behavioural psychotherapy

8/13/2019 behavioural psychotherapy

1/14

Behavioural Psychotherapy 1991, 19, 6-19

The Importance of Behaviour in theMaintenance of Anxiety and Panic:A Cognitive AccountPaul M SalkovskisUniversity of Oxford Department of Psychiatry WarnefordHospital Oxford

The theoretical and empirical basis of commonly accepted propositionsconcerning the role of behaviour in the practice of behavioural psycho-therapy for anxiety problems is considered. A number of problems areidentified, and an alternative, more explicitly cognitive hypothesis isdescribed. According to this cognitive account, there is both a closerelationship and specific interactions between threat cognitions andsafety seekingbehaviour .For any individual, safety seeking behaviourarises out of, and is logically linked to, the perception of serious threat.Such behaviour may be anticipatory (avoidant) or consequent (escape).Because safety seeking behaviour is perceived to be preventative, andfocused on especially negative consequences (e.g. death, illness, humili-ation), spontaneous disconfirmation of threat is made particularlyunlikely by such safety seeking behaviours. By preventing disconfir-mation of threat-related cognitions, safety seeking behaviour may be acrucial factor in the maintenance of anxiety disorders. The implicationsof this view for the understanding and treatment of anxiety disordersare discussed.

Th e ado ptio n of a subjective /cognitive elem ent in the thr ee systems analy-sis (Rachman and Hodgson, 1974) probably began the process of accept-ance by behaviour therapy that cognitions might be involved in the mainten-ance of psychological problems. More recently, there has been furtherprogression (or, as Blackburn, 1986, has suggested, an evolution), with cogn itive-behav iour th er ap y increasingly becoming a dom inating influ-ence on the practice of behavioural psychotherapy (Hawton et ai 1989).The reasons for the ready adoption of cognitive approaches by behaviourtherapists are many (Salkovskis, 1986); perhaps most important has been

Acknowledgements: The ideas in this article were developed in collaboration with DavidM. Clark, to whom the author is greatly indebted. The author is also grateful to the MedicalResearch Council of the United Kingdom for their support.

Reprint requests to Dr P. M. Salkovskis, Department of Psychiatry, Warneford Hospital,Oxford OX3 7JX, England, UK.

1991 British Association for Behavioural Psy chothe rapy


2/14

Behaviour and cognitions in anxiety 7the insistence of cognitive theorists that therapy be based on empiricallytestable hypotheses and the generally complementary relationship betweenprevious behavioural theories and the newer cognitive ones. For example,Clark's (1986) cognitive analysis of the relationship between sensationsperceived during panic attacks, their catastrophic misinterpretation and theconsequent intense anxiety has proved an important complement to pre-vious behavioural treatments of agoraphobia, in which panic attacks oftencould not be dealt with adequately by an exclusive focus on the modificationof avoidant behaviour.

Ho we ver, there remain a num ber of tensions between behav iour th erapyand cognitive therapy at the theoretical level. Nowhere is this tension moreapparent than in accounts of avoidance behaviour, anxiety and the role ofexposure. The failure of cognitive theories of anxiety to provide a compre-hensive account of the role of anxiety-related behaviour has probably beenthe greatest obstacle to the more general adoption of cognitive theories andtreatments by behavioural psychotherapists. The result has been that, attimes, cognitive-behavioural treatment of anxiety disorders has tended tobe a hybrid of techniques drawn from both traditions, without a consistentset of guiding principles to facilitate assessment and intervention.

The purpose of the present paper is to reconsider the theoretical basis ofthe relationship between cognition and behaviour, and how this interactionrelates to the cognitive-behavioural hypotheses of anxiety in which cog-nition of threat or danger is afforded a primary role. In order to do this,the behavioural theory of anxiety will be reviewed and evaluated, and somespecific extensions proposed. The cognitive hypothesis is also comparedwith learning theory modifications based on a pre pa red ne ss accountintended to deal with some of the same problems. The clinical implicationsof the cognitive-behavioural hypothesis of avoidant behaviour will be con-sidered.The theoretical basis of behaviour therapy: some problemsThe most influential behavioural theor y of phobia s is the tw o process mo del(Rachman, 1977), which proposes that phobias initially arise as a result ofthe phobic stimuli having previously been associated with aversive conse-quences or situations. The subsequent failure of phobias to extinguish issaid to be due to both (i) the way in which avoidance behaviour preventsthe occurrence of actual exposure to the feared stimuli and (ii) the way inwhich exposure to feared stimuli is terminated or shortened by escapebehaviour (when exposure to the feared stimulus unavoidably occurs)(Rachman, 1977). In turn, the persistence of avoidance and escapebehaviours is accounted for by the negative reinforcement associated with


