that it was both the reduction of threat and the availability of control that are manipulated in these experiments. Furthermore, in the broader context of laboratory-based research into panic attacks, it is worth remembering Rachman’s (1988) caution that the experimental conditions under which these panics are induced offer, by their nature, far greater control than is available in clinically occurring panics. According to Clark (1993a), however, the key variable is the reduction of catastrophic misinterpretation of bodily sensations, with the provision of control in these experiments simply further adding to the reassurance that the sensations are not dangerous. In a study that provides a clinical context in which to consider the relative contribution of both types of cognition, Schmidt, Trakowski, and Staab (1997) compared CBT-treated and untreated patients with panic disorder on pre- and posttreatment responses to five repeated vital-capacity inhalations of 35% CO2/65% O2. Measures in this study included panic attack frequency, SUDS ratings of subjective anxiety, psychophysiological indices of arousal, panic-related appraisals of the likelihood, consequences and perceived self-efficacy in coping with panic, and anxiety sensitivity. Anxiety sensitivity is a construct that refers to fears of anxiety symptoms based on the beliefs that these symptoms may have harmful consequences (Reiss & McNally, 1985) and thus is linked to, although separate from, the concept of catastrophic misinterpretation of bodily sensations (McNally, Hornig, Hoffman, & Han, 1999). Although there were no pretreatment differences between groups on any of these measures in the Schmidt et al. study, at posttreatment, patients who had participated in CBT recorded significantly fewer panic attacks during posttreatment inhalations. CBT-treated patients also reported significant pre- to postchanges on anxiety sensitivity and panic-related appraisals, whereas there were no changes on any of these indices reported by untreated patients. Schmidt et al. reported that there were relatively few posttreatment differences between treated and untreated patients on psychophysiological measures of arousal. However, patients with high levels of anxiety sensitivity and low levels of panic coping self-efficacy were at least at eight times greater risk for experiencing panic during the posttreatment assessment. Patients who continued to believe that anxiety symptoms were associated with a high probability of negative consequences, and patients who had a low level of belief in their ability to cope with panic attacks were at greater risk of panic attacks in response to the CO2 challenge. However, the strong negative correlation evident between levels of anxiety sensitivity
534
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
and panic self-efficacy in the results reported by Schmidt et al. highlights the importance of establishing whether positive cognitions contribute to the prediction of panic severity beyond the influence of negative cognitions. 3.3. The role of catastrophic misinterpretation of bodily sensations in maintenance of treatment outcome Clark’s (1986) final prediction suggests that sustained improvement in panic disorder depends on reduction of catastrophic misinterpretation of bodily sensations that has occurred during treatment. The available evidence mostly supports this prediction (Clark, 1996, 1999). For example, Clark et al. (1997) found that catastrophic misinterpretation of bodily sensations (as measured by the Bodily Sensations Interpretation Questionnaire) predicted maintenance of outcome at follow-up, controlling for panic/ anxiety composite scores at the end of treatment. Similarly, in a comparison of treatment outcome between CBT and Applied Relaxation (AR), Westling and Ost (1995) found successful treatment was characterized by a reduction in the catastrophic interpretation of bodily sensations in panic patients relative to the control group. They also found that compared to panickers at posttreatment, nonpanickers had reduced their cognitive threat bias at posttreatment and at follow-up. In contrast, Stoler and McNally (1991) found that patients judged to be recovered from panic disorder with agoraphobia were more similar to symptomatic patients than control subjects in their cognitive bias towards catastrophic misinterpretation of bodily sensations. Results from the sentence completion task that Stoler and McNally used in their study provide a useful insight into the potential role of panic selfefficacy in symptom reduction. Sentence completions were classified as either biased (threat-related) or unbiased. Biased interpretations were further classified as to whether or not subjects indicated efforts at adaptive coping with the perceived threat. Despite persistence of cognitive biases, analysis of responses of recovered patients revealed a pattern of adaptive coping with perceived threat. In contrast, symptomatic patients failed to exhibit an adaptive coping style. The interesting conclusion suggested by the Stoler and McNally study is that, even with the persistence of cognitive bias towards threat, perceived ability to cope adaptively with threat may serve to maintain treatment gains. In summary, although there are a number of studies that support each of the predictions that Clark has made regarding the role of catastrophic misinterpretation of bodily sensations in panic disorder, review of this evidence suggests that attention to the role of panic self-efficacy may serve to strengthen the case for a cognitively based approach to panic disorder. A number of other areas relevant to cognitive accounts of panic disorder that highlight the limitations to catastrophic misinterpretation of bodily sensations as the sole cognitive explanation of panic disorder are reviewed in the following section.
4. Limitations to catastrophic misinterpretation of bodily sensations as an explanation of panic disorder: the link between bodily sensations, cognitions, and affect A central assumption of Clark’s (1986) explanation of panic disorder is that there is a logical link between bodily sensations, affect, and cognitions that results in panic attacks. In its most straightforward form, this link suggests that bodily sensations give rise to catastrophic misinterpretation of bodily sensations that in turn produces panic. According to Clark, the key factor maintaining this link is the presence of catastrophic misinterpretation of bodily sensations, and thus reduction of this negative cognition should explain clinical change. The following section reviews evidence that suggests instead
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
535
that the link between bodily sensations, cognitions, and affect that results in panic may also be maintained by low levels of positive cognitions concerning coping or control. 4.1. The link between cognition and affect According to Clark (1986, 1996, 1999), the reason that people experience panic is that they interpret bodily sensations as a signal for imminent physical or psychological catastrophe. The ‘‘normalizing rationale’’ of the cognitive approach therefore is to suggest that anyone who makes such an appraisal of imminent danger would be extremely anxious and likely to respond in a manner designed to minimize the risk. As Salkovskis and Hackmann (1997) have commented: ‘‘the awfulness of many of the feared catastrophes appears to be self-evident’’ (p. 46). Thus, the presence of cognitions regarding imminent danger should be reliably associated with panic or anxiety and avoidance. However, as Rachman (1990) has documented, there is instead a considerable diversity of reactions manifested when people face situations that are literally life-threatening. People can parachute from planes, deactivate bombs, and face combat conditions without panicking. In these situations, people may experience fear, yet enter or remain in the situation despite this fear, or even more intriguingly, simply experience relatively little fear in these obviously dangerous situations. Importantly, fear appears to decrease even when the perceived danger of the situation remains the same. Salkovskis and Hackmann (1997) have noted that the anxiety reactions experienced in life-threatening situations are not likely to be regarded as abnormal. In this sense, the absence of panic in these situations may in part reflect the operation of an alternative explanation for anxiety symptoms. Despite this explanation, however, it seems clear that the actual danger of these situations remains, thus challenging the inevitability of overwhelming anxiety associated with danger-related cognitions that is posited by the construct of catastrophic misinterpretation of bodily sensations. According to Rachman (1990), what appears to distinguish people operating in such situations is a high level of self-confidence, a concept similar to self-efficacy. Thus, even in situations of actual and acute danger, the presence of positive cognitions concerning coping or control may enable people to tolerate cognitions concerning imminent danger and perform adequately. This observation can be translated in two ways that are relevant to the experience of panic. Firstly, an initial appraisal of threat or danger may not inevitably escalate into incapacitating anxiety, even in the absence of a plausible reinterpretation of the threat stimuli itself. Instead, it appears that the presence of positive cognitions concerning coping or control may somehow prevent this escalation, potentially either through simply broadening attentional focus (Fredrickson, 2001), or by allowing an individual to encapsulate and thereby contain threat-related cognitions within a broader, implicational level of self-referent meaning associated with control and coping (e.g., Power & Dalgleish, 1997; Teasdale, 1993). Secondly, in clinical terms, the presence of panic self-efficacy may enable a reinterpretation of threatening stimuli to take place, with a concomitant reduction of catastrophic misinterpretation of bodily sensations. Treatment strategies associated with CBT often involve asking patients to enter and remain in situations that are likely to trigger negative cognitions regarding imminent danger. Presence of panic self-efficacy that allows them to tolerate this experience seems, at a minimum, likely to be of considerable assistance in assisting patients to disconfirm the feared catastrophe. The threat stimuli that are most clearly involved in panic disorder are bodily sensations. It is therefore of importance to consider ways in which positive cognitions concerning coping or control may be involved in interpretation and response to these sensations.
