J. Behnv. Thu. & Exp. Psychiar. Vol. 24, No. 4, pp. 279-288. 1993.

Elsevier Science Ltd

Printed in Great Britain
COO-7916/93 $6.00 + 0.00

00057916(93)E0006-S

A CRITIQUE OF COGNITIVE THERAPY FOR ANXIETY DISORDERS*

S. RACHMAN
Department of Psychology, The University of British Columbia

Summary - Progress in the cognitive-behavioral treatment of anxiety disorders is reviewed.

Significant advances have been made in treating panic disorders and there are promising signs of an
expansion of cognitive theory and therapy to other disorders, notably hypochondtiasis, obsessional
disorders and circumscribed phobias. Nevertheless, some difficult obstacles have emerged to
present serious problems for the prevailing cognitive theory.

The emergence of cognitive methods of therapy is out the ideas of CBT and the evidence that has
a progressive and promising development, and been recruited to support them.
despite numerous obstacles, already has informed
and refreshed our ideas about anxiety disorders.
The purpose of this critique is to evaluate the Panic: Cognitive Explanations and Deductions
progress of cognitive intervention for anxiety
disorders, and to do so in a constructive manner. The leading proponents of a cognitive approach
The infusion of cognitive concepts and to the treatment of panic are Barlow (1988) who
procedures into clinical psychology has generated emphasizes the internal as well as the external
a revival of interest in clinical work and some cues for panic, Beck (1988) and Clark (1986;
remarkable therapeutic claims have been asserted, 1988). According to Clark, “panic attacks result
including excellent results with panic disorders from the catastrophic misinterpretation of certain
(Margraf, Barlow, Clark, & Telch, 1993). As far bodily sensations” (1986, p. 462). It follows that
as anxiety disorders are concerned, the greatest (1) a reduction in catastrophic cognitions should
theoretical and clinical progress has been made in reduce the episodes of panic, and (1 a) that the
applying cognitive theories and methods to the elimination of the catastrophic cognitions should
explanation and treatment of panic, and lead to the cessation of panic. (2) The substitution
consequently, the greatest attention will be of noncatastrophic interpretations of these bodily
devoted to these developments. Some progress has sensations should reduce episodes of panic. There
also been made in tackling other anxiety disorders is another, less obvious deduction, namely that (3)
in this way, notably specific phobias: claustro- a reduction in the relevant bodily sensations
phobia; obsessional compulsive disorders; and should reduce the opportunities for catastrophic
hypochondriasis (approached as a disorder of misinterpretations and hence, should be followed
anxiety). by a decline in the episodes of panic. (4)
These developments have also drawn critical Reduction/elimination of the catastrophic
fire, and I will turn to the critics after first setting cognitions should be followed by a generalized

*Invited address, APA Conference, Division 12, Toronto, 1993.

Requests for reprints should be addressed to S. Rachman, Department of Psychology, The University of British Columbia,
2136 West Mall, Vancouver, B.C., Canada V6T 124.

219
280 S. RACHMAN

Table 1

Srlrctrd Deductions from Cognitive Theon

I. Reduction in catastrophic cognitions should reduce episodes of panic.

la. Elimination of catastrophic cognitions should lead to cessation of panic episodes.
2. Substitution of noncatastrophic interpretations of bodily sensations should reduce episodes of panic.
3 Reduction in relevant bodily sensations reduces opportunities for misinterpretation and hence reduces episodes of panic.
4: Reduction/elimination of the catastrophic cognitions should be followed by a generalized reduction in fear and panics.
5. The enduring reduction of these cognitions should be accompanied by a lasting reduction in the occurrence/intensity of the
episodes of panic.
sa. A return of these cognitions, or a restoration of their high believability. will be followed by a return of fear and a recurrence
of episodes of panic.