3/14

8 P. M. Salkovskisthe om ission or term ina tion of anxiety ( anxiety relie f ). The clinicalimportance of this issue lies in the way exposure to feared stimuli is con-ducted; patients are invariably instructed to remain in the phobic situationuntil their anxiety has begun to decline and never to leave the situationwhen their anxiety was increasing or even remained above the initial level(e.g. Mathews, Gelder and Johnston, 1981). This theoretical view formedthe basis for the development of graded exposure as the treatment of choicefor phob ic anxiety (Rachm an, 1990a). H ow eve r, therap eutic success is anotoriously misleading way of evaluating theories, especially when theeffectiveness of the rap y has itself played a pa rt in the dev elop ment of thattheory, as has been the case for exposure based treatment. More specificquestions arising from behavioural theory are therefore considered next.(See also Rachman, 1990b, for a further critique in the context of simplephobias.)Does escape in the face of anxiety strengthen anxiety and later escape?Rachman and his colleagues have recently attempted to test directly the keytheoretical premise that, once a phob ia has developed, longer-term fearreduction (extinction) is prevented by escape behaviour (de Silva and Rach-man, 1984; Rachman, Craske, Tallman and Solyom, 1986). Rachman andcolleagues reasoned that, if the two process model is a valid basis to accountfor the persistence of phobic behaviour and for its effective treatment, thenit should be possible to use an instructional set which involves a directcontrast between treatment by exposure on the one hand with an escapeprocedure. Subjects were told to enter the feared situation and either (i)remain until their anxiety declined (as in exposure based treatments) or (ii)leave when their anxiety rose to 70 on a hundred point scale (as is hypothes-ized to normally occur). Unfortunately, the manipulation did not fullysucceed, and escape rarely took place in the experimental condition. Bothgroups showed anxiety reduction. There was, however, some evidence thatthe group given the escape option experienced (non-significantly) greaterreductions of anxiety than those told to remain in the situation, an appar-ently paradoxical finding not consistent with the prediction of thebehavioural model. Although methodological considerations limit the con-clusiveness of these studies (e.g. the subjects allowed to esca pe seldomdid so), they are nonetheless difficult to account for in terms of purelybehavioural accounts; in particular, they call into question the standard do n' t escape instructions as a basic com pon ent of exposure treatment.

In fact, Rachman opts for a cognitive account, suggesting that perceivedcontrol over the possibility of panic may have been a key factor (cf. Rach-man, 1990a, chapter 17; Sanderson, Rapee and Barlow, 1989). It remains


4/14

Behaviour and cognitions in anxiety 9difficult to explain, given the low levels of actual escape, why the perceptionof control over escape should have produced an apparent anxiety reducingeffect compared to the more normal situation experienced by these patients,in which they presumably would feel more readily able to escape (in thatthe experimental situation would be expected to produce a certain amountof social pressure not to escape; see also Rachman, 1990a, chapter 18, p.277). Further research on this topic is clearly very important given that theresults of the first two experiments were so tentative.Why are anxiety control strategies other than exposure sometimes helpful?If the conditioning hypothesis of anxiety is correct, then it also follows thatprocedures which the patient uses to control his or her response to fearfulstimuli (or the stimuli themselves) should be counter-productive (includingescape as outlined above, anxiety management treatments, distraction, cog-nitive therapy and so on). Indeed , Borko vec (1982) described the imp ortanceof what he termed functional CS expo sure , and of dealing with anybehaviours which interfere with such exposure. He pointed out that, forextinction to take place in a rapid and enduring fashion, it was not onlyimportant for the patient to be exposed to the conditioned anxiety stimulusfor long perio ds, bu t also for the patie nt to be actively en ga ge d wi th thestimulus; i.e. attending to it. Later descriptions of em otion al pr oces sing(Rachman, 1981) took a similar view, but extended the definition of thestimulus to include fear responses themselves. The idea of functional CSexposure makes sense in terms of the behavioural theo ry of the maintenanceof phobic anxiety, and as such is a helpful clarification. However, it high-lights a prob lem in accou nting for the effectiveness of trea tm ents w hich doinvolve anxiety control strategies (e.g. in the treatment of Panic Disorderwithout avoidance; Clark, 1988; social anxiety; Mattick and Peters, 1988;Mattick, Peters and Clarke, 1989),

