Annotated Bibliograghy
budsimpson
ACOGNITIVEAPPROACHTOMEASURINGAND.pdf
Pergamon
Behav. Res. Ther. Vol. 32, No. 3, 355-367, 1994 pp. Copyright 0 1994 Elsevier Science Ltd
00057967(93)EOO12-T Printed in Great Britain. All rights reserved
0005-7967/94 $6.00 + 0.00
A COGNITIVE APPROACH TO MEASURING AND MODIFYING DELUSIONS
P. D. J. CHADWICK’* and C. F. LOWE* ‘School of Psychology, University of Birmingham, Edgbaston, Birmingham B15 2TT, U.K
2University College of North Wales, University of Wales, Bangor, U.K.
(Received 11 May 1993)
Summary-The present paper summarizes and integrates with the existing literature the results of three studies we have conducted on the measurement and modification of delusions. The findings of two of these studies have been presented previously; the third is briefly reported here. A total of 12 people with delusions took part. Ten participated in two investigations that used between-subject multiple-baseline designs; the remaining two, each of whom held three distinct delusions, took part in a study using an across-beliefs multiple-baseline design. A variety of dimensions of delusional experience were monitored over baseline periods of at least 4 weeks, and two distinct cognitive interventions were used: a structured verbal challenge and a planned empirical test. Our focus in the present article is on intervention and the process of change as people come to question and sometimes reject their delusions. We also address related issues, including problems of measurement (i.e. demand characteristics, independent validation), the connection between depression and delusions and the prediction of treatment response. We conclude with specific recommendations for cognitive therapy for delusions.
THE COGNITIVE PERSPECTIVE ON DELUSIONS
Our theoretical perspective on delusions has been influenced by two main sources, the literature on verbal self-regulation of behaviour (Lowe, 1979; Lowe, Horne & Higson, 1987) and Maher’s (1974, 1988) work on delusions. Our interest has been on the common features of delusional thinking. Thus, our experimental strategy has been directed at a variety of delusion types (paranoid, grandiose, control and identity). Maher (1974, 1988) proposed that a delusion can be regarded as a normal attempt to make sense of an abnormal perceptual experience. A clear paradigm case would be a delusion that was secondary to auditory hallucinations, the argument here being that the hallucinations puzzled and perhaps distressed the individual concerned, and so he or she searches for a meaningful explanation of them. The delusion would arise from this effort after meaning, and would be invested with the psychological force of having rid the individual of the sense of bewilderment. According to Maher, the reasoning processing that produces delusions does not differ from that which produces so-called ‘normal’ beliefs; it is just that bizarre perceptions demand bizarre explanations.
Maher’s contention that a delusion may be rational, although incorrect, has been questioned recently with the discovery that people with delusions have biased reasoning (see Bentall, Kinderman & Kaney, 1994, this issue pp. 331-341). Under certain experimental conditions people with delusions appear to show bias in their attributional style, in their judgement of covariance, and in their probabilistic reasoning (Garety, 1991). However, this raises a number of consider- ations. First, it is sometimes difficult to interpret such findings vis-ci-vis rationality. For example, Huq, Garety and Hemsley (1988) using a neutral task which required subjects (Ss) to make inferences about the likely ratio of different coloured beads in a jar, investigated the probabilistic reasoning of a group of deluded people, a group of mixed psychiatric patients who were not deluded and a group of matched controls. People with delusions were found to require the least beads to be drawn from the jars before forming their conclusions and also to express most confidence in their decisions, and yet this ‘jumping to conclusions’ was nearer optimum reasoning than the caution displayed by other Ss. Also, it is not always clear how specific findings in analogue studies might apply to delusional thinking. For example, how might it be established if delusions are
*Author for correspondence.
355
356 P. D. J. CHADWICK and C. F. LOWE
formed on the basis of less information than, say, religious beliefs or depressive beliefs? Second, the evidence is for bias, not deficit; it might reasonably be inferred that the observed bias is a consequence of delusional behaviour, rather than the delusion being a consequence of the bias.
TO distinguish between rational and irrational thinking is not straightforward, Indeed, it may be that debate about the rationality of delusions runs the risk of obscuring the main value in Maher’s theory-i.e. that in delusions the primary underlying motivation is attributional, and that they may be thought of as an attempt to make sense of particular events. (It has been an integral part of our approach to challenging delusions that the person be encouraged to construe his or her delusion as a reaction to, and an attempt to make sense of, particular events, and that an alternative perspective from which to understand these events be supplied and evaluated.) Viewing delusions in this way has some interesting implications. One is that although delusions may not be culturally shared, they are grounded in external events. In an important sense the person, not a hidden pathology, constructs the belief. A related implication, which is born out in our research, is that people can presumably describe some, or perhaps all, the experience (‘evidence’), their delusions were invoked to explain, and their accounts might shed light on the process of delusion formation.