reduction in fear and panics. (5) The enduring cognitions, if any, of the panic-free
reduction of these cognitions should be patients.
accompanied by a lasting reduction in the (3) There is, so far, no quantitative evidence
occurrence/intensity of episodes of panic. (5a) A that pertains directly to the third deduction,
return of these cognitions or a restoration of their that is, the extent to which successfully
high believability, will be followed by a return of treated patients make fresh but benign
fear and a recurrence of episodes of panic. interpretations of their bodily sensations.
Most panic patients endorse or describe more Clinical reports suggest that this change
than one catastrophic cognition, and also endorse does take place.
several bodily “symptoms”; in practice, the (4) After successful treatment of panic
diagnosis of a panic disorder requires the disorder, patients reported fewer bodily
endorsement of at least three such bodily sensations than they did prior to treatment
sensations. Leaving aside for the moment the (Clark et al. 1992; Margraf-Schneider,
complexities of deciding which of the several 1991).
cognitions are critical and which of the bodily The evidence pertaining to the last three
sensations are critical, we can turn to the evidence deductions is positive but limited in quantity and
pertaining to these deductions. derived mainly from the clinical experiment on
(1) There is evidence that after successful claustrophobics referred to below (Booth &
treatment of panic disorder, patients do Rachman, 1992). There is also a large and
report significantly fewer negative growing body of support for elements of the
cognitions than they did prior to treatment underlying theory (Clark, 1993; Ehlers, 1993;
(e.g., Clark, Salkovskis, Hackmann, & Rapee, 1993).
Celder, 1992; Margraf & Schneider, 199 I ). There are indications that the improvements in
(2) There is also some indirect evidence to reducing panic are stable, and that there is a
support the second deduction. An positive correlation between reduction of the
encouragingly high percentage of patients negative cognitions during treatment and the
are free of panics at the conclusion of maintenance of improvements (see Clark et al.
treatment. In the most recent controlled 1992; Craske & Rodriguez, 1993; Margraf &
studies, 90, 86, 71, 86 and 75% of patients Schneider, 1991). Similarly, Shafran et al. ( 1993)
who received cognitive therapy were panic- found a correlation of 0.47 (p = 0.05) between
free at the end of treatment. However, we cognitions at post-test and fear at follow-up in a
do not have specific information about the group of treated claustrophobic subjects. It is
nature and number of the remaining however, possible that cognitions predict the
 A Critique of Cognitive Therapy for Anxiety Disorders 281

stability of change even though they are not the Alternative Interpretations: Cause, Consequence
vehicle of change; also, Margraf found other or Correlate?
predictors of stability, for example, in the
patient-therapist relationship. All of these results The decline in cognitions, and in bodily
are consistent with the cognitive approach to sensations, observed after successful treatment is,
therapy. of course, open to more than a single
It is worth noticing that even after successful interpretation. The decline in cognitions, and/or in
therapy, a number of patients continue to report bodily sensations may produce the reduction of
some negative cognitions (e.g., Margraf & the panics. But it is possible that the decline in
Schneider, 1991); presumably these are non- cognitions, and in bodily sensations, are
critical, or if they are critical, they are no longer consequences of the reduced episodes of panic,
strongly believed. The patients also continue to and not the cause. It is also possible that the
report bodily sensations, and these remaining decline of cognitions is a correlate of the reduction
sensations too, must be non-critical, or now in the episodes of panic [some critics have
interpreted in a benign manner. From this, one can suggested that the cognitions and their decline
deduce that some cognitions are critical (the so- may be mere epiphenomena (e.g., Seligman, 1988;
called key cognitions; see later) and some bodily Wolpe & Rowan, 1989)]. Wolpe and Rowan
sensations too may be critical. In testing the major argued that the first episode of panic is an
therapeutic deductions it will be necessary to unconditioned anxiety response, and the panic
identify the critical cognitions and bodily disorder arises from fear conditioned to stimuli
sensations in advance. If the critical cognitions, associated with the initial episode.
identified in this way prior to treatment, persist One reason for giving serious consideration to
even after fully successful treatment has been these alternative explanations arises from the fact
completed, the cognitive theory will be damaged. that in Margraf and Schneider’s (1991) study the

tromt treatemnt tramt

cognitiretherapy

Figure 1. The reduction of claustrophobia:

therapy (Booth & Rachman, 1992).
a comparison
ll.L
POti-

Exponua based
therapy
II
Pm-
tmatmmlt tmatmcnt treatlmat

Iutaroceptire therapy

of three methods. Negative cognitions declined after direct or indirect

282 S. RACHMAN

patients who received pure exposure treatment brief account. Clark and his colleagues have
without cognitive manipulations showed shown that people who experience panic attacks
improvements as large as the patients receiving make a greater number and more intensely
pure cognitive therapy in which exposures were negative interpretations of bodily sensations than
excluded. Moreover, the cognitions declined to the do other people. There are meaningful links
same extent in both groups. between cognitions and bodily sensations
In a similar vein, Booth and Rachman (1992) (Rachman & Levitt, 1988; Marks et al., 1991).
found that the fears and negative cognitions of Varying the instructions given to panic patients
claustrophobics showed large declines, regardless prior to undergoing the physiological lactate
of whether they received exposure treatment or provocation test affects the probability that an
pure cognitive treatment without any exposure episode of panic will occur, panics can be induced
(see Figure 1). It appears that negative cognitions by cognitive or other psychological manipulations
can decline after a direct attack or after an indirect (Rachman, 1990), etc.
attack. The theory has been criticized for being non-
exclusive, unrelated to traditional cognitive
Effects of Indirect Treatment on Cognitions psychology (e.g., Seligman, 1988), incomplete
(Rachman, 1990), unable to account for important
A satisfactory cognitive explanation needs to phenomena (Klein & Klein, 1989), and
account for the declining cognitions that occur indistinguishable from conditioning theory and no
after a nondirect treatment, such as exposure. The advance on that theory (Wolpe & Rowan, 1989).
most obvious possibility is that with each These complex theoretical matters will be sorted
exposure, the patient acquires fresh, out in time, but to return to the therapeutic
disconfirmatory evidence (e.g., no heart attack, did mechanisms of cognitive therapy, we need to
not lose control). The accumulation of this ascertain whether or not the reduction/elimination
personal, direct, disconfirmatory evidence of key cognitions is the critical element in this
weakens the catastrophic cognitions. However, form of therapy. We already know that the direct
one is nevertheless left to ponder why the direct modification of cognitions can be a sufficient
assault on cognitions is not more effective than the condition for treatment success, but we also know
indirect, incidental effects of exposure, as seen in that direct modification is not a necessary
the studies by Margraf and by 6st (199 1). condition for such success (e.g., exposure alone
There is an even more difficult problem. It is can be as effective as cognitive therapy,
desirable to accommodate the therapeutic effects imipramine and other medications produce
of drugs such as imipramine, and the fact that after therapeutic improvements, etc.).
this treatment as well, negative cognitions show a
small but significant decline (Clark et al., 1992). “Noncognitive ” Panics