The differentiation of co pi ng from avoidance is an almost unm ention edproblem which has been implicit in behaviour therapy from the early devel-opm ent of the expo sure prin ciple (M arks, 1987b) and its antecedents(Paul, 1966)1. Why do some behaviours such as distraction, which is saidto constitute avoidance (and which therefore should be blocked in thecourse of effective exposure-based therapy), become an effective and valuedcomp onent of a treatmen t such as anxiety manag ement wh en presented tothe patient as therapy? What is the difference between anxiety managementprocedures, which are effective in producing significant reductions in severeclinical anxiety, and similar strategies which such patients are reported ascommonly using prior to treatment (Butler, Cullington, Hibbert, Klimesand Gelder, 1987). We need to be able to conceptualize this apparently


5/14

10 P. M.Salkovskisfundamental difference between coping behaviour andavoidance behaviour,when both can be topographically similar. The cognitive hypothesis ofanxiety is able to offer a specific and empirically testable solution to thisclinically crucial issue.A cognitive account of avoidanceFrom a cognitive perspective, all anxiety-related phenomena arise fromcognitions of threat or danger. The cognitive account of anxiety and phobiasis also a learning account, consistent with current cognitive adaptations oflearning theo ry (van den H o u t and M erkelbach, 1991; Pow er, 1991,Resco rla, 1988). T hu s, cognitive and behavio ural formu lations of the acqui-sition of anxiety problems and phobias are identical, only differing in theterminology used to describe what is learned. However, some differencesemerge in terms of how anxiety is maintained. Th is includes the avoidantand escape behaviour characteristic of anxiety disorders. For example,according to the cognitive hypothesis of panic (Clark, 1986), the cognitionsinvolved in the production of acute panic attacks involve the misinterpret-ation of bodily sensations as a sign of imminent catastrophe such as inter-preting palpitations as a sign of an impending heart attack, or unreality asa sign that one is about to lose control of one's behaviour and behave in acrazy or uncontrolled way. The hypothesis proposes not only that patientsmake such misinterpretations, but also that the catastrophic attribution isstrongly believed. The person who believes that they are at considerableand imminent risk of experiencing a socially or physically threatening eventof catastrophic proportions would, quite logically, attempt to prevent, avoidor escape the perceived catastrophe. As will be described later, there isnoth ing intrinsically abnorma l abou t such safety seeking behav iour .

Thus , the key issue distinguishing cognitive and behavioural accounts ofavoidant and escape behaviour concerns what is being avoided and what issought; that is, according to the behavioural account, avoidance is of fearedstimuli and the patient seeks relief from the anxiety which has becomeassociated with particular stimuli. On the other hand, the alternative cogni-tive hypothesis proposes that the focus of avoidance concerns feared out-comes or consequences perceived as threatening and the patient is seekingsafety. Th e scope of safety seeking be ha vio ur encom passes full avoidanceof feared situations and behaviours occurring within feared situations.Failure to learn from experience: a problem for both cognitive andbehavioura l accoun ts of anxiety associated with panic attacks?In a detailed discussion of psychological perspectives on panic, Seligman(1988) posed an im po rtan t ques tion , wh ich he described as a central weak-


6/14

Behaviour and cognitions in anxiety 11ness in both the Cognitive and Pavlovian theories of the anxiety disorders:neither theory clearly distinguishes the rational from the irrational, theconscious from the unconscious 2 . Seligman argues the need for tw odistinguishable processes, obeying different laws . Particularly relevant toanxiety disorders, he prop ose s, is pr ep ar ed learning (which is biologicallyrelevant, irrational and not readily modified by cognitive means). By impli-cation, prepared learning is also relatively unconscious. He proposes as anexample the case of the patient who has experienced regular panic attacksfor a decade or more. This person

may have had about 1000 panic attacks. In each one, on the cognitive account,he misinterpreted his racing heart as meaning that he was about to have a heartattack, and this was disconfirmed. Under the laws of disconfirmation that Iknow, he received ample evidence that his belief was false, and he should havegiven it up. On the Pavlovian account, he has had 1000 extinction trials in whichthe CS was not followed by the US-UR . . . His panics should have extinguishedlong ago . . . but neither theo ry explains w hy the belief did not extinguish inthe face of disconfirmation long ago. What is it about cognitive therapeuticprocedures which makes them effective disconfirmations, and about the Pavlov-ian exposure procedures that make them effective extinction procedures? (Selig-man, 1988, p. 326).