536
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
4.2. The link between bodily sensations and cognitions According to Clark (1986, 1996), physiological arousal acts as a trigger for catastrophic misinterpretation of bodily sensations in people with panic disorder, and thus should precede or accompany catastrophic misinterpretation. There is relatively little evidence to suggest that people with panic disorder either experience more fluctuations in physiological functions, or are generally more physiologically reactive (and thus more prone to experience bodily sensations) than other people (Ehlers, 1993). On the contrary, it seems that when exposed to similar bodily sensations, people with panic disorder are more likely to experience anxiety and catastrophic cognitions (e.g., Pauli et al., 1991). Even when comparison subjects from a normal population are experiencing chest pain of demonstrated medical concern, they appear significantly less likely to experience catastrophic cognitions than patients with panic disorder (Fraenkel, Kindler, & Melmed, 1997). Such findings are, of course, consistent with the role of catastrophic misinterpretation of bodily sensations in mediating the experience of panic. However, it is of interest to consider why such comparison subjects from the normal population do not experience the same level of catastrophic cognitions as people with panic disorder. One possibility is the existence of positive illusions, such as exaggerated perceptions of control or mastery, which have been suggested as being not only normative, but also as having considerable protective value with regards to both mental and physical health (Taylor & Brown, 1988; Taylor, Kemeny, Reed, Bower, & Gruenewald, 2000). Arguably, comparison subjects may have been less likely to experience catastrophic cognitions regarding bodily sensations because of the presence of positive cognitions concerning coping or control (see also, Folkman & Moskowitz, 2000). An alternative view, however, is that these comparison subjects simply had no prior underlying association of bodily sensations with danger. Thus, when experiencing chest pain, these subjects were simply not primed to interpret these sensations in a negative way. Certainly, Clark (1988) has postulated that the tendency to catastrophically misinterpret bodily sensations is a ‘‘relatively enduring cognitive trait . . . amplified when a individual enters an anxious state’’ (p. 77). As a consequence, most people at risk for panic attacks presumably should exhibit these interpretive biases prior to the development of clinical symptoms. However, results from a recent study fail to support this hypothesis. Anxiety sensitivity has been established as a cognitive risk factor for the development of panic attacks (Schmidt, Lerew, & Jackson, 1997) and thus, is of particular value in experiments that are concerned with the responses of people who are at risk of, but have not yet developed, panic disorder. McNally et al. (1999) used a nonpanicking sample to investigate the relationship between scores on the Anxiety Sensitivity Index and a range of interpretative, memory and attentional tasks similar to those that have been used to measure threat-related cognitive biases in patients with panic disorder. If these threat-related cognitive biases predate the development of the disorder, then people who are identified as being at risk of developing the disorder presumably should show greater evidence of these threat-related cognitive biases than people who are considered to be a low risk of developing the disorder. In contrast, McNally et al. found little evidence to support the association between risk of developing panic disorder and cognitive biases for threat, concluding instead that these threat-related cognitive biases may be the result, rather than the precursor, of panic disorder. Two aspects of the McNally et al. (1999) study are of particular importance in the current context. Firstly, their results do suggest that more marked interpretive biases for internal cues are evident at higher levels of anxiety. Secondly, McNally et al. speculated that either absence of these cognitive
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
537
biases, or more interestingly, presence of what they described as ‘‘positivity’’ biases, may protect people with high anxiety sensitivity from developing panic disorder. Thus, schematic processing, which in Beck et al.’s (1985) terms reflects an interaction between negative and positive cognitions, that results in higher levels of anxiety is likely to be more strongly associated with interpretative threat biases, such as catastrophic misinterpretation of bodily sensations. But as a consequence, presence of positive cognitions concerning coping or control may determine the impact of these biases, either by inhibiting the development of the initial triggering anxiety, or by offsetting their situational influence, even in people who have a marked underlying interpretative bias associating bodily sensations with danger. Positive cognitive factors, such as panic self-efficacy, may thus be helpful in explaining, and ultimately decreasing, the frequency with which people with panic disorder experience catastrophic misinterpretation of bodily sensations when experiencing physiological arousal. Paradoxically, the evidence to support the proposition that panic attacks are necessarily associated with marked physiological arousal is at best mixed. Many studies looking at the role of bodily sensations in panic disorder have focused on the role of cardiovascular states. Although a number of earlier studies suggested that panic attacks were associated with elevated heartbeat (e.g., Taylor et al., 1986; Taylor, Telch, & Havvik, 1983), more recent studies have failed to clearly support this association (Kenardy, Oei, Weir, & Evans, 1993; Khawaja & Oei, 1999). Attempting to account for these discrepant results, it has been argued that panic disorder patients are simply more accurate in heartbeat detection (HBP; Clark, 1999) and thus presumably may be more sensitive to minor variations in HBP. However, a recent reanalysis of this literature (Van der Does, Antony, Ehlers, & Barsky, 2000), which combined the results of seven HBP studies, does not support this contention. Instead, Van der Does et al. concluded that although accurate HBP is more likely to be found in people who have continuous or frequent episodes of clinical anxiety, it does not appear to be specific to panic disorder. Of course, not all individuals with panic disorder report cardiac-related concerns, and for those who do not, evidence pertaining to HPB may be less relevant (Ehlers, 1993). Similarly, methodological limitations in studies testing interoceptive accuracy in HPB may account for some of these findings (Zoellner & Craske, 1999). A more broadly based test of the specificity of interoceptive acuity to panic disorder, however, was reported by Rapee (1994). He found that panic-disordered patients were no better than normal controls at detecting increases in CO2 concentrations in inhaled air, despite the fact that such increases should impact across a range of bodily sensations. Accuracy does not, of course, necessarily equate to awareness or even more importantly, to interpretation. It has been frequently assumed in the literature that the ability of individuals to accurately detect bodily sensations is associated with increased attention (Schmidt, Lerew, & Trakowski, 1997), and thus, a greater probability of making a catastrophic misinterpretation of bodily sensations. However, failure to support such a relationship does not by itself reflect on the role of catastrophic misinterpretation of bodily sensations in mediating panic (McNally, 1999). Van der Does et al. (2000) argued, for example, that their findings regarding HBP in panic disorder should be understood in terms of schemabased processing, suggesting that once a patient with Panic Disorder perceives a situation as threatening, an ‘anxiety’ schema is activated, and that perception of symptoms is more guided by the schema (that is, past information) than based upon present physiological status. (p. 61)
538
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
Similarly, Davey (1995) has drawn attention to the possibility that it is the presence of repetitive processing of interoceptive cues that may explain the role of these cues in the enhancement of anxiety. Thus, the initial actual intensity, or indeed accuracy of awareness of bodily sensations, may again be less important in the experience of panic than the schematic factors which mediate the experience of panic (see also, McNally, 1999). This finding is in fact consistent with Beck’s schema model, which suggests that cognitive biases occur after the schema has been activated (Beck et al., 1985). Again, according to Beck, however, it is the interaction between negative cognitions and positive cognitions at the schematic level that produces anxiety. In this sense, the presence of positive cognitions concerning coping or control may in part mediate the effect of interoceptive cues and awareness upon panic, a possibility that may help to account for the clinical evidence regarding the variant occurrence of panic discussed in the next section. 4.3. The link between bodily sensations and affect Clinically, it is common for patients with panic disorder to report that at times, they ‘‘feel more confident’’ and therefore, less prone to panic. This phenomenon is evident in an experimental study (Rachman, Levitt, & Lopatka, 1987) that used standardized behavioural tests that involved patients entering feared situations to elicit panic episodes (defined as when the patient reported either having panicked or come close to panicking). Although their results in the main supported the association of bodily sensations, catastrophic cognitions, and panic, Rachman et al. reported that in over 50% of these trials, no panic was recorded. However, bodily sensations typically associated with panic attacks were recorded in all of these nonpanic trials, albeit to a lesser extent than in trials that did result in panic. In view of evidence reviewed earlier in this paper that suggests that the actual intensity of bodily sensations may not be a major factor in producing panic, it is pertinent to consider alternative explanations for the absence of panic on these trials. Several studies suggest that it is clusters, or combinations of bodily sensations, that appear to be most clearly linked to specific catastrophic misinterpretations (Chambless, Beck, Gracely, & Grisham, 2000; Kenardy, Evans, & Oei, 1992; Marks, Basoglu, Alkubaisy, & Sengun, 1991). Therefore, it could be argued that these bodily sensations experienced in nonpanic trials did not result in panic because they were not occurring in meaningful relation to other sensations. Looking at combinations of symptoms, however, Rachman et al. (1987) found that the same combination of symptoms in some trials resulted in panic, yet in other trials, no panic. The only difference observed in regard to these different outcomes for the same combinations was the presence of catastrophic cognitions. Thus, for example, combination of palpitations, shortness of breath, and dizziness would not result in panic unless it was accompanied by a catastrophic cognition. On the face of it, this observation seems to provide strong support for Clark’s (1986) explanation of panic disorder, in that it is consistent with the proposition that it is the misinterpretation of sensations (rather than the sensations themselves) that provoke panic. However, these sensations were being experienced in these nonpanic trials by untreated patients, who demonstrably were prone to catastrophic misinterpretation of bodily sensations. Despite the presence of both bodily sensations and a cognitive bias towards misinterpreting such sensations, patients did not experience panic on each trial. Clearly, patients with panic disorder do not inevitably panic when they experience panicogenic bodily sensations. Instead, there appears to be other factors that may influence and possibly offset their tendency to catastrophically misinterpret these sensations on occasions.