Maher’s perspective agrees well with the writing of Vygotsky (1962) and Luria (1961) who charted the developmental process by which language gradually acquires a regulatory function and comes to organize and drive behaviour. The regulatory function of language is surprisingly potent. For example, simply giving objects a label appears to allow young children to use them in a way that was not possible before, and also helps to abstract the principle that guided this organization and apply it elsewhere (Luria, 1961). Delusions, like other beliefs, alter our relation to the environment, because they organize and structure future experience. It is important to recognize that a beliefs regulatory potency does not hinge on its accuracy, nor on the degree to which it is the result of reason, nor on its adaptive value. Several studies of human learning have shown that the rules people form about contingencies of reinforcement are often not based on correct inference and do not produce optimum responding, yet they exercise a unique control over behaviour and are extremely resistant to change (Lowe, 1979, 1983; Horne & Lowe, 1993). According to this account, there are important respects in which delusions are like other beliefs. This new emphasis on continuity, rather than discontinuity, is nowhere more evident than in the literature on defining and measuring delusions.
Traditionally, delusions have been defined on the basis of empirical claims of discontinuity- e.g. as beliefs that were undeniably false, that were held with total and unshakable conviction, that were not shared by others with the same cultural background and that were based on incorrect inference (DSM-III-R; APA, 1987). Individually these criteria have been disputed: thus, a delusion need not be false (Brockington, 1991), it need not be held with absolute or unshakable conviction (Brett-Jones, Garety & Hemsley, 1987; Watts, Powell & Austin, 1973) and it need not be based on incorrect inference (Garety, 1991). The criterion relating to the unusual content of delusions also may be questioned, since research has demonstrated how difficult it is to rate the ‘bizarreness’ of delusions (Kendler, Glazer & Morgenstern, 1983). Traditional criteria have also been challenged by a radical and exciting call to define delusions as points on a continuum with normality, the position on this continuum being influenced by dimensions of delusional experi- ence such as degree of belief conviction and the extent of preoccupation with the belief (Strauss, 1969).
As well as stressing continuity, this new perspective also places great emphasis on the individual and on individual differences. For this reason, in our three studies we used multiple-baseline methodology (Kazdin, 1982) with its combination of experimental control together with the power to reveal individual diversity. Before, during, and after therapy, we measured the degree of conviction with which the delusion was held, the degree of preoccupation with the delusion, the concurrent degree of anxiety and the extent to which the person observed confirmations and disconfirmations. We also gave the Beck Depression Inventory (BDI; Beck, Rush, Shaw & Emery, 1979) during baseline, intervention, and follow-up to assess for unwanted effects of the loss or weakening of a delusion. One of the most striking aspects of our findings on the modification of delusions was the very individual nature of the process and outcome of change, and for this reason we strongly favour the use of sophisticated single-subject methodology in this field.
Cognitive approach to delusions 357
ARE DELUSIONS OPEN TO MODIFICATION?
On occasion, psychiatric opinion appears to be absolutistic and very difficult to change. For example, it is often asserted that delusions cannot be modified or are utterly insensitive to reason (see Garety, 1985) and yet there are empirical grounds for rejecting this opinion. There has been a modest number of studies, including our own, reporting attempts to weaken delusions, with generally favourable results (Alford, 1986; Beck, 1952; Fowler & Morley, 1989; Hartman & Cashman, 1983; Hole, Rush & Beck, 1973; Johnson, Ross & Mastria, 1977; Milton, Patwa & Hafner, 1978). Whilst it is important to admit some methodological shortcomings, and the need to develop better measures and therapies, it seems unwise to deny a priori these findings and their implications.
It might be more reasonable to assert that delusions are difficult to modify, sometimes fiendishly so. This position would acknowledge that the class of beliefs called delusions varies considerably along a number of dimensions, and it would encourage examination of the multitude of factors which might be thought to influence therapeutic outcome. It would also encourage an exploration of whether delusions associated with a diagnosis of schizophrenia are more difficult to modify than political or religious beliefs, or the core beliefs associated with conditions such as depression and anorexia.
It is, of course, as fallacious to accept research uncritically as to reject it out of hand. Research on modifying delusions is in its infancy, and there are methodological and conceptual problems that render many of the findings suggestive rather than definitive. We are nevertheless optimistic that psychological therapy has a role to play, as we hope our findings will demonstrate, but we recognize the need for caution and for critical appraisal of current theories and methods.