Seligman thus highlights the apparent failure of people experiencing fre-quent panic attacks to take advantage of naturally occurring disconfir-mations (extinction experiences). He suggests that his well-known conceptof preparedness (Seliman, 1971) can account for this; he argues that preparedlearning follows different rules to unprepared associations. Panic and mostphobias, he suggests, involve highly prepared associations which are particu -larly resistant to extinction.In order to account for the failure of anxious patients to take advantageof naturally occurring disconfirmations, the cognitive hypothesis postulates

a functional and internally logical link betw een co gnitio n and beh avio ur(Clark, 1988; Salkovskis, 1988, 1989c). Assuming for the moment that thesame rules of logic apply to panic as to other areas of human behaviour,then the logical response to threat is to take action designed to preventperceived negative outcom es tha t are believed to be im m inen t. Th at is, aperson panicking because he believes that a catastrophe is imminent will doanything he believes he can to prevent the catastrophe. The person afraidof fainting sits, the per son afraid of hav ing a hea rt attack refrains fromexercising, and so on. By doing so, the patient not only experiences immedi-ate relief, but also unw ittingly pr ot ec ts his or her belief of the potentialfor disaster associated with particular sensations. Each panic attack, ratherthan being experienced as a disco nfirm ation, becom es ano ther ex ample of


7/14

12 P.M. Salkovskisnearly being overtaken by disaster; I have been close to fainting so manytimes: I have to be careful, or one of these times I won't be able to catchit. Th is means that the apparen t failure of panic patien ts to take advantageof natural disconfirmations may be because the non-occurrence of fearedcatastrophes, when associated with safety seeking behaviour, does not con-stitute an actual disconfirmation, and may sometimes be perceived as con-firmation of a near miss .

Thus , the avoidant behaviour of panic patients is normal and logical, inthe sense that the fact that the reader of this article avoids drinking poisonis normal and logical. It could be argued that the reader has good reasonfor believing that drinking poison is harmful, despite the lack of directexperience of the effects of poison; but then, according to the cognitivehypothesis, the panic patient also has good reason for his or her catastrophicmisinterpretations. This view also helps explain the reluctance of the panicpatient to carry out exposure: for the panic patient to engage in unrestrictedexposure, convincing evidence contradicting the feared catastrophe isrequired. The reader would require some specific encouragement to drinkfrom a bottle marked D A N G E R : DE AD LY P O I S O N , and might reasonablybe apprehensive when doing so, even when convinced by, for example,another person modelling drinking from the bottle. Some of the beliefs heldby anxiety patients are specific and idiosyncratic: for example, patients mayoverestimate their personal anxiety sensitivity, such that they believe thateven a small amount of anxiety could result in heart failure for themselves.Although it may be difficult for an observer to understand the fear ofan agoraphobic patient concerning entering the supermarket, this simplyindicates that the observer does not share the idiosyncratic beliefs of thepatient. Thus, a young child is unlikely to understand the horror his parentexperiences on seeing him unscrewing the lid of a bottle of paraquat. It iseasier to understand the avoidance of agoraphobic patients if one simplyreflects upon why he or she will not enter a situation where he or shebelieves that death, loss of control or insanity may occur as a result.

This account of the cognitive basis of avoidant and escape behaviourhas much in common with Rachman's (1984) safety signal perspective ofago raph obic avoidan ce, in that it specifically extends the scop e of such safetysignals to behav iours , wh ich could best be described as safety seekingbehaviours . A major problem which must be faced, however, is that theproposed link between specific avoidance behaviours and cognition dependson a m ore mentalistic analysis of the in ten tio n of the behaviour. That is,the cognitive definition of avoidant behaviour (as opposed to adaptive co p in g beha viour whic h has the effect of reducin g anxiety in the longerterm) depends on an understanding of what is being avoided and what the


8/14

Behaviour and cognitions in anxiety 13intention of avoidance is apparen tly representing a major dep arture fromprevious behavioural conceptualizations of avoidance, although perhapsconsistent with other recent developments in learning theory (e.g. Mackin-tosh, 1983; Power, 1991). However, before further conclusions can bedrawn with any confidence, the relevance of the proposed behaviour-cog-nitions link needs to be examined m ore directly and the distinctions betweenavoidance and coping behaviour defined particularly carefully. Research onthese topics is currently under way.