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
539
An issue of relevance here is the relationship of panic (or fear) to anxiety. From a cognitive perspective, panic and anxiety are viewed essentially on a continuum of the same emotion which Rapee (1996, p. 83) has described as ‘‘. . .differing possibly along such parameters as temporal features and intensity’’ (although emotion theorists such as, e.g., Antony & Barlow, 1996, argue for a qualitative distinction between the two states). According to Clark (1986), a number of different emotional states or innocuous events (e.g., exercise) may produce the bodily sensations that give rise to catastrophic misinterpretation (as opposed to simply anxiety). However, prospective monitoring by patients with panic disorder suggests that many panic attacks appear to surge out of ongoing anxiety states (Basoglu, Marks, & Sengun, 1992). Indeed, this concept of panic attacks occurring as discrete episodes in the context of ongoing anxious apprehension is central to Barlow’s (1988) influential emotion theory-based model of panic disorder. The debate about the relationship of panic and anxiety has a number of important implications for the field as a whole (Rapee, 1996). However, its relevance in the present context is to again suggest that the inclusion of panic self-efficacy may be required to increase the explanatory power of a cognitive mediational account of panic disorder. Specifically, the role of panic self-efficacy may be firstly, to explain the translation of anxious appraisal to catastrophic misinterpretation of bodily sensations on the same continuum, and secondly, to account for situations in which the major components of the link between cognition, bodily sensations, and affect may be present, and yet fail to culminate in panic. 4.4. The impact of metacognitive processes on the link between bodily sensations, cognitions, and affect Recent theoretical interest in the metacognitive aspects of cognition raises a further possibility regarding the contribution of positive cognitions to the experience of panic. A number of researchers (e.g., Power & Dalgleish, 1997; Teasdale, 1999a, 1999b; Wells & Matthews, 1996) have argued that the specific or thematic content of thought represents only one level of possible analysis in cognitive approaches to emotional disorder, and that a useful extension of these approaches may be found in examining other aspects of thought. These other aspects of thought concern the metacognitive beliefs that people have about their own mental processes (Hackmann, 1997). For example, a patient may have the thought ‘‘I am going to faint’’ and feel anxious. But according to metacognitive approaches, the anxiety experienced as a result of this thought will be additionally influenced by other dimensions, such as believability, frequency, sense of control (over the thought), etc. Low levels of panic self-efficacy may thus play a role in the link between bodily sensations, cognitions, and affect by resulting in a perceived inability to disrupt or control the repetitive nature of catastrophic misinterpretation of bodily sensations. It may be that simply experiencing a catastrophic cognition in the presence of bodily sensations is not by itself an adequate explanation of schematic processing in the experience of panic. A fleeting thought of danger by itself may be of less consequence than the perceived inability to evaluate, discount, or somehow control this danger-related thought. Consistent with this line of reasoning, Fehm and Margraf (2002) reported that relative to patients with social phobia, and normal controls, patients with panic disorder showed specific deficits in suppression of thoughts associated with physical or psychological catastrophes. In keeping with the cognitive concept of panic as an escalating circular process, a perceived lack of control over negative cognitions that associate bodily sensations with danger may have two consequences. Firstly, this perceived lack of control may appear to substantiate the content of the catastrophic misinterpretation of bodily sensations, and secondly, it may directly contribute to anxious arousal that appears to provide even further confirmatory evidence for the feared catastrophe.
540
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
In summary, the concept that there is a link between bodily sensations, cognition, and affect in panic disorder is both logically and intuitively compelling. To date, however, efforts to explicate this link have largely focused on the role of negative danger-related cognitions. Evidence reviewed in the preceding sections suggests instead that an understanding of this link may be considerably strengthened with reference to the role of positive cognitions concerning coping or control. Cognitions regarding imminent danger can be tolerated without inevitable escalation of incapacitating anxiety through the presence of self-efficacy, and panic self-efficacy is likely to ultimately assist in disconfirming feared catastrophes. Similarly, positive cognitions concerning coping or control may prevent either the actual or perceived registration and intensification of bodily sensations, or simply allow people to experience bodily sensations without further escalation of anxiety and catastrophic misinterpretation of bodily sensations. Finally, level of panic self-efficacy may determine when a fleeting thought of danger can be dismissed or tolerated, as opposed to engaged with, and reified. However, it is in the areas that most directly impact on an understanding of change processes in CBT that the limitations of catastrophic misinterpretation of bodily sensations as the predominant focus of cognitively based explanations of panic disorder are most apparent.
5. Limitations to catastrophic misinterpretation of bodily sensations as an explanation of symptom reduction in panic disorder One aspect of Clark’s account of panic disorder that has attracted considerable criticism (e.g., McNally, 1994; Seligman, 1988) concerns its ability to explain why people with panic disorder repeatedly catastrophically misinterpret bodily sensations. As Seligman (1988) has commented: A person who has had Panic Disorder for a decade may have had about 1,000 panic attacks. In each one, on the cognitive account, he misinterpreted his racing heart as meaning that he was about to have a heart attack, and this was disconfirmed. Under the laws of disconfirmation that I know, he received ample evidence that his belief was false and he should have given it up. (p. 326) Consideration of why people with panic disorder persist in their catastrophic misinterpretation of bodily sensations, and whether reduction of this negative cognition alone is sufficient to overcome the disorder, goes to the heart of understanding the cognitive changes that may be involved in successful treatment. In a recent review, Clark (1999) suggested a number of processes that may explain the persistence of various specific negative cognitions in anxiety disorders, and thereby more generally explain the persistence of the disorders themselves. Central to Clark’s argument is that a reduction of these negative cognitions will result in a reduction of symptoms. The areas that he suggested that are of particular relevance to the persistence of catastrophic misinterpretation of bodily sensations in panic disorder concern safety-seeking behaviour and threat-related attentional and memory biases. 5.1. Safety-seeking behaviour In Clark’s original formulation of panic attacks, agoraphobia was described as ‘‘a fear of panic rather than a fear of specific situations’’ (Clark, 1986, p. 467). Development of the concept of safety-
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
541
seeking behaviour (Salkovskis, 1988; Salkovskis, Clark, & Gelder, 1996) represents an elegant attempt to provide a direct link between catastrophic misinterpretation of bodily sensations and avoidance of situations in a cognitive account of panic disorder. According to Salkovskis, safety-seeking behaviour refers to either overt or covert behaviours whose purpose is either the prevention or minimization of feared catastrophes. This concept of safety-seeking behaviour refers both to the marked situational avoidance that is characteristic of agoraphobia as well as to a range of more subtle behaviours which prevent disconfirmation of feared consequences and thus partly account for the persistence of catastrophic misinterpretation of bodily sensations. Although more recently, Salkovskis and Hackmann (1997) have argued that the phenomenon of agoraphobia may be additionally related to a range of beliefs that constitute threat appraisals maintaining the disorder, their analysis retains a key role for catastrophic misinterpretation of bodily sensations, consistent with Clark’s (1986) original formulation. Attempts to account for the persistence of catastrophic misinterpretation of bodily sensations are also, therefore, at least partly a measure of the extent to which the presence of this negative cognition can adequately explain the avoidance and more general disability associated with panic disorder. As reviewed below, however, it seems likely that panic self-efficacy may also play a role. Salkovskis et al. (1996) examined catastrophic misinterpretation of bodily sensations and avoidance behaviour in a large sample of patients with panic disorder. Their results demonstrated that catastrophic misinterpretation of bodily sensations appears to be logically linked to the behaviours used to avoid or minimize the occurrence of the feared catastrophe (Salkovskis et al., 1996). For example, patients who feared fainting would characteristically hold on to objects or other people, whereas fears concerning heart attacks were associated with sitting down and keeping still. However, evidence of such links demonstrates only the association of catastrophic misinterpretation of bodily sensations with behaviour, rather than establishing the causal role of safety-seeking behaviours in the maintenance of this negative cognition, or maintenance of the disorder itself. Attempting to investigate such a causal role in a subsequent study, Salkovskis, Clark, Hackmann, Wells, and Gelder (1999) experimentally manipulated safety-seeking behaviours in order to examine the subsequent impact of these behaviours on catastrophic misinterpretation of bodily sensations and anxiety in a sample of patients with panic disorder. They instructed their experimental group to drop safety-seeking behaviours during a 15-min exposure. In contrast, the control group was given what the researchers described as ‘‘habituation’’ rationale during the 15-min exposure, which simply instructed these patients to remain in the situation as a way of reducing their anxiety. The results of Salkovskis et al.’s study indicated that catastrophic misinterpretation of bodily sensations and anxiety were significantly reduced in the experimental group relative to the control. However, some caution is needed before interpreting these results as supportive of the role of catastrophic misinterpretation of bodily sensations alone in the cognitive mediation of panic and panic-related avoidance. The methodology used in Salkovskis et al.’s (1999) study is highly similar to that used by Williams (1990) and Williams and Falbo (1996) in a series of studies investigating the role of self-efficacy in panic disorder. However, for Williams, the goal of dropping what he describes as ‘‘defensive coping rituals’’ (Williams, 1990, p. 97) is part of a number of therapeutic strategies used to increase selfefficacy. Specifically, for Williams, the purpose of instructing patients to desist in their maladaptive safety behaviours is so that they may engage in active coping behaviours that will increase their selfefficacy. It is therefore possible that dropping safety behaviours in the Salkovskis et al. study not only
542
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