Conditions for Change In their investigations of the specific links

between bodily sensations and negative cognitions
No satisfactory evaluation of the effects of as they occur in episodes of panic, Rachman,
cognitive therapy can ignore the context from Levitt, and Lopatka (1987) came across a small
which the theory emerges and the evidence number of episodes in which the panic patients
pertaining to the validity of the cognitive reported an absence of negative cognitions. They
explanation for the nature and the causes of panic attributed these occurrences to measurement
disorder (see Clark, 1988; Craske & Rodriguez, problems but expressed some disquiet because all
1993; Rapee, 1993; Rachman, 1990; Rachman & of the patients who had a “noncognitive panic”
Maser, 1988; Seligman, 1988; Teasdale, 1988), had also reported at least one cognitive panic. It
but for present purposes there is space for only a was not merely subject error. It remains an
 A Critique of Cognitive Therapy for Anxiety Disorders
unsolved problem and it is no easy matter to
conceive a disconfirmatory experimental test.
Complexities
These analytical tasks are easier to identify than
to study because there are several obstacles to
progress. As mentioned earlier, one is obliged to
consider the probable existence of key cognitions,
critical cognitions that are responsible for the
panics and whose removal is necessary for
elimination of the panics. During treatment,
patients may express not one, but several,
cognitions and endorse several cognitions from
widely used standard lists, such as that constructed
by Chambless (1988). Also, during the course of
treatment, other and often idiosyncratic cognitions
emerge. It is unlikely that all of these cognitions
are equally critical, and we know that even after
successful treatment some patients continue to
endorse a diminished number of negative
cognitions. Given that treatment is successful,
these remaining cognitions cannot be regarded as
critical. How then can we determine, in advance,
as we will need to do, which cognitions are critical
and which are not?
The key cognitions may serve many functions.
They may serve to drive fear and probably to
stitch together different types of fearful cognitions.
There are various ways in which one can define
and measure key cognitions, none of them simple
or straightforward. And as mentioned earlier, the
detection of a non-cognitive episode is difficult.
There is another layer of complexity because
combinations or clusters of cognitions are more
influential than single cognitions (Rachman, Levitt
& Lopatka, 1987; Marks et al., 1991). The prob-
ability of panic is greater when the person has two
or more threatening cognitions.
The Timing of Events
A serious obstacle to severe tests of the theory
arises from the need for control over the timing of
events. If the reductions in negative cognitions are
no more than correlates of panic reduction, or if
the cognitive changes follow rather than precede
283
the reduction of panic, we need to study the
sequence of events with care. Reductions in fear
are easier to observe and record, but they can
occur slowly, over weeks rather than minutes. In
cases of panic, the measures typically range over
days or weeks (e.g., the number of panics recorded
per week or even per month). So if the patient
records a decrease in panics, say from four per
week to one per week, when exactly did this
decline take place?
Cognitive changes can be even more difficult to
track. It is true that in treatment, major changes
can occur suddenly (e.g., Gst, 1989; Rachman &
Whittal, 1989), and it is therefore possible to
record the change in these instances. In many,
perhaps most, occurrences, clinical or
experimental, the cognitive shifts are slow to
develop, changing over weeks rather than minutes
[e.g., the cognitive therapy group in the Booth-
Rachman (1992) study)]. To make matters worse,
the changes in fear and in fearful cognitions can,
and undoubtedly often do, occur even when the
affected person is separated from and out of
contact with the fear-provoking stimulus
(Rachman, 1990). It is not possible to determine
precisely when the change occurred, assuming, of
course, that there is a complete change in the first
place. So we are left with the awkward task of
timing the sequence of changes in the cognitions
and in the episodes of panic, knowing that these
changes may take place over an extended period
and that the determination of a precise point of
change will in many instances be impossible. We
also have some evidence that cognitive shifts can
initiate a process of change that becomes evident
some time later. The existence of a so-called “dual
belief system”, in which the intensity and
believability of fearful cognitions varies greatly
from situation to situation (see Booth & Rachman,
1992) is one more complication. It confuses
measurement and can mislead patient and therapist
alike.
In the midst of these complexities and
obstacles, it is worth drawing attention to the fact
that useful progress has been made in developing
methods to tackle these critical questions. For
example, Salkovskis, Clark and Hackmann (199 1)
284 S. RACHMAN