In summary, from a cognitive perspective, it is argued that panic is infact a rational response to a given set of circumstances, and that there isfundamentally no difference between the type of anxiety and rea so nin ginvolved in panic as opposed to other types of anxiety. The evidential basisfor the misinterpretations made in panic may be inaccurate but as hasalready been described, the important point is that the patient experiencingpanic has a logical basis for these misinterpretations. The rationality orirrationality of a pe rso n's beliefs to the outside observer is not the keyissue. The degree of anxiety is proportional to the immediate personal andidiosyncratic appraisal of threat; in the case of panic attacks, the catastrophicnature of the misinterpretations generates spectacular levels of anxietyincomprehensible to the observer who does not share the patient's assump-tions (and bodily sensations). An equivalent would be the victim of apractical joke who is held up by a masked figure carrying an inoperativereplica weapon; the informed observer who is aware that the weapon isharmless might find the extreme fear and panic of the victim irrational. Inthe next section, the role of such behaviour in a range of anxiety problemswill be briefly considered.Clinical manifestations of safety seeking behaviourThe cognitive account described above deals not only with generalizedavoidance and escape, but also explains some of the specific types ofbehaviour observed to be associated with particular anxiety problems. Forexample, it accounts for medical consultation and reassurance seeking inpatients anxious about their health (Salkovskis, 1989a; Salkovskis and War-wick, 1986; Warwick and Salkovskis, 1990); the link between suchbehaviour and the fear of illness is obvious. Less obvious is the relationshipbetween cognition and behaviour in people suffering from obsessional prob-lems. However, the cognitive-behavioural analysis of obsessions (Salkov-skis, 1985, 1989b) suggests that the cognitions involved concern theinterpretation of the occurrence or content of intrusive thoughts as a signof personal responsibility for further action. It therefore follows that thebehaviours of such patients should reflect attempts to prevent themselves


9/14

14 P. M. Salkovskisfrom being responsible for adverse consequences that might arise from notacting on the content of the thought (prevention of harm through, forexample, washing and checking). Often, this type of behaviour may manifestas covert neu tral izin g (Salkovskis and W estbro ok, 1989). An otherphenomenon of interest is concern over the actual occurrence of upsettingthoughts, leading to counter-productive attempts at active thought sup-pression (Lavy and van den Hout, 1990; Salkovskis, 1989b; Wegner, 1989).In social anxiety, self focus (i.e., away from the social context) may also beinvolved in the maintenance of clinical anxiety.

Recent work carried out by our own group suggests that similar cog-nition/behavioural interactions may even apply to simple phobics (usuallyregarded as the anxiety problem with the least cognitive involvement). Forexam ple, spider phob ics repo rt high belief ratings on items such as T hespider will attack m e and If I can 't escape from the spider I will goinsane . Consideration of the behaviour of spider phobics (c.f. Watts andSharrock, 1984) suggest that the behaviour of such people is often notsimply directed at preventing contact with feared stimuli, but also at pre-venting disasters if such contact takes place. Research into this type ofassociation may yet prove therapeutically and theoretically useful (Rach-man, 1990b).Clinical implicationsSome of the most important implications of this proposed cognition-behaviour link concern treatment. In general, cognitive behavioural treat-ment emphasizes the need to deal with idiosyncratic factors that are bolster-ing the continued misinterpretation of bodily sensations. For example, apatient w ho becomes confused d uring panic describes mentally holdin gon to my sanity . In each successive panic he becomes more convinced thathe would have gone mad were it not for this effort. Another example isthe patient who, hundreds of very severe panic attacks later, still believesin each new attack that she is about to go crazy, pass out or die. Clinically,questioning of such patients reveals that no disconfirmation has occurredbecause the patients believe that they have, in every instance, been able totake successful preventative action. As described above, each panic becomesa nea r m iss and further confirma tion of the risk. On ce such behaviouralresponses are identified, the patient can be helped to begin the process ofre-appraisal by withholding such protective responses and learning the trueextent of risk. In many instances, this may involve suggesting to the patientthat they challenge their worries by actively trying to bring about the feareddisaster; for example, going into the supermarket and trying to faint ortrying to go mad. This helps the patient to discover that their efforts to