decreased catastrophic misinterpretation of bodily sensations, but also increased panic self-efficacy. Again, the important question here is whether changes in both cognitions may mediate reduction of both panic and panic-related avoidance. A related question concerns the association between the actual occurrence of panic attacks and avoidance in panic disorder. Reports of treatment outcome in CBT tend to focus on the presence or absence of panic attacks as their primary measure (for example, see Clark, 1999). However, the evidence more generally seems to suggest that factors beyond panic phenomena contribute to the considerable avoidance and disability observed in some individuals with panic disorder (Bandelow, Hajak, Holzrichter, Kunert, & Ruther, 1995; Hollifield, Katon, Skipper, & Chapman, 1997; Telch, Schmidt, Jaimez, & Jacquin, 1995). Certainly, there is a strong body of research to suggest that most typically panic attacks precede avoidance (for review, see Clum & Knowles, 1991). However, in their review of the relationship between panic attacks and avoidance, Craske and Barlow (1988) found that neither the severity or frequency of panic attacks differentiated between individuals with either minimal and extensive avoidance (see also Clum & Knowles, 1991). If fears of the physical or psychological consequences of anxiety were related to avoidance, then it would be expected that individuals with extensive avoidance would score more highly on the Anxiety Sensitivity Index. However, neither Craske and Barlow (1988) nor Clum and Knowles (1991) found evidence to support this relationship in their reviews of the literature. Similarly, Telch, Brouillard, Telch, Agras, and Taylor (1989) found that thoughts about the physical consequences of panic failed to distinguish between individuals with minimal and extensive avoidance. Instead, Telch et al. reported that individuals with minimal avoidance in their sample were characterized by comparatively higher perceived self-efficacy in coping with panic attacks. Similarly, Hoffart (1995a) found that perceived self-efficacy for entering feared situations was a better predictor of situational avoidance than the strength of catastrophic beliefs. Thus, an increase in panic self-efficacy is likely, at a minimum, to be involved in change processes that result in reduction of the avoidance and disability that accompanies panic attacks. However, a more detailed description of the means by which panic self-efficacy may be involved in change processes related to the core features of panic disorder emerges from contemporary models of information-processing, reviewed in the following section. 5.2. Attentional and memory biases: information-processing models Central to the role that both attentional and memory biases are hypothesized to play in the cognitive model of panic disorder is the notion that individuals selectively attend to, and recall threat-related stimuli. These biases are thought to maintain catastrophic misinterpretation of bodily sensations by heightening the perception of threat (e.g., Beck et al., 1985), thus contributing to a feedback loop (e.g., Clark, 1994) that prevents disconfirmation of these catastrophic misinterpretation of bodily sensations. As noted earlier, considerable evidence suggests that patients with panic disorder are characterized by cognitive biases for threat. According to Clark, such biases partly explain the persistence of catastrophic misinterpretation of bodily sensations. However, the most relevant question to examine in context of change processes concerns the changes that need to occur during treatment in order to either eliminate, reduce, or offset the impact of these biases. Despite the emphasis on measuring the reduction of catastrophic misinterpretation of bodily sensations as an indicator of change processes in CBT for panic disorder, contemporary models of information processing suggest instead that these change processes are
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
543
also likely to involve an increase in panic self-efficacy. Two such models that have particular relevance to the area of anxiety are reviewed below. 5.2.1. Beck and Clark’s (1997) information-processing model of anxiety Beck and Clark (1997) have presented an information-processing model of anxiety, based on an elaboration of the cognitive theory of anxiety initially advanced by Beck et al. (1985). The first ‘‘orienting’’ stage of Beck and Clark’s model is characterized by automatic processing, whose purpose is to identify stimuli and determine the relative allocation of system resources to processing information regarding the stimuli. Allocation of system resources which occurs as a function of this stage is determined by perceived valence and personal relevance of the stimuli, with stimuli identified as representing potential threat given priority in subsequent processing. The second stage involves the activation of the ‘‘primal mode,’’ described by Beck and Clark (1997, p. 52) as ‘‘a cluster of interrelated schemas embodying more primitive and immediate cognitive/ affective/behavioural/physiological patterns aimed at meeting evolutionary derived objectives, such as survival, safety, security, procreation and sociability.’’ In anxiety, this primal mode is characterized by threat and is associated with autonomic arousal, behavioural mobilization and inhibition (i.e., ‘‘fight or flight’’), feelings of fear and hypervigilance for threat cues. Much of this information processing may take place outside of awareness and involuntarily, reflecting automatic processing. However, the presence of controlled processing is denoted by the amount of attentional resources captured by the primal mode, resulting in narrowing or constriction of cognitive awareness, which according to Beck and Clark blocks off more constructive or reflexive modes of thinking. Much of this controlled processing is occupied with initial semantic analysis of the threat, indexed by negative, danger-laden cognitions. The third and final stage of this model is that of ‘‘secondary elaboration,’’ in which other schemas emerge that are relevant to the individual. This stage is mostly characterized by controlled processing, although the activation of the primal mode results in continued involuntary cognitive processing of threat. The other schemas activated at this stage are concerned with evaluation of self in the current context, representing in particular the individual’s assessment of coping resources and capacity to deal with the perceived threat. Importantly, Beck and Clark (1997) suggest that anxiety will decline at this third stage if the action of these other schemas enables a reappraisal of the situation, in which the probability or severity of threat is decreased relative to perceived ability to cope. Based on this analysis, Beck and Clark argued that the goal of CBT is to deactivate the primal threat mode and strengthen the impact of controlled processing at the final stage in order to promote a more constructive and reflexive processing. Consideration of Beck and Clark’s (1997) information-processing model emphasizes the point that cognitive approaches to treatment of panic disorder must take into consideration factors beyond negative cognitions that associate bodily sensations with danger. Instead, the process of change during treatment is likely to involve the development and increase of positive cognitions concerning control or coping, such as panic self-efficacy. Beck and Clark’s model implies, however, that the impact of panic selfefficacy will occur at a latter stage in information processing. In contrast, Mansell (2000) has argued that automatic processing itself may be influenced by alterations of higher-order meaning, whereas Matthews and Wells (2000) have questioned the automaticity of threat-biases, suggesting instead that these biases may be more clearly determined by strategic processing. Clearly, considerably more experimental work is needed to address the complex issues underlying aberrant information processing in panic disorder. However, the most important inference from research, to date, in this area, is the possibility that changes
544
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
in panic self-efficacy are likely to be highly relevant in mediating symptom reduction in the treatment of panic disorder, either through impacting on appraisal of information that the individual attends to, or the attentional biases themselves. 5.2.2. Lang’s bioinformation model Recent elaborations of Lang’s (1979) influential bioinformational theory of emotion in regard to the anxiety disorders suggest a similar conclusion. Foa and Kozak (1986) contributed to the heuristic value of the bioinformational model in anxiety disorders by providing a more cognitively based explanation of changes that occur when networks associated with fear (fear structures) are modified by CBT. According to their analysis, representations of excessive responses (e.g., avoidance), together with inappropriate S–S and S–R associations (e.g., money–germs; spider–escape) are embedded within pathological fear structures, which makes them unrealistic but resistant to change. According to Foa and Kozak, the task of therapy for anxiety disorders was to alter these fear structures by incorporation of corrective information that is incompatible with the pathological elements of the structure. In particular, they argued that successful therapy works by altering the meaning elements that characterize the pathological fear structures. However, Mineka and Thomas (1999) have criticized the emphasis upon the reduction of negative, danger-related cognitions in applications of the bioinformational model to the anxiety disorders. They point instead to a considerable body of experimental and clinical evidence suggests that a key component in the experience of pathological fear is the sense of uncontrollability (see also Barlow, 1988, 1991; Shear, 1991 for a similar position). Mineka and Thomas consequently suggested a further elaboration of the bioinformational model, which is based on an integration of Foa et al.’s position and theories, such as Bandura’s (1983, 1988) self-efficacy model that emphasizes the importance of an individual’s self-appraisal of his or her capacities to deal with the specific situational demands. However, Mineka and Thomas suggested there were two limitations to the construct of selfefficacy in capturing changes in treatment of anxiety disorders. Firstly, they questioned the extent to which self-reported cognitive appraisals may serve as a veridical measure of cognitive processes occurring during exposure to anxiety-provoking stimuli. Secondly, they suggested that the emphasis on behaviour in Bandura’s (1977) exposition of self-efficacy failed to capture the broader range of domains relevant to the area of anxiety. In particular, Mineka and Thomas argued that it is perceptions of control regarding emotional responses that may be crucial. Both of these criticisms need to be addressed in the current context of considering the inclusion of panic self-efficacy within a cognitive mediational account of panic disorder. Although the relative merits of emotion-based theories of anxiety versus cognitively based approaches have generated considerable controversy, a recent discussion of this issue in the literature suggests that the similarities between the two approaches are likely to prove greater than their differences (Antony & Barlow, 1996; Rapee, 1996). Similarly, debate regarding the validity of conscious cognitive appraisals in describing aberrant processing of anxiety-related stimuli is likely to continue. As McNally (2001) has suggested, methodological pluralism in this area seems warranted, given that to date it is the cognitive appraisal theories that have had the greatest impact in developing useful treatment approaches. Mineka and Thomas (1999) are correct in their understanding of the emphasis upon behaviour in an earlier work on self-efficacy by Bandura (1977), and indeed this emphasis has been the focus of criticism from other researchers in the area (Barlow, 1988; McNally & Foa, 1996). In more recent elaborations of the construct, however, Bandura has expanded the concept of self-efficacy to incorporate the broad range of cognitive, emotional, and behavioural domains that contribute to personal control. This sense of
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
545
control can be defined as ‘‘a theory about oneself in relation to one’s environment, and a concern with causality, whether outcomes are a consequence of one’s own behaviour or tend to occur independently of that behaviour’’ (Brim, 1976, p. 243). While still fundamentally concerned with a sense of control, the construct of self-efficacy refers additionally to self-referent expectancies individuals hold about their own capacities. As Cervone (2000) noted, ‘‘people are unlikely to take action to control events if they doubt their own capability to execute requisite behaviours’’ (p. 31). Panic self-efficacy thus emerges as a highly relevant construct to examine the role of positive cognitions in panic disorder and as discussed below, there are two interrelated questions that need to be addressed.