were able to show in a preliminary study that confidently announce that the fear has gone. At
cognitive treatment that is focused on the (key?) some point between sessions, the processes set off
cognitions produces larger and quicker changes by cognitive inputs reached a conclusion; in this
than therapy which is more broadly aimed. In sense, the sessions initiate emotional processing
addition, we have early indications from the (Rachman, 1980, 1990) that reaches completion
claustrophobia experiment that the believability of only after an interval in which the fearful person
the cognitions plays a major part in therapy. A has no contact with the phobic stimulus and
single strongly believed cognition may be more usually cannot recall having made deliberate
important than four or five cognitions with low attempts to facilitate the fear-reduction between
believability (Shafran, Booth & Rachman, 1992). sessions. This apparent delay in the effects of the
This same experiment produced evidence of the inputs was also encountered in the experimental
centrality of certain cognitions; when the reduction of claustrophobia described by Booth
presumably key cognition of “feeling trapped” and Rachman (1992). The comparison group
was reduced, most of the other cognitions received exposure-only treatment and made rapid
collapsed and fear declined sharply. In 10 out of progress; the cognitive group made slow early
13 claustrophobic Ss, an absence of believable progress but caught up with the gains of the
cognitions at post-test was associated with a total exposure-only group directly after the re-test in
loss of claustrophobia. And zero fear was never which the Ss were required to re-enter the closet,
reported in the presence of believable negative i.e., to be re-exposed (see Figure 1). In seeking to
cognitions. Moreover, high belief in the cognition explain the therapeutic action of cognitions it is
“trapped’ correlated with the return of fear. essential to look beyond the events that are known
to occur during particular sessions. True, the
Phobias sessions often are followed by an immediate
reduction in fear, but in other instances a session
As recently as 1987, Last concluded her review initiates a process that continues well past the
of cognitive methods for phobias with pessimism. duration of the session and the effects are not
There was little evidence to support their efficacy, apparent until some period has passed.
but within the last two years some promising signs A second feature of these rapid “glass-jar”
have appeared. Slight progress has been made in changes in fear carries a caution. The fearful
reducing snake/spider fears by them (Rachman & person’s confident report that the fear has been
Whittal, 1989) and it was discovered that a sudden eliminated is not always followed by an equally
and complete elimination of fear occurs in some rapid reduction in avoidance behavior.
cases, reminding one of “learning by insight”. In Desynchronous changes sometimes take place,
these instances the fearful person experiences a thereby complicating the task of explaining the
sudden transition and is certain the fear has been mechanisms of change. Pending further progress,
eliminated - e.g., “I am no longer frightened of the simplest course to follow at present is to
snakes; it has gone.” concentrate first on finding an explanation for the
Two interesting features of these so-called cognitively produced changes in self-reported fear.
“glass jar” experiences, of a sudden, complete In the comparative experimental analysis of the
elimination of fear, need to be taken into account reduction of claustrophobia referred to above, a
in trying to identify the mechanisms of change. pure exposure condition proved as powerful in
First, the elimination of fear is sometimes reducing claustrophobia as cognitive therapy
recognized and reported by the person to have (Booth & Rachman, 1992). Detailed analyses of
taken place between sessions. He/she might the cognitive changes did however provide
complete a session with say a 45% fear reaction to support for the cognitive theory as elaborated from
the phobic stimulus remaining. On returning for Clark’s theory (1986, 1988). High fear (and panic)
the next session, perhaps a week later, they was always accompanied by negative, fearful
 A Critique of Cognitive Therapy for Anxiety Disorders 285

cognitions; moreover, zero fear was never reported responsibility (see Rachman, 1993). We now have
in the presence of believable cognitions (Shafran, experimental evidence to support this view, at
Booth, & Rachman, 1992). Changes in the number least in respect of obsessions and of compulsive
and the believability of the cognitions were closely checking.
associated with reductions in fear. Continued There is a need for additional trials, but the
belief in a central fearful cognition was, on the early results are promising. Arntz (1992),
other hand, associated with maintenance of the Emmelkamp and Beens (1991), van Oppen (1992)
fear. It was concluded that claustrophobia and Salkovskis (1992) have all reported successful
comprises a number of cognitions “centred on key cognitive therapy, but in none of the comparisons
thoughts of trappedness, suffocation and loss of with conventional behavior therapy was it shown
control” (Shafran, Booth, & Rachman, 1992, p. to be superior. As in the therapeutic research on
75). The removal of the belief in the pertinent panic disorder, the dysfunctional cognitions of
cognition was followed by “a dramatic reduction obsessional patients treated with “pure exposure”
in claustrophobia” (p. 75). have also been reported to decline.
Encouraging though these results are, one must
remember that the noncognitive comparison Hypochondriasis
treatment, exposure-only, produced equally large
reductions in claustrophobia. If the reduction of There are early signs that (noncognitive)
fearful cognitions is indeed the mechanism behavioral treatment of this disorder may tell the
involved in successful treatment, it has to be same story as that emerging from research on
conceded that the cognitions can be removed as panic. Prompted by the novel re-analysis carried
effectively by indirect noncognitive methods. That out by Warwick and Salkovskis (Salkovskis &
concession is comforting but in turn raises fresh Warwick, 1986; Warwick & Salkovskis, 1987;
problems, not least of which is skepticism about 1990), researchers in Holland compared cognitive
the need for cognitive therapy. and pure behavioral treatment in a series of case-
studies, only to find a slight superiority for the
Obsessional Disorders behavioral method (Visser & Bouman, 1992). As
in related research, dysfunctional beliefs declined
In Salkovskis’ (1985) refreshing cognitive after behavioral treatment, equally well.
analysis of obsessional compulsive disorders Given their common ancestry (Salkovskis &
(OCD), the affected person’s construal of the Clark, 1993), it is no surprise that the cognitive
behavior and urges was taken as the starting point analyses of hypochondriasis (Hs) and of panic
and the conclusion of the problem. Salkovskis disorder are similar and that the results of
focussed attention on the explanations which the cognitive therapy of panic disorder find an echo in
affected person provides for his/her OC urges, the work of Visser and Bouman. The cognitive
behavior and motives. In this way he succeeded in theory of Hs shares the boldness, simplicity and
filling a previously empty stage. Previously, the breadth that characterizes the theory of panic
nature and significance of the specific content of disorder. It is argued that “bodily signs and
the obsessions and compulsions remained symptoms are perceived as more dangerous than
unexamined. they really are, and that a particular illness is
As a result of this analysis therapists now believed to be more probable than it really is,”
devote serious attention to the cognitive contents (Warwick & Salkovskis, 1990, p. 110). In panic
of the OCD and attempt to modify the maladaptive disorder, the affected person is assumed to make a
ideas. Salkovskis also emphasized the importance catastrophic misinterpretation of bodily sensations
of the person’s sense of responsibility, claiming and hence panic (Clark, 1986). Importantly, the
that in most cases the OCD is associated with panic theory pertains to expectations of imminent
an excessive and absurdly broad sense of catastrophe, e.g., “I am having a heart attack”. The
286 S. RACHMAN