10/14

Behaviour and cognitions in anxiety 15prevent these disasters have been misdirected; this can also help the patientto then re-interpret their usually considerable past experiences of suchanxiety prov okin g situations as true disconfirmations instead of nea rm isses . This partic ular ty pe of association m ight be a fruitful one forresearch investigations, as controlling one's mind is a common response inpanic patients as well as other groups, including non-clinical subjects. Giventhe readiness with which patients can reproduce this type of behaviour aseither safety seeking or coping when requested in the laboratory, experi-ments concerning the use of the same degree of effort directed at differenttargets (i.e. reduction of anxiety vs preservation of sanity) could be used toassess the validity of the cognitive basis of such behaviour and its putativeanxiety-preserving effects when used as a safety seeking behaviour.The present analysis also suggests ways of combining cognitive pro-cedures and brief exposu re in a wa y which sho uld be particular ly effectivein bringing about belief change without repeated andprolonged exposurebeing necessary. Thus, exposure sessions are devised in the manner ofbehavioural experiments, intended as an information gathering exercisedirected towards invalidation of threat-related interpretations. Most pre-vious studies in which cognitive and behavioural treatments have beencombined (such as those reviewed in Marks, 1987a, b) have used cognitiveprocedures as a way of dealing with general and ba ck gr ou nd life stresses,most of which tend not to be directly relevant to the specific experience ofanxiety subject to exposure. The particular strategy of using exposure as anexercise in testing alternative non-threatening interpretations of experiencewou ld be predicted to succeed better than brief exposure with sup plem en-tar y (threat-irrelevant) cognitive change proc edu res. That is, such a studyshould show that gen era l cognitive therap y com bined with exposure hasan additive effect, whilst anxiety focused cognitive therap y wo uld beexpected to multiply the effect of exposure, resulting in maximal cognitivechange through behavioural experiments. Thus, according to the cognitivehypothesis, the value of behavioural experiments transcends mere exposure;such experiments allow patient and therapist to collaborate in the gatheringof new information assessing the validity of non-threatening explanation ofanxiety and associated symptoms.ConclusionThe difference between the account outlined in this article and previousbehavioural explanations is that, in anxiety disorder, it is not invariably thefeared stimuli that are being avoided, but may more commonly be threaten-ing consequences of particular situations or of anxiety itself. In mostinstances, if a per son is convinced that the e nd -po int of exposu re to a feared


11/14

16 P.M. Salkovskisstimulus will be anxiety alone, then avoidance is relatively unlikely, otherthan the desultory avoidance normally associated with mildly unpleasantemotional states. If, on the other hand, the person is anxious as a result ofa perceived threatening or catastrophic event, then avoidance of that threatis both rational and advisable. An important extension of this view concernsthe issue of anxiety sensitivity i.e. the tendency to believe that anxiety perse is dangerous, which has been demonstrated to be a feature of panic,especially panic associated with agoraphobia (McNally and Lorenz, 1987;Reiss, Peterson, Gursky and McNally 1986). Anxiety will be the focus ofavoidance if the subject strongly believes that anxiety can itself result inharm. That is, the person who believes that anxiety itself can result inserious physical or social harm would reasonably avoid situations where heor she might become anxious. This phenomenon would be expected to bea major mechanism involved in the readiness to acquire phobic behaviour;for example, the transition from Panic Disorder to situational (agoraphobic)avoidance.

Safety seeking behav iou rs as described here are hypo thesized as havingthe subjective effect of sa vi ng the per son from the threa t involved inanxious stimuli and situations, in the sense that the person comes to believethat their behaviour stands (and has stood) between them and a likelydanger. It is not proposed that learning theory approaches to avoidancebehaviour be abandoned, but rather that it be the incorporation of cognitiveaccounts of how learning takes place and what is learned (Rescorla, 1988;see also Pow er, 1991; van den H ou t and M erkelbach, 1991).

Th e cognitive hypoth esis of avoidance may help account for the ex traordi-nary efficacy of intensive graded exposure (e.g. Ost, 1989; Ost, Salkovskisand Hellstrom, 1990), and provides a framework to understand the keyissue of the difference between a coping (adaptive) response and an avoid-ance (anxiety maintaining) response. The key issue concerns the questionof what the person is avoiding. If the cognitive accou nt is correc t, thenavoidance responses are those behaviours which are intended to avoid disas-ter but thereby also have the secondary effect of preventing the disconfir-mation that would otherwise take place. On the other hand, copingresponses are those behaviou rs br oug ht to bear by a person intending todeal with anxiety alone with no further fears about the consequences ofthe anxiety and so on. The second strategy is not intended to prevent threat,and therefore will not interfere with disconfirmation; in fact, it would beexpected to enhance cognitive change because the strategy is based on analternative, non-threatening account of symptoms and situations whichreceives logical support from the patients' experience.