6. The relationship between panic self-efficacy and catastrophic misinterpretation of bodily sensations in panic disorder 6.1. Prediction of panic severity As we have noted elsewhere (Casey, Oei, Newcombe, & Kenardy, 2004), although several studies point to the conclusion that both panic self-efficacy and catastrophic misinterpretation of bodily sensations are individually important in panic disorder (Michelson, Schwartz, & Marchione, 1991; Stoler & McNally, 1991; Telch, Silverman, & Schmidt, 1996), other studies suggest that only one or the other of these factors may be critical in determining panic severity (Borden, Clum, & Salmon, 1991; Hoffart, 1995b; Williams, 1995). One explanation of the contradictory findings in this area may involve methodological differences in the measurement of the relevant constructs in studies reported to date. For example, one method of measuring cognitions sometimes used by researchers in this area is to ask clients to ‘‘think aloud,’’ and then to report on independent ratings of content of these thoughts (e.g., Williams, Kinney, Harap, & Liebmann, 1997). However, this method may be compromised by a number of problems. Patients may have difficulties in verbalizing cognitions (McNally & Foa, 1996), while other evidence suggests more specifically that raters who do not suffer panic disorder themselves may fail to accurately identify the extent of negative cognitions (for review of this issue, see Khawaja & Oei, 1998). Similarly, attempts to measure self-efficacy that focus solely upon the behavioural domain (e.g., Hoffart, 1998) may fail to capture important aspects of this construct. A number of researchers have suggested that the avoidance behaviour characteristic of patients with agoraphobia is associated with low selfefficacy (e.g.,Cox, Endler, Swinson, & Norton, 1992; Craske, Rapee, & Barlow, 1988; Salkovskis & Hackmann, 1997). Another plausible explanation for some of the differing results in the literature to date may therefore concern the absence or presence of agoraphobia in the samples used to examine the relationship of the two cognitive factors. To address these issues, Casey, Oei et al. (2004) examined the relative contribution of catastrophic misinterpretation of bodily sensations and panic self-efficacy to the prediction of panic severity using a measure of panic self-efficacy that required patients to rate their confidence in controlling panic attacks not only in regard to situations, but also when experiencing thoughts and sensations associated with panic attacks. Furthermore, the role of panic self-efficacy in predicting panic severity was evaluated in this study only after controlling for the contribution of negative cognitions that associate bodily sensations with danger, using the Brief Bodily Sensations Intepretations Questionnaire (Clark et al., 1997), which is a well-recognized measure of catastrophic misinterpretation of bodily sensations. Each of the cognitive factors was found to independently predict panic severity in a large sample of patients
546
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
with panic disorder (with and without agoraphobia) awaiting treatment in a multicentre trial of CBT. Catastrophic misinterpretation of bodily sensations predicted panic severity in these patients after controlling for the influence of panic self-efficacy. Conversely, panic self-efficacy predicted panic severity after controlling for the influence of catastrophic misinterpretation of bodily sensations. Importantly, panic self-efficacy continued to independently predict panic severity (in addition to catastrophic misinterpretation of bodily sensations), after controlling for the presence or absence of agoraphobia. There was no evidence to support the utility of a single higher-order construct representing the interaction of the two cognitive factors in predicting panic severity. Although in need of replication, the findings of this study suggest that both catastrophic misinterpretation of bodily sensations and panic self-efficacy should be examined in examination of change processes associated with CBT for panic disorder. 6.2. Cognitive mediation of treatment effects: the role of panic self-efficacy and catastrophic misinterpretation of bodily sensations Reduction of panic severity has been shown to be associated with both decreases in danger-related cognitions and increases in self-efficacy (e.g., Bouchard, Gauthier, Laberge, & French, 1996; Cote, Gauthier, Laberge, & Cormier, 1994; Lidren, Watkins, Gould, & Clum, 1994). However, studies that report changes in both factors associated with treatment leave open the possibility that changes in one factor may be simply a coeffect of changes in the other factor (cf. Bandura, 1988). Moreover, studies that show cognitive changes are associated with successful outcome at the end of the treatment cannot discount the possibility that the change in cognitions has occurred as a consequence of improvement (Marks, Basoglu, & Noshirvani, 1994). Demonstration that symptom reduction is correlated with cognitive change in the course of therapy, or even more convincingly, that it is predicted by cognitive change during therapy would provide stronger support for the cognitive position (Oei, Llamas, & Devilly, 1999; Rachman, 1994). To date, only a handful of studies can be identified that have attempted to examine the temporal relationship of cognitive change and symptom reduction. Last, O’Brien, and Barlow (1985) examined the relationship between cognitions and anxiety in six agoraphobics during an exposure session. They found a significant, positive relationship between negative thinking (characterized as percentage of negative or maladaptive thoughts during exposure) and anxiety, with larger percentages of maladaptive thoughts corresponding with higher ratings of anxiety. Similarly, van Hout, Emmelkamp, and Scholing (1994) examined negative self-statements during exposure and found that the frequency of these statements best differentiated between least and most improved patients with panic disorder. Oei, Duckham, and Free (1989) have warned however that experimental artifacts (such as therapist reinforcement) may confound such changes in cognition reported through use of tape-recorded verbalisations and/or thought-listing. Attempting to overcome the various limitations of thought production techniques, van Hout, Emmelkamp, Koopmans, Boegels, and Bouman (2001) recently developed a measure called ‘‘the Agoraphobic Self-Statements Questionnaire,’’ which asks respondents to rate the frequency of both positive and negative self-statements. These self-statements were gathered by the researchers via thought-listing procedures in the first stage of development of this measure from agoraphobic patients undergoing behavioural avoidance tests. Although the study reported by van Hout et al. represents a useful attempt to expand the focus of research in this area by examining changes in both negative and
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
547
positive cognitions occurring during treatment, their results suggest limitations to the use of selfstatements as a way of accomplishing this goal. van Hout et al. reported that the frequency of both negative and positive self-statements decreased over the treatment period, and that further, there was no significant relationship between the positive self-statements and any of the symptom measures used in the study. A preferable strategy, therefore, is the measurement of specific factors derived from the theoretical constructs that inform the model of change processes (Oei et al., 1989). Adopting such an approach, other studies have tracked changes in theory-specific negative cognitions across the treatment period. Poulton and Andrews (1996) demonstrated that ratings of the likely occurrence of physical danger (‘‘you will have a heart attack’’) or psychological danger (‘‘you will go insane’’) reduced across a structured, intensive 3-week treatment period in (N = 45) patients diagnosed with panic disorder. However, these researchers failed to test whether these changes predicted symptom reduction. Salkovskis, Clark and Hackmann (1991) used a multiple baseline design to compare the use of focal cognitive therapy to nonfocal cognitive therapy in treatment of seven patients diagnosed with panic disorder. Whereas focal cognitive therapy focused on altering misinterpretations of sensations, nonfocal cognitive therapy omitted discussion of misinterpretations of sensations but used cognitive techniques to address stressors in the patients’ lives. Measures reported in this study were ratings of belief in catastrophic misinterpretations (that were taken by the therapists at baseline; and before and after each session) and panic attack frequency derived from a daily panic diary. Results indicated that use of focal cognitive therapy was associated with the reduction of panic frequency for six out of seven participants reported in this clinical series, whereas nonfocal cognitive therapy was not associated with reduction in panic frequency. However, neither of these two studies investigated the potential role of changes in positive cognitions during the treatment period. In a recent study (Casey, Newcombe, & Oei, in press), this question was examined by examining the differential role of changes in catastrophic misinterpretation of bodily sensations and panic self-efficacy during the treatment period in prediction of posttreatment panic severity through a comparison of a Standard CBT versus a Waitlist Condition. Changes in catastrophic misinterpretation of bodily sensations and panic self-efficacy contributed significantly more to prediction of panic severity than did assignment to either Standard CBT or a Waitlist Condition. Importantly, changes in both panic selfefficacy and catastrophic misinterpretation of bodily sensations independently predicted panic severity at posttreatment, suggesting that the effects of CBT upon panic severity were mediated by changes in both cognitive factors.
7. Theoretical implications: an ICM of panic disorder A strong conclusion suggested by the evidence reviewed in this paper is that the cognitive model may be usefully revised to take account of both negative and positive cognitions. Clark’s (1986) cognitive account of panic disorder postulates a vicious circle in which a sequence of events occur that are initiated by the perception of threat, and culminate in the occurrence of panic. Within this sequence described by Clark, catastrophic misinterpretation of bodily sensations serves as the cognition that mediates both panic attacks and the avoidance and disability associated with disorder. According to Clark, it is this vicious circle of events that maintains panic disorder. In contrast, evidence reviewed in this paper supports an ICM, in which both catastrophic misinterpretation of bodily sensations and panic selfefficacy independently contribute to maintenance of panic severity. On the basis of this evidence, it is
548
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
possible to reconceptualize the maintenance of panic disorder again as a sequential, repetitive process, but as one that involves both catastrophic misinterpretation of bodily sensations and panic self-efficacy as the mediating cognitions. Such a reconceptualization suggests that at a minimum, in the final cognitive pathway postulated in Clark’s circular account of panic, the ongoing occurrence of panic is influenced by both catastrophic misinterpretation of bodily sensations and panic self-efficacy, as shown in Fig. 2. Evidence reviewed earlier regarding the links between cognition, bodily sensations, and affect however would suggest that potentially the role of panic self-efficacy could be incorporated somewhat earlier within this reconceptualization of a cognitive account of panic. For example, there is broad agreement regarding the initiating role of the perception of threat in the vicious cycle that culminates in panic. According to both Bandura (1988) and Beck et al. (1985), however, even this initial event in the sequence of panic would be likely to involve both negative, danger-related cognitions and positive cognitions concerning control or coping. These cognitions would presumably then influence subsequent responses in the chain of events that determines the occurrence and maintenance of panic. For example, some studies suggest that positive cognitions may impact upon indices of physiological arousal, a finding that has intriguing implications for the role of panic self-efficacy in a condition such as panic disorder, whose hallmark is anxiety regarding physiological arousal. In a series of studies, Bandura et al. (reviewed in Bandura, 1988) have documented the relationship of lower self-efficacy to increased autonomic arousal, while other studies suggest that the experience of positive emotions can speed physiological recovery from the arousal associated with negative events (Fredrickson & Levenson, 1998). Along these lines, a model of change processes in treatment of panic disorder, which incorporates the role of both panic self-efficacy and catastrophic misinterpretation of bodily sensations, would suggest at a minimum that an increase in panic self-efficacy would operate as an additional point of exit from the vicious circle that maintains the disorder.
Fig. 2. Integrative Cognitive Model of panic attacks.