Hs theory pertains to threats to one’s health or interpretation of the events, improvement will
well-being that can be equally catastrophic but are follow. The future of CBT in the treatment of
more remote (e.g., “This bump on my skin may panic disorder will have major ramifications for
develop into a cancer.“), but the underlying the cognitive approaches to all of the anxiety
mechanisms are assumed to be common to both disorders.
disorders. It follows therefore that the action of
psychological treatment for Hs, as for panic
disorder, should be directed at the identification of Conclusions
the dysfunctional misinterpretations of bodily
sensations and signs, their modification and To conclude, the cognitive explanation for the
replacement with a more benign (and accurate) results of cognitive therapy is the best supported at
interpretation of sensations and signs. In the present. Indeed, there is no plausible alternative
treatment of Hs, greater emphasis is needed on the explanation for the effects of cognitive therapy at
probability estimates made by patients, because of present.
their postulated tendency to overpredict the The advance of cognitive therapy, and its
likelihood of health disasters (see Rachman & successes in treating panic, gave rise to the first
Arntz, 1990 on overpredictions of aversive and only major psychological alternative to the
events). Salkovskis and Warwick (1986) and then widely accepted biological theory of panic
Warwick and Salkovskis (1990) shaped their (see Rachman & Maser, 1990). The next big
cognitive therapy of Hs accordingly and the early debate will center on the first and only major
results are promising, as are the findings from a psychological alternative to the then widely
completed but not yet fully analyzed controlled accepted biological theory of panic (see Rachman
trial (personal communication, Salkovskis, 1993). & Maser, 1990). The debate will also incorporate
The postulated mechanisms of change in competing explanations for obsessional-
treating Hs are similar to those for panic, but differ compulsive disorders.
in detail, with greater emphasis on remote threats Most of the available data are from studies of
and on the correction of inflated probabilities of panic, occasionally supplemented by other work.
illness/injury. Confirmation of the cognitive Therapeutic reductions in the frequency/intensity
mechanisms of panic disorder would strengthen of panic are associated with reductions in negative
the likelihood of similar therapeutic actions in cognitions. The magnitude of the therapeutic
treating hypochondriasis, just as a disconfirmation improvements is impressive. The results with
would reduce the plausibility of a cognitive other disorders are promising but scant. In
explanation for the reduction of Hs. Anxious addition, there is a growing amount of indirect
people are easily able simultaneously to fear evidence, mainly in the form of support for the
imminent and remote threats to their health and underlying theory of panic causation. Cognitive
well-being, and hence a high degree of co- therapy is legitimately deduced from the theory,
morbidity is to expected between Hs and panic and the theory itself is gaining some support
disorder (Rachman, 199 I ). (Rapee, 1993). This cohesiveness of the theory
In summary, the cognitive approaches to and therapy is an added source of strength.
phobias, OCD and hypochondriasis have a It remains to be shown that the reduction of
common core that is derived mainly from the panic is conditional on the reduction of critical
cognitive theory of panic. Intense anxiety arises negative cognitions. The temporal relations
when the person makes a catastrophic between cognitive change and panic reductions
misinterpretation of internal or external events need close investigation, and a variety of other
such that a serious threat is perceived. The idea is questions need to be answered. On a broader scale,
that if the misinterpretation can be identified, we need to determine the extent to which the
modified and then replaced by a more benign cognitive explanation can account for the
 A Critique of Cognitive The1‘spy for Anxiety Disorders 287