12/14

Behaviour andcognitions inanxiety 17Notes

1. The theoretical inconsistency here may have been masked by the way thatbehavioural theories gradually shifted to pur e exp osure views away from theoriesbased on reciprocal inhibition (Wolpe, 1958), which hadoriginally required a com-peting anxiety-inhibiting response.

2. At least someofthese issues conc erning therelative inaccessibility of cognitiveprocesses, discussed in detail by Williams etal. (1988) can be resolved by makingthe important distinction between the m easurement of cognitive processesand th ecognitive events which can beregardedas theoutcome of those processes. It shouldbe possible to measure reliably b oth aspects of cognitive functioning .

ReferencesBECK, A. T., EMERY, G. and GREENBERG, R. L. (1985). Anxiety Disorders and

Phobias.New Yo rk: Basic Book s.BLACKBURN, I. M. (1986). The cognitive revolution: an ongoing evolution.

Behavioural Psychotherapy 14 274-277.BORKOVEC, T . (1982). Facilitation and inhibition of functional CS exposure in the

treatment of phobias . In J. Boulougouris (Ed). Learning Theory A pproachestoPsychiatry. New York: Wiley.BUTLER, G., C U L L I N G T O N , A., H I B B E R T , G., K L I M E S , I., and G E L D E R , M. (1987).Anxiety management for persistent generalized anxiety. British Journal ofPsy-chiatry 151, 535-542.

CLARK, D. M. (1986). Cogn itive therap y for anxiety. Behavioural Psychotherapy14, 283-294.

CLARK, D. M. (1988). A cognitive model of panic. In S. Rachman and J. Maser(Eds.).Panic: Psychological Perspectives. Hillsdale: Erlbaum.

H A W T O N , K., SALKOVSKIS, P., K I R K , J. and C L A R K , D. M. (1989). CognitiveBehaviour Therapyfor Psychiatric Problems:A Practical Guide. Oxford: OxfordUniversity Press.

LAVY, E. H. and VAN DENHOUT, M. A. (1990). Th ou gh t sup pression inducesintrusion. Behavioural Psychotherapy 18 4).

MCNALLY, R. J. (1986). Preparedness andphobias: a review. Psychological Bulletin101, 283-303.

MCNALLY, R. J. and LORENZ, M. (1987). Anxiety sensitivity in agoraphobics. Jour-nalofBehavior Therapy andExperimental Psychiatry 18, 3-11.

MACKINTOSH, N . J. (1983) Conditioning andAssociative Learning. New York:Oxford University Press.

MARKS, I. M. (1987a). Behav ioral aspects of panic disorder. American JournalofPsychiatry 144 1160-1165.MARKS, I. M. (1987b). Fears Phobias andRituals. NewYork: O xford University

Press.


13/14

18 P.M. SalkovskisMATHEWS, A. M., GELDER, M. G. and JOHNSTON, D. W. (1981). Agoraphobia:

Nature andTreatment. New York : Guilford Press.MATTICK, R. P. and PETERS, L. (1988). Treatm ent of severe social phob ia : effects

of guided exposure with andwith out cognitive restructuring.Journal of Consult-ingand Clinical Psychology 56, 251-260.

MATTICK, R. P., PETERS, L. and CLARKE, J. C. (1989). Exposure and cognitiverestructuring for social pho bia: a controlled study. Behavior Therapy 20, 3-24.

OST, L. G. (1989).One session treatment for specific phobias. Behaviour Researchand Therapy 27,1-8.

OST, L. G., SALKOVSKIS, P. M. and HELLSTROM, K.(1990). Onesession therapist-directed exposure vsself-exposurein thetreatment ofspider phobia. Manuscriptsubmitted for publication.

PAUL, G. L.(1966). Insight versus Desensitisation inPsychotherapy. Stanford CA:Stanford University Press.POWER, M. J. (1991). Cognitive science and behavioural psych otherapy : where

behaviour was, there shall cognition be? Behavioural Psychotherapy 19,20-41.RACHMAN, S. (1977). The conditioning theo ry of fear acquisition: a critical exami-

nation. Behaviour ResearchandTherapy 15, 375-387.RACHMAN, S. J. (1981). The primacy of affect: some theoretical implications.

Behaviour ResearchandTherapy 19, 279-290.RACHMAN, S.J.(1984) Ago raphob ia: a safety signal perspective, Behaviour Research

and Therapy 22,59-70.RACHMAN, S.J. (1990a). FearandCourage (2nd edn) New York: Freeman.RACHMAN, S. J. (1990b). The determinants and treatment of simple phobias.