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
549
It is possible, however, to conceptualize the process of change during CBT for panic disorder in terms that have somewhat broader theoretical implications. An issue of ongoing debate within the broader CBT literature concerns the extent to which treatment works by changing the content of thought, as opposed to the way in which it operates through changing the form of processing thought (Teasdale et al., 2001). Persons (1993) described these two alternative perspectives on the process of change in CBT in the following terms: The schema change model proposes that cognitive therapy helps patients by teaching them that their central underlying attitudes, or schemata, are distorted and/or maladaptive. . .the compensatory skills model proposes that cognitive therapy helps patients by teaching them cognitive and behavioural skills for use when they experience. . .negative emotional experiences. (pp. 123–124) The first of these models refers to the underlying cognitive theory proposed by Beck et al. (1985) and Beck, Rush, Shaw, and Emery (1979) which emphasizes the role of changes in the content of thought, and from which Clark’s (1986) cognitive account of panic derives. The second of these models was first articulated by Barber and DeRubeis (1989), but is most closely aligned to a long tradition within CBT that emphasizes the importance of teaching coping skills. However, recent elaboration of Beck et al.’s cognitive theory of anxiety disorder in light of information-processing theories (Beck & Clark, 1997) suggests a way in which this compensatory skills model can be considerably updated. Inherent within the model of Beck and Clark is the concept that the initial processing of threat triggers a narrowing or constriction of cognitive awareness, which blocks off more constructive or reflexive modes of thinking. Consequently, a primary goal in therapy is to strengthen controlled processing which involves an individual’s assessment of coping resources. In these terms, an important step in a compensatory skills model is to broaden the scope of cognitive awareness, and in this context it is useful to consider two lines of recent research. Teasdale (1999a, 1999b has argued for the role of metacognitive insight as an important variable in CBT research, defining this insight as ‘‘ experiencing thoughts as events in the field of awareness, rather than as direct readouts on reality’’ (p. 146). In a similar vein, Safran and Segal (1990) have emphasized the importance of ‘‘decentering’’ in CBT, that they define as, ‘‘a process through which one is able to step outside of one’s immediate experience, thereby changing the very nature of that experience’’ (p. 117). As Hollon (1999) has commented, ‘‘we are not used to thinking about what we think’’ (p. 306). One of the primary aims of CBT is to teach patients to identify and to challenge thoughts associated with psychological distress. This explicit emphasis in CBT on assisting patients to comprehend, and make use of the distinction between facts and beliefs can therefore be seen as training in compensatory skills which broadens the scope of cognitive awareness, and thus enables the use of more constructive or reflexive modes of thinking. However, it is this focus in CBT that also enables a more fundamental change in underlying beliefs to occur, and in this sense, it seems likely that the process of change will involve the operation of both a schema change and a compensatory skills model. Secondly, in the context of emerging interest in positive factors in psychology (Gilham & Seligman, 1999), Fredrickson (2000, 2001, #1412) and Fredrickson and Levenson (1998) have proposed a broaden-and-build theory of positive emotions. Although implications of this theory have not as yet been clearly articulated in clinical terms, the broaden-and-build theory provides a more focused perspective on the role that increases in positive cognitions, such as panic self-efficacy, may play in treatment of psychological disorders. According to Fredrickson, the experience of positive emotions both broadens people’s momentary
550
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
thought–action repertoires, and builds enduring personal resources. Although she has couched her theory within the general framework of emotion theory, Fredrickson has argued for a reciprocal relationship between positive cognitions and positive emotions, and thus provides for an analysis of cognitive change processes within CBT that strengthens the case for an independent role for positive cognitions. She has described this role in the following terms: One function of certain positive emotions may not be to spark specific action (as do many negative emotions), but instead may be to loosen the hold that these negative emotions gain on an individual’s mind and body by dismantling, or undoing this psychological and physiological preparation for specific action. (Fredrickson & Levenson, 1998, p. 193) According to this analysis, an increase in positive cognitions concerning control or coping may in turn produce an increase in positive emotions, and thus create an upward spiral characterized by a reciprocal relationship between positive emotions and positive cognitions. Evidence consistent with this proposition comes from a study that used ambulatory monitoring procedures to examine cognitions, physiological arousal, and subjective arousal in panic-disordered patients during exposure to feared situations (Kenardy et al., 1993). Results of this study indicated that, whereas reporting of negative thought sequences appeared stable over time, reporting of positive thoughts appeared to lead to an increasing likelihood of reporting further positive thoughts. In regard to panic disorder, therefore, a reciprocal, escalating relationship between positive cognitions and emotions may serve to broaden, in Beck and Clark’s (1997) terms, the scope of cognitive awareness and further enable the use of more constructive modes of thinking. Again, the most likely outcome of such a process may be both a shift in underlying beliefs, as well as an increase in compensatory skills, suggesting that both a decrease in negative cognitions and an increase in positive cognitions will play a role in reducing symptomatic severity. Such a conclusion is highly consistent with the empirical findings reported in this paper. Furthermore, evidence that catastrophic misinterpretation of bodily sensations and panic self-efficacy appear to play an independent role in panic disorder may provide a useful template for investigation of the contribution of negative and positive cognitions in other psychological disorders.
8. Summary and conclusion To date, cognitive mediational accounts of panic disorder have focused on the construct of catastrophic misinterpretation of bodily sensations, despite the suggestion that positive cognitions concerning coping or control may be important in this disorder. Clark (1986, 1996) has outlined a number of predictions in regard to the role of catastrophic misinterpretation of bodily sensations in panic disorder and, as reviewed in this paper, evidence consistent with these predictions supports the hypothesized role of this construct in panic disorder. However, it was argued that a number of studies suggest each of these predictions may be strengthened by the inclusion of panic self-efficacy to provide a more coherent account of the role of cognitions in panic disorder. Similarly, evidence regarding the link between cognitions, bodily sensations, and affect in panic disorder suggests that panic self-efficacy is likely to be involved in both determining and ultimately disrupting this link. Finally, review of processes argued to be central in effective treatment of this disorder suggested that panic self-efficacy appears
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
551
highly likely to be involved in change processes that result in the reduction of panic severity. Although empirical studies in this area have largely been characterized by a failure to adequately examine the role of both negative and positive cognitions, more recent studies suggest that both catastrophic misinterpretation of bodily sensations and panic self-efficacy play an important and independent role in panic disorder. The theoretical implications of this evidence were discussed in terms of an integrated cognitive account of panic disorder that draws upon both schema change and compensatory skills models of change processes in CBT, and as such, have considerable implications for research and treatment of other psychological disorders. References Antony, M. M., & Barlow, D. H. (1996). Information-processing and emotion theory views of panic disorder: Overlapping and distinct features. In R. M. Rapee (Ed.), Current controversies in the anxiety disorders ( pp. 94 – 98). New York: Guilford Press. Bandelow, B., Hajak, G., Holzrichter, S., Kunert, H. J., & Ruther, E. (1995). Assessing the efficacy of treatments for panic disorder and agoraphobia: 1. Methodological problems. International Clinical Psychopharmacology, 10, 83 – 93. Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84, 191 – 215. Bandura, A. (1983). Self-efficacy determinants of anticipated fears and calamities. Journal of Personality and Social Psychology, 45(2), 464 – 468. Bandura, A. (1988). Self-efficacy conception of anxiety. Anxiety Research, 1, 77 – 98. Barber, J. P., & DeRubeis, R. J. (1989). On second thought: Where the action is in cognitive therapy for depression. Cognitive Therapy and Research, 13(5), 441 – 457. Barlow, D. H. (1988). Anxiety and its disorders: The nature and treatment of anxiety and panic. New York: Guilford Press. Barlow, D. H. (1991). Disorders of emotion. Psychological Inquiry, 2, 58 – 71. Basoglu, M., Marks, I. M., & Sengun, S. (1992). A prospective study of panic and anxiety in agoraphobia with panic disorder. British Journal of Psychiatry, 160, 57 – 64. Beck, A. T., & Clark, D. A. (1997). An information processing model of anxiety: Automatic and strategic processes. Behaviour Research and Therapy, 35(1), 49 – 58. Beck, A. T., Emery, G., & Greenberg, R. L. (1985). Anxiety disorders and phobias: A cognitive perspective. New York: Basic Books. Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive therapy of depression. New York: Guilford Press. Borden, J. W., Clum, G. A., & Salmon, P. G. (1991). Mechanisms of change in the treatment of panic. Cognitive Therapy and Research, 15, 257 – 272. Bouchard, S., Gauthier, J., Laberge, B., & French, D. (1996). Exposure versus cognitive restructuring in the treatment of panic disorder with agoraphobia. Behaviour Research and Therapy, 34, 213 – 224. Brim, O. G. J. (1976). Life span development of the theory of oneself: Implications for child development. In H. W. Reese (Ed.), Advances in Child Development and Behavior, vol. 11. New York: Academic Press. Casey, L. M., Oei, T. P. O., Newcombe, P. A. (in press). Cognitive Mediation of Panic Severity. Cognitive Therapy and Research. Casey, L. M., Oei, T. P. O., Newcombe, P. A., & Kenardy, J. (2004). The role of catastrophic misinterpretation of bodily sensations and panic self-efficacy in predicting panic severity. Journal of Anxiety Disorders, 18, 325 – 340. Cervone, D. (2000). Thinking about self-efficacy. Behavior Modification, 24(1), 30 – 56. Chambless, D. L., Beck, A. T., Gracely, E. J., & Grisham, J. R. (2000). Relationship of cognitions to fear of somatic symptoms: A test of the cognitive theory of panic. Depression and Anxiety, 11, 1 – 9. Clark, D. M. (1986). A cognitive model of panic. Behaviour Research and Therapy, 24, 461 – 470. Clark, D. M. (1988). A cognitive model of panic attacks. In S. Rachman, & J. D. Maser (Eds.), Panic: Psychological perspectives ( pp. 71 – 89). Hillsdale, NJ: Lawrence Erlbaum Associates. Clark, D. M. (1993a). Cognitive mediation of panic attacks induced by biological challenge tests. Advances in Behaviour Research and Therapy, 15(1), 75 – 84.