COGNITIVE THERAPY FOR DELUSIONS: AN EXPERIMENTAL ANALYSIS
We carried out three studies on the use of cognitive therapy (CT) to weaken delusional thinking in people with a diagnosis of schizophrenia. In total, 12 outpatients took part, 8 men and 4 women, all of whom were on stable neuroleptic medication. They ranged in age from 28 to 63 years (average 36). All described clear delusions, which had been voiced at varying points over at least the preceding 2 years, though usually far longer. However, it should be noted that though delusions may be mentioned over a number of years, this is not necessarily an accurate guide as to how long they have been held (Strauss, 1991).
CT usually pursues cognitive change through a combination of verbal challenge and planned reality testing (Beck et al., 1979). In our research on delusions we used each of these strategies in separate interventions, and manipulated the order in which they were introduced in an attempt to understand more about how each works. Both interventions were applied within an atmosphere of ‘collaborative empiricism’ (Beck et al., 1979). Rather than being told that they were wrong, individuals were encouraged to see their delusions as being only one possible interpretation of events and they were asked to consider and evaluate an alternative view. Also, because of our dissatisfaction with current definitions of schizophrenia and delusion, in speaking to the individuals we did not use these labels to describe their experience.
The verbal challenge comprised four elements. To minimize the likelihood of psychological reactance (Brehm, 1962) the evidence for the beliefs was challenged first; the order in which each piece of evidence was challenged was inversely related to its importance to the delusion (see Watts et al., 1973). An integral part of this discussion involved the therapist making clear to individuals how strongly held beliefs can exert a profound influence over their behaviour and interpretation of events. Second, the internal consistency and plausibility of the belief system was questioned, and all irrational or inconsistent features were highlighted. Third, following Maher, an alternative perspective was offered, namely, that the belief was formed in response to, and as a way of making sense of, a specific experience-often this included a primary symptom, but in several cases it was hypothesized that the delusion was in part a response to important life events. For example, a link was drawn and discussed between patient DR’s belief that he could communicate with people from the past, and thereby prevent accidents and disasters, and the accidental death of his father shortly
358 P. D. J. CHADWICK and C. F. LOWE
before the onset of symptoms. Lastly, the individual’s delusion and the therapist’s alternative were assessed in light of the available information.
The main feature of our reality testing intervention was that it involved the individual planning and performing an activity that could invalidate the delusion, or some part of it (Hole et al., 1973). Beck et al. (1979) called such activities behavioural experiments, conveying that they were performed in order to test a hypothesis. Strategically, the primary purpose of both empirical testing and verbal challenge is to bring about cognitive change, although empirical testing is thought to be the more powerful intervention. As Beck put it:
“There is no easy way to ‘talk the patient out’ of his conclusions that he is weak, inept, or vacuous By helping the patient change certain behaviours, the therapist may demonstrate to the patient that his negative, overgeneralized conclusions were incorrect.” (Beck el al., 1979, p. 118. original emphasis.)
The effects of the interventions on three outcome measures (conviction, preoccupation and
Conviction
Preoccupation
Anxietv TD
100 5
4
100
50
0
100
50
s
5 0 ._ ‘, ‘5 100
g V
50
0
EE
50 3
2
1
0 0 Baseline Verbal Reality Follow-up
challenge tesring
Fig. 1. Bar graph summarizing 6 individuals’ (HM, TD, WH, EE, MM & DR) weekly conviction, preoccupation and anxiety scores during baseline, verbal challenge, reality testing and follow-up.
Cognitive approach to delusions
HI
IC onviction
359
100 5
4
3 50
2
1
0 0 x
100 5 z ._
4 2
3 ‘c) g 50 2 5
C z 1 .8
‘- 6 0 0 ; ._ 2 100 5 :: 6
4 s EL
50 3
2 z 5
1 ::
0 Ozz 100 5
4
3 50
2
1
” 0 Baselille Reality Verbal Follow-up
testing challenge
Fig. 2. Bar graph summarizing 4 individuals’ (HJ. LJ, CE & FD) weekly conviction, preoccupation and anxiety scores during baseline, reality testing. verbal challenge and follow-up.
anxiety) are shown in Figs 1-3. Figure 1 summarizes data from 6 individuals who received the verbal challenge first, 3 of whom went on to receive the reality testing second [for detailed sessional results see Chadwick and Lowe (1990)]. Figure 2 summarizes data for a further 4 individuals who all received the reality testing first and the verbal challenge second [see Chadwick and Lowe (in press) for details]. The two studies reported in Figs 1 and 2 used multiple-baseline designs across individuals. Figure 3 summarizes two single case studies, in both of which the verbal challenge was used in a multiple-baseline across delusions design to challenge three delusions; the three delusions were challenged successively at intervals of 4 or more weeks in the sequence of ordering used in Fig. 3 [for a full account, see Lowe and Chadwick (1990)]. Briefly, the delusions were as follows: BP believed that a woman, Amanda, was reading his mind and controlling his life, and that in past lives he had been Leonardo da Vinci and Jesus Christ; BG who was 51 years old, believed that she was only a teenager, that she was the daughter of Princess Anne, and that the Government was controlling her in elaborate and fantastic ways.