Advances inBehaviour ResearchandTherapy 12, 1-30.RACHMAN, S. J., CRASKE, M., TALLMAN, K. and SOLYOM, C. (1986). Does escape

behaviour strengthen agoraphobic avoidance.A replication. Behavior Therapy 17,366-384.

RACHMAN, S. J. and HODGSON, R.(1974). Sy nchro ny and desynchrony infearandavoidance. Behaviour Research andTherapy 12, 311-318.

REISS, S., P E T E R S O N , R. A., GURS KY, G. M., and M C N A L L Y , R. J. (1986). Anxietysensitivity, anxiety frequency and the prediction of fearfulness. BehaviourResearchandTherapy 24, 1-8.

RESCORLA, R. A. (1988). Pavlovian c ond itioni ng: it s not what you think it is.American Psychologist 43, 151-160.

SALKOVSKIS,P. M.(1985). Obsessive-compulsive proble ms: a cognitive-behaviouralanalysis. Behaviour Research andTherapy 23, 571-583.

SALKOVSKIS, P. M. (1986). The cognitive revolution: new wayforward, backw ardsomersault or full circle? Behavioural Psychotherapy 14,278-282.

SALKOVSKIS, P. M. (1988). Phenomenology, assessment and the cognitive modelofpanic.In S.RACHMAN and J. MASER (Eds).Panic: P sychological Perspectives. NewJersey: Erlbaum.

SALKOVSKIS, P. M. (1989a). Som atic disord ers. In K. Hawton , P. M. Salkovskis,J.


14/14

Behaviour andcognitions inanxiety 19W. Kirk and D. M. Clark (Eds). Cognitive-behavioural Approaches to AdultPsychiatric Disorder: A Practical Guide. Oxford: Oxford University Press.

SALKOVSKIS, P. M. (1989b). Cognitive-behavioural factors and the persistence ofintrusive thoughts in obsessional p roblem s. Behaviour Researchand Therapy 27,677-682.

SALKOVSKIS,P. M. (1989c). Cognitive models andinterventions inanxiety. CurrentOpinion inPsychiatry 2,795-800.

SALKOVSKIS,P. M. AND WARWICK, H. M. C. (1986). Morbid preoccupations, healthanxiety and reassurance: a cognitive behavioural appro ach to hypochondriasis .Behaviour Researchand Therapy 24, 597-602.

SALKOVSKIS,P. M. andWESTBROOK, D.(1989). Beha viour the rapy with obsessionalruminations: can failure be turne d into success? Behaviour Researchand Therapy27 147-160.SANDERSON, W. C , RAPEE, R. M. and BARLOW, D. H. (1989). The influence of anillusion ofcontrol on panic attacks induced viainhalation of 5.5% carbon dioxideenriched air. Archives of General Psychiatry 46, 157-164.

SELIGMAN, M. E. P. (1971). Phobias and preparedness . Behaviour Therapy 2,307-320.

SELIGMAN,M. E. P. (1988). Competing theories ofpanic. In S.Rachman and J. D.Maser (Eds). Panic: Psychological Perspectives. New Jersey: Erlbaum.

SILVA,P. andRACHMAN, S. J. (1984). Does escape behaviour strengthen agora-phobic avoidance? A preliminary study. Behaviour Research and Therapy 22,87-91.VAN DEN HOUT, M. and MERCKLEBACH, H. (1991). Classical conditioning: stillgoing strong. Behavioural Psychotherapy 19, 59-79.

WEGNER, D. M.(1989). White Bearsand Other U nwanted Thoughts. New York:Viking.

WARWICK, H. M. C. and SALKOVSKIS, P. M. (1990). Hypochondriasis. BehaviourResearchand Therapy 28 105-118.

WATTS,F. N. andSHARROCK,R. (1984). Questionn aire dimensions ofspider pho bia.Behaviour Researchand Therapy 22,575-580.

W I L L I A M S , J. M. G., W A T T S , F. N., M A C L E O D , C. and M A T H E W S , A. M. (1988).Cognitive PsychologyandEmo tional Disorders. Chichester: Wiley.

WOLPE, J. (1958). Psychotherapy by Reciprocal Inhibition. Stanford, CA: StanfordUniversity Press.

behavioural psychotherapy

Documents

Transcript of behavioural psychotherapy