552
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
Clark, D. M. (1993b). Cognitive mediation of panic attacks induced by biological challenge tests. Special issue: Panic, cognitions and sensations. Advances in Behaviour Research and Therapy, 15, 75 – 84. Clark, D. M. (1994). Cognitive therapy for panic disorder. In B. E. Wolfe, & J. D. Maser (Eds.), Treatment of panic disorder: A consensus development conference ( pp. 121 – 132). Washington, DC: American Psychiatric Press. Clark, D. M. (1996). Panic disorder: From theory to theraphy. In P. M. Salkovskis (Ed.), Frontiers of Cognitive Therapy ( pp. 318 – 344). Guildford Press. Clark, D. M. (1999). Anxiety disorders: Why they persist and how to treat them. Behaviour Research and Therapy, 37(Suppl. 1), S5 – S57. Clark, D. M., Salkovskis, P. M., Hackmann, A., Middleton, H., Anastasiades, P., & Gelder, M. (1994). A comparison of cognitive therapy, applied relaxation and imipramine in the treatment of panic disorder. British Journal of Psychiatry, 164, 759 – 769. Clark, D. M., Salkovskis, P. M., Ost, L. G., Breitholtz, E., Koehler, K., Westling, B. E., Jeavons, A., & Gelder, M. (1997). Misinterpretation of body sensations in panic disorder. Journal of Consulting and Clinical Psychology, 65(2), 203 – 213. Clum, G. A., & Knowles, S. L. (1991). Why do some people with panic disorders become avoidant? A review. Clinical Psychology Review, 11(3), 295 – 313. Cote, G., Gauthier, J. G., Laberge, B., & Cormier, H. J. (1994). Reduced therapist contact in the cognitive behavioral treatment of panic disorder. Behavior Therapy, 25, 123 – 145. Cox, B. J. (1996). The nature and assessment of catastrophic thoughts in panic disorder. Behaviour Research and Therapy, 34, 363 – 374. Cox, B. J., Endler, N. S., Swinson, R. P., & Norton, G. R. (1992). Situations and specific coping strategies associated with clinical and nonclinical panic attacks. Behaviour Research and Therapy, 30(1), 67 – 69. Craske, M. G., & Barlow, D. H. (1988). A review of the relationship between panic and avoidance. Clinical Psychology Review, 8(6), 667 – 685. Craske, M. G., Rapee, R. M., & Barlow, D. H. (1988). The significance of panic-expectancy for individual patterns of avoidance. Behavior Therapy, 19, 577 – 592. Davey, G. C. L. (1995). Rumination and the enhancement of fear: Some laboratory findings. Behavioural and Cognitive Psychotherapy, 23(3), 203 – 215. Ehlers, A. (1993). Interoception and panic disorder. Advances in Behaviour Research and Therapy, 15(1), 3 – 21. Fehm, L., & Margraf, J. (2002). Thought suppression: Specificity in agoraphobia versus broad impairment in social phobia? Behaviour Research and Therapy, 40, 57 – 66. Foa, E. B., & Kozak, M. J. (1986). Emotional processing of fear: Exposure to corrective information. Psychological Bulletin, 99, 20 – 35. Folkman, S., & Moskowitz, J. T. (2000). Positive affect and the other side of coping. American Psychologist, 55(6), 647 – 654. Fraenkel, Y., Kindler, S., & Melmed, R. (1997). Differences in cognitions during chest pain of patients with panic disorder and ischemic heart disease. Depression and Anxiety, 4, 217 – 222. Fredrickson, B. L. (2000). Cultivating positive emotions to optimize health and well-being. Prevention and Treatment, 3. p. 193. Fredrickson, B. L. (2001). The role of positive emotions in positive psychology: The broaden-and-build theory of positive emotions. American Psychologist, 56(3), 218 – 226. Fredrickson, B. L., & Levenson, R. W. (1998). Positive emotions speed recovery from the cardiovascular sequelae of negative emotions. Cognition and Emotion, 12(2), 191 – 220. Gilham, J. E., & Seligman, M. E. P. (1999). Footsteps on the road to a positive psychology. Behaviour Research and Therapy, 37, 163 – 173. Hackmann, A. (1997). The transformation of meaning in cognitive therapy. In M. J. Power, & C. R. Brewin (Eds.), The transformation of meaning in psychological therapies: Integrating theory and practice ( pp. 125 – 140). Chichester, UK: Wiley. Hoffart, A. (1995a). Cognitive mediators of situation fear in agoraphobia. Journal of Behavior Therapy and Experimental Psychiatry, 26, 313 – 320. Hoffart, A. (1995b). A comparison of cognitive and guided mastery therapy of agoraphobia. Behaviour Research and Therapy, 33(4), 423 – 434. Hoffart, A. (1998). Cognitive and guided mastery therapy of agoraphobia: Long-term outcome and mechanisms of change. Cognitive Therapy and Research, 22(3), 195 – 207.
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
553
Hollifield, M., Katon, W., Skipper, B., & Chapman, T. (1997). Panic disorder and quality of life: Variables predictive of functional impairment. American Journal of Psychiatry, 154, 766 – 772. Hollon, S. D. (1999). Rapid early response in cognitive behavior therapy: A commentary. Clinical Psychology: Science and Practice, 6(3), 305 – 309. Kenardy, J., Evans, L., & Oei, T. P. (1992). The latent structure of anxiety symptoms in anxiety disorders. American Journal of Psychiatry, 149, 1058 – 1061. Kenardy, J., Oei, T. P., Weir, D., & Evans, L. (1993). Phobic anxiety in panic disorder: Cognition, heart rate, and subjective anxiety. Journal of Anxiety Disorders, 7, 359 – 371. Khawaja, N. G., & Oei, T. P. S. (1998). Catastrophic cognitions in panic disorder with and without agoraphobia. Clinical Psychology Review, 18(3), 341 – 365. Khawaja, N. G., & Oei, T. P. S. (1999). The psychobiological correlates of panic attacks during in vivo exposure. Behavioural and Cognitive Psychotherapy, 27(4), 353 – 367. Lang, P. J. (1979). A bio-informational theory of emotional imagery. Psychophysiology, 16(6), 495 – 512. Last, C. G., O’Brien, G. T., & Barlow, D. H. (1985). The relationship between cognitions and anxiety: A preliminary report. Behavior Modification, 9, 235 – 241. Lidren, D. M., Watkins, P. L., Gould, R. A., & Clum, G. A. (1994). A comparison of bibliotherapy and group therapy in the treatment of panic disorder. Journal of Consulting and Clinical Psychology, 62, 865 – 869. MacLeod, A. K., & Moore, R. (2000). Positive thinking revisited: Positive cognitions, well-being and mental health. Clinical Psychology and Psychotherapy, 7, 1 – 10. Maddux, J. E. (1995). Self-efficacy theory: An introduction. In J. E. Maddux (Ed.), Self-efficacy, adaptation, and adjustment: Theory, research, and application ( pp. 3 – 33). New York: Plenum. Maidenberg, E., Chen, E., Craske, M., Bohn, P., & Bystritsky, A. (1996). Specificity of attentional bias in panic disorder and social phobia. Journal of Anxiety Disorders, 10(6), 529 – 541. Mansell, W. (2000). Conscious appraisal and the modification of automatic processes in anxiety. Behavioural and Cognitive Psychotherapy, 28(2), 99 – 120. Marks, I., Basoglu, M., & Noshirvani, H. (1994). Cognitive therapy in panic disorder. British Journal of Psychiatry, 165, 556–557. Marks, M. P., Basoglu, M., Alkubaisy, T., & Sengun, S. (1991). Are anxiety symptoms and catastrophic cognitions directly related? Journal of Anxiety Disorders, 5, 247 – 254. Matthews, G., & Wells, A. (2000). Attention, automaticity, and affective disorder. Behavior Modification, 24(1), 69 – 93. McNally, R. J. (1994). Panic disorder: A critical analysis. New York: Guilford Press. McNally, R. J. (1999). Panic and phobias. In T. Dalgleish, & M. J. Power (Eds.), Handbook of cognition and emotion ( pp. 479 – 496). Chichester, UK: Wiley. McNally, R. J. (2001). On the scientific status of cognitive appraisal models of anxiety disorder. Behaviour Research and Therapy, 39(5), 513 – 521. McNally, R. J., & Foa, E. B. (1996). The limitations of self efficacy theory in explaining therapeutic changes in phobic behaviour. In R. M. Rapee (Ed.), Current controversies in the anxiety disorders ( pp. 369 – 372). New York: Guilford Press. McNally, R. J., Hornig, C. D., Hoffman, E. C., & Han, E. M. (1999). Anxiety sensitivity and cognitive biases for threat. Behavior Therapy, 30(1), 51 – 61. Michelson, L. K., Schwartz, R. M., & Marchione, K. E. (1991). States-of-mind model: Cognitive balance in the treatment of agoraphobia: II. Advances in Behaviour Research and Therapy, 13(4), 193 – 213. Mineka, S., & Thomas, C. (1999). Mechanisms of change in exposure therapy for anxiety disorders. In T. Dalgleish, & M. J. Power (Eds.), Handbook of cognition and emotion ( pp. 747 – 764). Chichester, UK: Wiley. Oei, T. P., Duckham, S., & Free, M. (1989). Does cognitive behaviour therapy support cognitive models of depression? Special issue: Depression: Treatment and theory. Behaviour Change, 6, 70 – 75. Oei, T. P. S., Llamas, M., & Devilly, G. J. (1999). The efficacy and cognitive processes of cognitive behaviour therapy in the treatment of panic disorder with agoraphobia. Behavioural and Cognitive Psychotherapy, 27(1), 63 – 88. Pauli, P., Marquardt, C., Hartl, L., Nutzinger, D. O., Holzl, R., & Strian, F. (1991). Anxiety induced by cardiac perceptions in patients with panic attacks: A field study. Behaviour Research and Therapy, 29(2), 137 – 145. Persons, J. B. (1993). The process of change in cognitive therapy: Schema change or acquisition of compensatory skills? Cognitive Therapy and Research, 17(2), 123 – 137.