Conviction
Of the 12 participants, 10 (83%) reported reductions in their level of belief conviction as a consequence of the interventions, and 5 had rejected their delusions altogether by the close of the intervention period. In 10 cases conviction remained stable over baseline periods of between 4 and 10 weeks. Two individuals maintained their baseline levels of conviction throughout the interven- tion phases. For all individuals these changes were corroborated during assessment interviews with independent clinicians, and in 5 cases there was additional authentication from either a second professional or relative.
The verbal challenge was a strong opening intervention. In the study depicted in Fig. 1, verbal challenge periods of &6 weeks provoked stable and large reductions in three (HM, TD & DR) of the 6 individuals and a substantial but transient reduction in a further person (EE). In spite of
360 P. D. J. CHADWICK and C. F. LOWE
this, when reality testing followed the verbal challenge, in 2 out of 3 cases (EE & MM) it improved upon the effectiveness of the verbal challenge alone, suggesting that reality testing is a strong intervention when it follows the verbal challenge. Both individuals shown in Fig. 3 had rejected their three delusions by the close of the 14 sessions of verbal challenge. However, it should be noted that in the case of BP, an apparent generalization effect meant that conviction in the Leonardo belief had already fallen to only 10% by the time it was challenged.
When reality testing was used as a first intervention over 224 weeks with 4 people (Fig. 2) only LJ showed a substantial drop in conviction, and this effect was transient with conviction returning to baseline level by the close of the reality testing period; her conviction fell again at the start of the verbal challenge period and this reduction was maintained. For the remaining 3 individuals, reality testing had little or no effect, although 2 (HJ & EE) showed a large fall in conviction during the subsequent verbal challenge period. For CE the verbal challenge, like reality testing, failed to alter conviction.
Verbal challenge
Follow-up y
Fig. 3. Bar graph summarizing 2 individuals’ (BP & BG) weekly conviction, preoccupation and anxiety scores for each of their three delusions during baseline, verbal challenge and follow-up.
Cognitive approach to delusions 361
Table I. BDI scores for all 12 participants at the final baseline session, the final intervention session and at the I-, 3-, and 6-month follow-ups (FU)
Individual Baseline Intervention I-Month FU 3-Month FU 6-Month FU
HM TD WH EE MM DR HJ LJ CE FD BP BG
I9 35 I8 24
20 23 28 9
2s I3 I7
I4 27 17 20 0
I9 I2 24 9
22 5 6
I7 27 8 6 0
17 IO 18 IO 20 ^ 5
5 5 0
IO IO I6 I2 20 6
II
3 27 7 5 0 4 8
IO 8 9
24 8
On the basis of these results, reality testing on its own would seem to be a weak intervention, perhaps because strongly held beliefs are ‘immunized’ against disconfirming events (Popper, 1977). This finding gives preliminary empirical support to a hitherto untested assumption about CT, namely, that empirical testing is more effective when it follows a period of verbal challenge (e.g. Trower, Casey & Dryden, 1988).
Preoccupation and anxiety
At the close of every session we used personal questionnaire methods to extract a retrospective rating of degree of preoccupation (the number of times the person thought about the delusion in the preceding week) and the degree of anxiety experienced at these times. When these scores are compared with the conviction scores (see Figs l-3), two clear patterns emerge. The first is that during baseline, scores on preoccupation, anxiety and conviction frequently varied independently (see Chadwick & Lowe, 1990) which supports the view that delusions are multidimensional phenomena. Second, during the intervention period more often than not the three measures tended to change together. For example, of the 10 people whose conviction ratings fell during the interventions, 6 also reported reduced preoccupation ratings and 6 reported lower anxiety scores. However, there were exceptions; in 3 cases, conviction and anxiety fell but preoccupation rose (LJ) or remained very high (TD & MM). On the other hand, the results from WH show how preoccupation and anxiety may fall even when conviction is unchanged. These exceptions again emphasize both how complex delusions are, and how clinical improvement may take many different forms.
Depression
Zigler and Glick (1988) postulated that paranoia might be a defence against low self-esteem, and that paranoid schizophrenia might therefore be defended depression. Since then, Bentall and his colleagues have amassed considerable support for the possibility that paranoia represents a defence against low self-esteem, although seemingly not against depression (see Bentall, 1994). Conse- quently, in our research the BDI was given once at the close of baseline, once at the close of the intervention and once at each follow-up meeting.