therapeutic effects of exposure treatment and depression and accessibility of memories. Journal of
Abnormal Psychology, 91,87-95.
pharmacological treatment. As a first step we need Craske, M., & Rodriguez, B. (1993). Behavioral treatment of
to investigate the cognitive changes that panic disorders and agoraphobia. Progress in Behavior
accompany those other treatments. While aiming Modification.
Ehlers, A. (1993). Interoception and panic disorder. Advances
for a parsimonious explanation of the effects of in Behavior Research & Therapy, 15, 3-22.
therapy, we should leave open the possibility that Emmelkamp, P., & Beens, H. (1991). Cognitive therapy with
a single explanation may prove unattainable. One OCD: A comparative evaluation. Behaviour Research &
Therapy, 29,293-301.
possible candidate is the emotional processing Klein, D., & Klein, H. (1988). The nosology of anxiety
model (Rachman, 1980,199O). disorders: a critical review of hypothesis testing about
The implications of the cognitive theory and spontaneous panic. In P. Tyrer (Eds.), Psychopharmacology
ofanxiety. Oxford: Oxford University Press.
the related research are manifold. Confirmation of Klosko, J., Barlow, D., Tassinari, R., & Cemy, J. (1990). A
the cognitive theory account of anxiety disorders, comparison of alprazolam and behaviour therapy in
namely that dysfunctional cognitions play the treatment of panic disorder. Journal of Consulting &
Clinical Psychology, 58, 77-84.
major role in generating and sustaining abnormal Last, C. (1987). Simple phobias. In L. Michelson & M. Ascher
anxiety and avoidance, will put us in a position to (Eds.), Anxiery and stress disorders. New York: Guilford
improve our current methods of treatment and, Press.
Margraf, J., Barlow, D., Clark, D., & Telch, M. (1993).
importantly, enable us to provide treatment for the
Psychological treatment of panic. Behavior Research &
full range of these disorders. However, if the Therapy, 31, 1-8.
cognitive theory is disconfirmed, we will be left Margraf, J., & Schneider, S. (1991). Outcome and active
ingredients of cognitive-behavioural treatments for panic
with a collection of embarrassingly successful disorder. AABT Conference, New York, November 1991.
therapy results that evade explanation, and two Marks, M., Basoglu, M., Alkubaisy, T., Senguia, S., & Marks,
richly talented psychologists will need re-training I. M. (1991). Are anxiety symptoms and catastrophic
cognitions directly related? Journal of Anxiefy Disorders, 5,
as coach-tour guides in Oxford.
241-254.
ijst, L. (1989). One-session treatment for specific phobias.
Behaviour Research & Therapy, 27, l-8.
Rachman, S. (1980). Emotional processing. Behavior Research
References C Therapy, 18,5 l-60.
Rachman, S. (1990). Fear and courage, 2nd Edition. W. H.
Amtz, A. (1992). Application of cognitive therapy for OCD. Freeman: New York.
World Congress Cognitive Therapy, Toronto. Rachman, S. (1991). A psychological approach to the study of
Barlow, D. (1988). Anxiety and its disorders. Guilford Press: co-morbidity. Clinical Psychology Review, II, 461464.
New York. Rachman, S. (1993). Obsessions, responsibility and guilt.
Beck, A. (1988). Cognitive approaches to panic disorder. In S. Behaviour Research & Therapy, 31,149-154.
Rachman & J. Maser (Eds.), Panic: psychological Rachman, S., & Arntz, A. (1991). The overprediction and
perspectives. New Jersey: Erlbaum. underprediction of pain. Clinical Psychology Review, II,
Booth, R. & Rachman, S. (1992). The reduction of 339-355.
claustrophobia. Behaviour Research & Therapy, 30, Rachman, S., & Levitt, K. (1988). Panic, fear reduction and
201-222. habituation. Behaviour Research & Therapy, 26, 199-206.
Chambless, D. L. (1988). Cognitive mechanisms in panic Rachman, S., K. Levitt, & Lopatka, C. (1987). Panic: the links
disorder. In S. Rachman and J. Maser (Eds.), Panic: between cognitions and bodily symptoms - 1. Behaviour
psychological perspectives. New Jersey: Erlbaum. Research & Therapy, 25,4ll-I24.
Clark, D. M. (1986). A cognitive approach to panic. Behaviour Rachman. S., & Maser, J. (Eds.). (1988). Panic: psychological
Research & Therapy, 24.461470. perspectives. New Jersey: Erlbaum.
Clark, D. M. (1988). A cognitive model of panic attacks. In S. Rachman, S., & Whittal, M. (1989). Fast, slow and sudden
Rachman & J. Maser (Eds.), Panic: psychological reductions in fear. Behaviour Research & Therapy, 27,
perspectives. New Jersey: Erlbaum. 613-620.
Clark, D. (1993). Cognitive mediation of panic attacks. Rapee, R. (1993). Psychological factors in panic disorder.
Advances in Behaviour Research & Therapy, 15, 75- Advances in Behavior Research and Therapy, 15,85-102.
84. Salkovskis, P. (1985). Obsessional compulsive problems: a
Clark, D. M., Salkovskis, P., Hackmann, A., & Gelder, M. cognitive behavioral analysis. Behaviour Research &
(1991, November). A comparison of cognitive therapy, Therapy, 5,571-583.
applied relaxation and imipramine in the treatment of panic Salkovskis, P. (1992). Cognitive models and therapy for OCD.
disorder. A.A.B.T. Meeting, New York. World Congress Cognitive Therapy, Toronto.
Clark, D., & Teasdale, J. (1982). Diurnal variation in clinical Salkovskis, P., & Clark, D. (1993). Panic disorder and
288 S. RACHMAN