554
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
Poulton, R. G., & Andrews, G. (1996). Change in danger cognitions in agoraphobia and social phobia during treatment. Behaviour Research and Therapy, 34(5 – 6), 413 – 421. Power, M., & Dalgleish, T. (1997). Cognition and emotion: From order to disorder. UK: Psychology Press. Rachman, S. (1988). Panics and their consequences: A review and prospect. In S. Rachman, & J. D. Maser (Eds.), Panic: Psychological perspectives ( pp. 259 – 303). Hillsdale, NJ: Lawrence Erlbaum Associates. Rachman, S. (1994). Psychological treatment of panic: Mechanisms. In B. E. Wolfe, & J. D. Maser (Eds.), Treatment of panic disorder: A consensus development conference ( pp. 133 – 148). Washington, DC: American Psychiatric Press. Rachman, S., Levitt, K., & Lopatka, C. (1987). Panic: The links between cognitions and bodily symptoms: I. Behaviour Research and Therapy, 25(5), 411 – 423. Rachman, S. J. (1990). Fear and courage. (2nd ed.). New York: W.H. Freeman. Rapee, R. M. (1994). Detection of somatic sensations in panic disorder. Behaviour Research and Therapy, 32, 825 – 831. Rapee, R. M. (1995). Psychological factors influencing the affective response to biological challenge procedures in panic disorder. Journal of Anxiety Disorders, 9(1), 59 – 74. Rapee, R. M. (1996). Information processing views of panic disorder. In R. M. Rapee (Ed.), Current controversies in the anxiety disorders ( pp. 77 – 93). New York: Guilford Press. Reiss, S., & McNally, R. J. (1985). Expectancy model of fear. In S. R. R. R. Bootzin (Ed.), Theoretical issues in behaviour therapy ( pp. 107 – 121). San Diego, CA: Academic Press. Safran, J. D., & Segal, Z. V. (1990). Interpersonal process in cognitive therapy. New York: Basic Books. Salkovskis, P. M. (1988). Phenomenology, assessment, and the cognitive model of panic. In S. Rachman, & J. D. Maser (Eds.), Panic: Psychological perspectives ( pp. 111 – 136). Hillsdale, NJ: Lawrence Erlbaum Associates. Salkovskis, P. M. (1998). Panic disorder and agoraphobia. In P. M. Salkovskis (Ed.), Comprehensive clinical psychology, vol. 6 ( pp. 399 – 437). Oxford, England: Pergamon/Elsevier Science. Salkovskis, P. M., Clark, D. M., & Gelder, M. G. (1996). Cognition – behaviour links in the persistence of panic. Behaviour Research and Therapy, 34, 453 – 458. Salkovskis, P. M., Clark, D. M., & Hackmann, A. (1991). Treatment of panic attacks using cognitive therapy without exposure or breathing retraining. Behaviour Research and Therapy, 29, 161 – 166. Salkovskis, P. M., Clark, D. M., Hackmann, A., Wells, A., & Gelder, M. G. (1999). An experimental investigation of the role of safety-seeking behaviours in the maintenance of panic disorder with agoraphobia. Behaviour Research and Therapy, 37(6), 559 – 574. Salkovskis, P. M., & Hackmann, A. (1997). Agoraphobia. In G. C. L. Davey (Ed.), Phobias—a handbook of theory, research and treatment ( pp. 27 – 61). Chichester: Wiley. Schmidt, N. B., Lerew, D. R., & Jackson, R. J. (1997). The role of anxiety sensitivity in the pathogenesis of panic: Prospective evaluation of spontaneous panic attacks during acute stress. Journal of Abnormal Psychology, 106(3), 355 – 364. Schmidt, N. B., Lerew, D. R., & Trakowski, J. H. (1997). Body vigilance in panic disorder: Evaluating attention to bodily perturbations. Journal of Consulting and Clinical Psychology, 65, 214 – 220. Schmidt, N. B., Trakowski, J. H., & Staab, J. P. (1997). Extinction of panicogenic effects of a 35% CO2 challenge in patients with panic disorder. Journal of Abnormal Psychology, 106(4), 630 – 638. Schniering, C. A., & Rapee, R. M. (1997). A test of the cognitive model of panic: Primed lexical decision in panic disorder. Journal of Anxiety Disorders, 11(6), 557 – 571. Seligman, M. E. P. (1988). Competing theories of panic. In S. Rachman, & J. D. Maser (Eds.), Panic: Psychological perspectives ( pp. 321 – 329). Hillsdale, NJ: Lawrence Erlbaum Associates. Shear, M. K. (1991). The concept of uncontrollability. Psychological Inquiry, 2, 88 – 93. Stoler, L. S., & McNally, R. J. (1991). Cognitive bias in symptomatic and recovered agoraphobics. Behaviour Research and Therapy, 29, 539 – 545. Taylor, C. B., Sheikh, J., Agras, W. S., Roth, W. T., Margraf, J., Ehlers, A., Maddock, R. J., & Goddard, D. (1986). Ambulatory heart rate changes in patients with panic attacks. American Journal of Psychiatry, 143, 478 – 482. Taylor, C. B., Telch, M. J., & Havvik, D. (1983). Ambulatory heart rate changes during panic attacks. Journal of Psychiatric Research, 17, 261 – 266. Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: A social psychological perspective on mental health. Psychological Bulletin, 103(2), 193 – 210. Taylor, S. E., Kemeny, M. E., Reed, G. M., Bower, J. E., & Gruenewald, T. L. (2000). Psychological resources, positive illusions, and health. American Psychologist, 55(1), 99 – 109.
L.M. Casey et al. / Clinical Psychology Review 24 (2004) 529–555
555
Teasdale, J. (1999a). Metacognition, mindfulness and the modification of mood disorders. Clinical Psychology and Psychotherapy, 6, 146 – 155. Teasdale, J. D. (1993). Emotion and two kinds of meaning: Cognitive therapy and applied cognitive science. Behaviour Research and Therapy, 31, 339 – 354. Teasdale, J. D. (1999b). Multi-level theories of cognition – emotion relations. In T. Dalgleish, & M. J. Power (Eds.), Handbook of cognition and emotion ( pp. 665 – 681). Chichester, UK: Wiley. Teasdale, J. D., Scott, J., Moore, R. G., Hayhurst, H., Pope, M., & Paykel, E. S. (2001). How does cognitive therapy prevent relapse in residual depression? Evidence from a controlled trial. Journal of Consulting and Clinical Psychology, 69(3), 347 – 357. Telch, M. J., Brouillard, M., Telch, C. F., Agras, W. S., & Taylor, C. B. (1989). Role of cognitive appraisal in panic-related avoidance. Behaviour Research and Therapy, 27(4), 373 – 383. Telch, M. J., Schmidt, N. B., Jaimez, T. L., & Jacquin, K. M. (1995). Impact of cognitive – behavioral treatment on quality of life in panic disorder patients. Journal of Consulting and Clinical Psychology, 63, 823 – 830. Telch, M. J., Silverman, A., & Schmidt, N. B. (1996). Effects of anxiety sensitivity and perceived control on emotional responding to caffeine challenge. Journal of Anxiety Disorders, 10(1), 21 – 35. Van der Does, A. J. W., Antony, M. M., Ehlers, A., & Barsky, A. J. (2000). Heartbeat perception in panic disorder: A reanalysis. Behaviour Research and Therapy, 38(1), 47 – 62. van Hout, W. J. P. J., Emmelkamp, P. M. G., Koopmans, P. C., Boegels, S. M., & Bouman, T. K. (2001). Assessment of selfstatements in agoraphobic situations: Construction and psychometric evaluation of the Agoraphobic Self-Statements Questionnaire (ASQ). Journal of Anxiety Disorders, 15(3), 183 – 201. van Hout, W. J. P. J., Emmelkamp, P. M. G., & Scholing, A. (1994). The role of negative self-statements during exposure in vivo: A process study of eight panic disorder patients with agoraphobia. Behavior Modification, 18, 389 – 410. Wells, A., & Matthews, G. (1996). Modelling cognition in emotional disorder: The S-REF model. Behaviour Research and Therapy, 34(11 – 12), 881 – 888. Westling, B. E., & Ost, L. G. (1995). Cognitive bias in panic disorder patients and changes after cognitive – behavioral treatments. Behaviour Research and Therapy, 33, 585 – 588. Williams, S. L. (1990). Guided mastery treatment of agoraphobia: Beyond stimulus exposure. In M. Hersen, R. M. Eisler, & P. M. Miller (Eds.), Progress in Behavior Modification, vol. 26 ( pp. 89 – 121). Newbury Park, CA: Sage Publications. Williams, S. L. (1995). Self-efficacy, anxiety, and phobic disorders. In J. E. Maddux (Ed.), Self-efficacy, adaptation, and adjustment: Theory, research, and application ( pp. 69 – 107). New York: Plenum. Williams, S. L., & Falbo, J. (1996). Cognitive and performance-based treatments for panic attacks in people with varying degrees of agoraphobic disability. Behaviour Research and Therapy, 34, 253 – 264. Williams, S. L., Kinney, P. J., Harap, S. T., & Liebmann, M. (1997). Thoughts of agoraphobic people during scary tasks. Journal of Abnormal Psychology, 106(4), 511 – 520. Zoellner, L. A., & Craske, M. G. (1999). Interoceptive accuracy and panic. Behaviour Research and Therapy, 37(12), 1141 – 1158.