These results offered no suggestion that the weakening or loss of a delusion had a detrimental effect (see Table 1). Indeed, quite the opposite trend was observed; for each of the 10 individuals whose conviction scores fell during the course of the interventions, the BDI score fell also. In 9 of these 10 cases, during the follow-up period the BDI scores either continued to fall or remained at the same level as at the close of the intervention. The one exception, BP, recorded his highest BDI score during the follow-up period. BP was also unique in showing an apparent ‘symptom substitution’, when soon after giving up his Amanda belief he reported believing that a famous sportswoman was now reading his mind and influencing his life in much the same way as had Amanda, and the high BDI score during follow-up corresponded with his own attempts to challenge the substituted belief (Lowe & Chadwick, 1990). Overall, however, the results of Table 1 lead us to endorse the conclusion drawn by Milton et al. (1978, p. 127), that “the fall in strength of delusions appears to be associated with a worthwhile reduction in overall psychiatric disturb-
362 P. D. J. CHADWICK and C. F. LOWE
ante”. This is also consistent with the finding that a fall in conviction scores was often accompanied by a reduction in anxiety scores.
In two ways our findings appear, at first glance, to contradict the view that paranoid schizophrenia is defended depression. First, 5 of the 6 individuals describing paranoid delusions (HM, TD, WH, HJ, CE & FD) were depressed at the onset of the study-Bentall himself has reported the ‘paradoxical’ finding that paranoid people typically display high depression as well as high self-esteem (Bentall, 1993). However, Zigler and Glick (1988) are in fact making two separate proposals. One is that paranoia defends against depression, and the second is that paranoia defends against low self-esteem. However, someone who scores highly for depression on standard measures such as the BDI may not necessarily have tow self-esteem, For example, it is our experience that paranoid individuals commonly feel sad, discouraged and often suicidal, believe that they are being punished, lose interest in others and experience vegetative symptoms, all without feeling worthless.
It may be, therefore, that paranoia should be thought of not as defended depression, but only as defended low self-esteem. This separation is consistent with diagnostic practice which recognizes that feelings of worthlessness are not a necessary feature of depression (e.g. DSM-III-R; APA, 1987). It also is consistent with the finding that paranoid people are commonly depressed; indeed, recent research into delusions that are secondary to auditory hallucinations has shown that people who believe their voices to be a form of punishment or persecution are significantly more likely to be depressed than those with non-paranoid beliefs (Chadwick & Birchwood, 1994b).
The second respect in which our data seem to contradict Zigler and Glick’s (1988) defended depression theory is the observed connection between a weakening of paranoid delusions and a reduction in BDI scores. All 4 individuals whose paranoid delusions were weakened by the interventions (HM, TD, HJ & FD) reported lower depression scores at the close of therapy and for 6 months there after.
Thus the evidence suggests that, in the short-term at least, the weakening of paranoid thinking does not precipitate either increased depression or, it might be inferred, low self-esteem. Of course, Zigler and Glick’s theory might be maintained by arguing that some of the individuals in our research might eventually fall into a ‘post-psychotic’ depression (DeLisi, 1990) and that unlike those depressions which existed alongside paranoia, these post-psychotic depressions would be characterized by low self-esteem. However, Zigler and Glick offer a theory of the psychological motivation for forming paranoid ideas, and it may well be that those conditions which gave rise to the paranoid delusions in the first place no longer pose the same degree of threat many years later. In other words, different conditions may be responsible for the formation and maintenance of paranoid delusions, and Zigler and Glick’s theory is not weakened if the loss of a paranoid delusion does not lead to low self-esteem.
Accommodation
It is often argued that delusions, indeed all strongly held beliefs, are maintained through a strong confirmation bias (Maher, 1988). A study of delusions by Hole et al. (1973) found evidence of a confirmation bias, and also showed that for some people this bias diminished as they became less certain that their delusions were true. Consequently, we measured accommodation (Brett-Jones et al., 1987), i.e. the extent to which people observed and assimilated events which confirmed or disconfirmed their delusions. Accommodation was measured at the start of every session by asking the individual if anything had happened over the past week to alter his or her belief in any way.