hypochondriasis. Advances in Behavior Research and Teasdale, J. (1988). Cognitive models and treatments for panic:
Therapy, 1.5, 2348. a critical evaluation. In S. Rachman & J. Maser (Eds.).
Salkovskis, P., Clark, D., & Hackmann, A. (1991). Treatment Panic: p.sychological perspecritzes. New Jersey: Erlbaum.
of panic attacks using cognitive therapy without exposure or Van Oppen, P. (1992). Behavior therapy vs cognitive therapy
breathing retraining. Behaviour Research & Therapy, 2Y, for OCD. World Congress Cognitive Therupy, Toronto.
161-166. Visser, S., & Bouman, T. (1992). Cognitive-behavioural
Salkovskis, P., & Warwick, H. (I 986). Morbid preoccupations, approaches to hypochondriasis. Behaviour Research &
health anxiety and reassurance: a cognitive-behavioural Therapy, 30, 301-306.
approach to hypochondriasis. Behaviour Research & Warwick, H.. & Salkovskis, P. (1990). Hypochondriaais.
Therapy, 24,597-602. BehmGxu Research & Therapy. 28, 105-I 18.
Seligman, M. E. P. (1988). Competing theories of panic. In S. Wolpe, J., & Rowan, V. (1989). Panic disorder: A product of
Rachman and J. D. Maser (Eds.), Panic: psychological classical conditioning. Behaviour Reseurch X Thenrpy, 27,
perspecfives. New Jersey: Erlhaum. 583-585.
Shafran, R., Booth, R., & Rachman, S. (1992). The reduction
of claustrophobia II - cognitive analyses. Behaviour
Research & Therapy, 30, 75-85.