We found that delusions were maintained in a remarkably similar way. In a total of nearly 100 baseline sessions, only 2 disconfirming events were reported, both by BG. And yet disconfirmation was occurring, and in quite direct ways. DR, for example, who believed that he could communicate with people in the past and thereby change the course of history, tried and failed to do this numerous times each day and yet never felt compelled to abandon the delusion. All 12 individuals frequently reported events which they perceived to support their delusions, though many of these were not obviously supportive. For example, BP reported how a young dishevelled woman had once asked him for money at a bus stop. Although at the time he did not recognize the woman, he later deduced that she had been Amanda in disguise and that she had planned the meeting to remind him of her presence. Much of the evidence was of this type, that is, reconstruction of past
Cognitive approach to delusions 363
events in line with a delusion. For instance, FD went through University enjoying satisfactory contact with his lecturers and only years later came to ‘realize’ that they had been reading his mind and persecuting him.
Six of the ten individuals whose conviction scores fell during CT reported at least one instance of disconfirmation during the intervention period; only BG experienced disconfirmation during the baseline period. These reports of disconfirmation fell into two categories. First, people recalled situations which initially they had seen in delusional terms, but then subsequently had construed instead as weighing against the delusions. For example, as part of his delusion HM believed that women in cafeterias were sexually interested in him. At the start of the second session of verbal challenge, he reported an occasion when a woman had sat by him in a cafeteria; at first he interpreted her behaviour as a sexual invitation, but then looking around and discovering that all the other seats were taken he was able to refute this interpretation for himself. Second, people recalled situations which at no point were interpreted in delusional terms, but were seen simply and immediately as going counter to their delusions. For example, LJ believed that she was an evil person who contaminated those around her. During the verbal challenge phase she reported how she had been walking in town with her son when her ex-husband (the boy’s father) had walked straight past them without stopping to talk, much to the distress of her son. She believed that this had been a deliberate attempt to upset her son and judged it to have been an ‘evil’ thing for her ex-husband to do and unlike any of her actions. At no time did this event actuate her belief that she was evil, even though prior to CT seeing her ex-husband did have this effect.
Thus, like other researchers, we observed how delusional thinking is typified by a bias towards confirmation, and how this bias may lessen during CT. It is important, however, not to infer from this specific bias in delusional thinking a bias or deficit in thinking generally. Our data offer no support for the existence of a general confirmation bias in deluded individuals-indeed, evidence exists that under certain experimental conditions, people with delusions may be fess attached to their formulations than psychiatric and ordinary controls (see Garety, 1991). So too, our data do not suggest that CT has an impact on fundamental thought processes. Whilst certain individuals appeared to display a weakened confirmation bias in their thinking within the context of a particular delusion, we would not assume that these individuals became less disposed to find support for other central beliefs-indeed, clinical practice suggests that delusions, depressive beliefs etc, can be surprisingly independent.
PROCESSES OF CHANGE IN CT FOR DELUSIONS
Before concentrating on delusions, it may be helpful to consider how CT in general works. The ‘primary premise’ (Beckham, 1990) of CT is that cognitive change precedes behavioural and affective change. To date, although there is ample evidence that CT can produce cognitive, behavioural and affective change (Hollon, Shelton & Loosen, 1991), the primary premise that cognitive change is the first to occur has not been established. Indeed, more recent formulations of the cognitive model throw into question whether a neat temporal relationship is likely to exist between these variables (Muran, 1991).
Currently CT is thought to influence beliefs either through activation or compensation (Barber & DeRubeis, 1989) and in our research we found evidence of both processes. Deactivation occurs when a new belief is formed which shares some attributes with the delusion and which is then accessed in preference to it, such that the delusional thought no longer occurs. Compensation occurs when delusional thoughts still occur, but the individual responds to them differently and uses strategies to restrict their regulatory influence on behaviour and feelings.
One means of measuring deactivation is to monitor the number of conscious delusional thoughts an individual has (i.e. preoccupation); if a delusion is deactivated then preoccupation scores should be highest during baseline. Six of the ten people whose conviction ratings fell during the interventions also reported reduced preoccupation ratings and their comments during sessions and at independent assessment interviews supported the conclusion that certain stimuli that had once prompted delusions now triggered different beliefs. For example, during baseline, BG’s auditory hallucinations invariably prompted delusional thoughts, and yet by the close of therapy this connection was broken and she usually saw the hallucinations as originating from her own mind.
364 P.D.J. CHADWICK~~~C. F. LOWE
BP made the same point regarding a pulsating in his temple which he formerly attributed to Amanda reading his mind and now attributed to tension.
It might therefore reasonably be argued that one process whereby CT weakened delusions was deactivation. However, in 2 people (MM & LJ) conviction fell and preoccupation did not, and even when the 2 measures fell together only 1 person (BP) consistently reported no preoccupation with his delusions. In other words, most people still thought about their delusions some of the time, albeit less often than prior to CT. What was noticeable was that many times people reported how they dealt with these thoughts differently as a consequence of CT-i.e. they displayed ‘compen- sation’. Usually this involved rehearsing those arguments against the delusion that the client had found most persuasive in CT, so halting what regulatory control the delusion might otherwise have exercised over future behaviour and affect. Less commonly, compensation even extended to testing the delusional interpretation of an event. A clear example of this was when HM scanned the cafeteria in which he was seated in order to test his impression that a woman had chosen to sit by him as a form of sexual invitation.
Our impression is that deactivation and compensation are not competing mechanisms, but different features of one dynamic process. At the start of therapy, delusions regulate behaviour and affect predominantly without effort or awareness on the part of the person, and verbal control is so well-established as to be ‘fossilized’ (Vygotsky, 1962). As CT progresses, this process is brought back into awareness and new ways of seeing events are discussed. Gradually these alternative beliefs become part of the person’s effortful (i.e. compensatory) thinking and, with time, become increasingly automatic (i.e. deactivation has occurred). To call deactivation a separate process seems puzzling, because the very notion of a new belief assuming regulatory control seems to imply some amount of compensation; conversely, prolonged compensatory thinking would seem certain to lead to reduced delusional thinking. Thus, it may be that non-delusional thinking evolves in the same way as delusional thinking, that is, a dynamic progression through predominantly effortful thinking to predominantly ‘fossilized’ automatic thought, but that this balance will always remain sensitive to events and mood.
PREDICTING TREATMENT RESPONSE
An issue that all individuals who participated in our studies raised at the independent assessments was the importance of a sound therapeutic relationship. All participants stated that a good therapeutic relationship had been established, and that they had welcomed the opportunity to discuss their delusions without being criticized or judged. This may in part explain the fact that attendance was very good at all stages of the research. It is likely that a good therapeutic relationship is necessary if a person is to benefit from CT, although judging by the fact that conviction scores were unchanged over long baseline periods, it is unlikely to be sufficient in itself to weaken beliefs.
Counting preparatory interviews, from 6 to 12 sessions were spent with each person before beginning to challenge the delusions. This long listening period allowed the relationship to grow before intervention began, perhaps thereby limiting psychological reactance. Thus, whilst several participants reported feeling anxious when the disputing did begin, not one dropped out. Only 1 person (WH) was hostile during therapy, once stating that the researcher was the type who would like to hold a machine gun to him; in his case therapy had no noticeable impact on the delusion (see Fig. 1).
Reaction to hypothetical contradiction (RTHC) was developed by Brett-Jones et al. (1987) as a measure of people’s potential for accepting disconfirmation. A plausible hypothetical event that contradicted the delusion was described and the person was asked how if at all such an occurrence would alter the belief. For example, one of BG’s delusions was that she was the daughter of Princess Anne, and she was asked if her belief would be altered in any way if Princess Anne met her and said that the belief was false. EE, who believed that Elvis Presley on his death had taken over EE’s mind and body, was asked if this belief would be altered in any way if Elvis appeared on television and said that he had stage-managed his death to avoid the glare of the public eye. RTHC was measured twice during baseline to see if people’s susceptibility to hypothetical contradiction was related to their susceptibility to CT.
Cognitive approach to delusions 365
On the basis of our data, it would appear that RTHC does indicate treatment response: 8 of the 12 participants responded on at least one occasion that the instance of hypothetical contradiction would lead them either to reject their delusions or to lower their degree of belief conviction and all 8 did lower their degree of conviction in the face of the interventions. Conversely, 2 of the 4 people who stated that the instance of hypothetical contradiction would not affect their delusions did not lower their belief conviction during the subsequent interventions. However, the relationship was not one-to-one; 2 individuals gave negative responses to RTHC that were at odds with their positive response to treatment.
It is our impression that some individuals are far more prepared than others to question their delusions and that results on the RTHC measure reflect this. (In this regard it is important to note that all participants in our research volunteered to discuss their beliefs and this may well have an implication for external validity.) Whilst a change in belief conviction is undoubtedly related to weight of counter-evidence, CT is rarely so persuasive that people have no option but to relinquish their core beliefs. Indeed in our research the belief conviction of 2 Ss was unmoved by the interventions (WH & CE) and a further 4 Ss’ conviction was untouched by the first intervention they received (MM, HJ, CE & FD). In part, openness seemed to be linked to the degree of affect that the delusions still commanded. For example, 2 people with delusions related to long-standing auditory hallucinations (HM & BG) gave the impression at the outset of our study of no longer being ‘caught in their voices power’ (Bauer, 1979) to the degree that they had once been; they had perhaps come to expect that their voices promises and threats were unlikely to be realized. In this vein, BG recalled how many years before the present research she had packed her suitcases and waited in vain by the front door when the voices had told her that the royals were ‘coming for her’. This finding suggests that delusions might not necessarily become harder to modify with time.
Although at the outset it was not clear if treatment might be hindered by the presence of other symptoms, this was not what we found. The 2 people who displayed mild thought disorder (EE & HJ) both responded well. Those 3 people for whom the delusions were secondary to auditory hallucinations (HM, BG & HJ) also responded well to CT. These findings are consistent with those of two recent experiments (Chadwick & Birchwood, 1994a) which have shown that CT may be a useful treatment approach for drug resistant auditory hallucinations, or voices. In the first experiment it was shown how differences in 26 individuals’ affective and behavioral responses to voices reflected vital differences in the beliefs they held about them. Whilst all people believed their voices to be omnipotent and omniscient, additional beliefs about identity and meaning led to some voices being construed as benevolent, others as malevolent. Without fail, ‘malevolent’ voices provoked fear and were resisted and ‘benevolent’ voices were courted. In the second experiment the authors described four examples of the treatment of drug-resistant voices using an adapted version of the cognitive therapy applied to delusions. Where patients were on medication, it was held constant whilst beliefs about the voices’ omnipotence, identity, and purpose were systemati- cally disputed and tested. Large and stable reductions in belief conviction were reported, and these were associated with reduced distress, increased adaptive behaviour, and unexpectedly, a fall in voice activity.
METHODOLOGICAL PROBLEMS
It is important to be precise about what is and is not being claimed of CT for delusions. All participants were on established drug treatments. Although the delusions had apparently resisted medication, medication may nevertheless have had a significant impact on these people’s behaviour and response to CT, Whilst there is evidence that CT can be used to weaken delusions in people diagnosed as schizophrenic who are not on medication (Chadwick & Birchwood, 1994a), the influence of the latter should not be underestimated.
One of the problems inevitably facing research on CT with any disorder is that beliefs are ‘private events’. In our research, for example, it is possible that the reported changes in conviction were responses to demand characteristics, and that at the covert level the delusion was unchanged. The fact that all individuals also reported the changes in belief conviction to at least one person who was not part of the study makes this possibility less likely. Clearly, a behavioural measure of
366 P. D. J. CHADWICK and C. F. LOWE
change, in addition to that of self-report, would be a distinct advantage in future studies, although as we have discussed elsewhere, one may not always be possible (Lowe & Chadwick, 1990).
Another methodological issue is the timing of the measurement of conviction, which was invariably at the close of every session. This raises two questions. First, were baseline conviction scores inflated as a consequence of discussing matters to do with the delusions for an hour or more? Second, were the changes reported at the close of the intervention sessions maintained between sessions? A single case experiment investigating specific cognitive change in CT with depression (Chadwick, in press) provides evidence relating to both questions. It was found that during baseline the person rated his four depressive beliefs just as highly on the Sunday between sessions as he did at the close of sessions. In relation to the second question, early on in the intervention period reductions in conviction scores were not maintained in between sessions, but this effect quickly disappeared and there was once again no difference between scores recorded at the close of sessions and at weekends. The practice of training individuals to score their belief conviction themselves in between sessions would be worth implementing in future research on delusions.
Just as these practical considerations mean that the internal validity of our studies is not beyond doubt, so too their external validity should be questioned. For example, we believe that it would be premature to conclude, on the basis of our findings, that delusions are by and large modifiable in people with a diagnosis of schizophrenia. Such a conclusion ignores the complexity that lies at the heart of contemporary understanding of delusions, of schizophrenia, and indeed of CT. Rather, the data suggest that delusions are complex multidimensional phenomena which need to be studied individually, and that modifiability is an interaction between the individual, the therapist, and the therapy. The weight of evidence, however, now indicates that we should put aside the question of whether delusions can be modified, and concentrate instead upon the more interesting questions surrounding the processes of maintenance and change.
Acknowledgements-We wish to express our gratitude to the 12 people who took part in the research, and to the many colleagues who contributed to the success of the project.
REFERENCES
Alford, B. A. (1986). Behavioural treatment of schizophrenic delusions: a single case experimental Therapy, 17, 63144.
APA; American Psychiatric Association (1987). Diagnostic and statistical manual of menial disorder- Washington, DC: APA.
analysis. Behavior
-revised (3rd edn).
Barber, J. P. & DeRubeis, R. J. (1989). On second thought: where the action is in cognitive therapy for depression. Cognitive Therapy and Research, 13, 441441.
Bauer, (1979). The function of hallucinations: an inquiry into the relationship of hallucinatory experience to creative thought. In Keup, M. (Ed.), Origin and mechanisms of hallucinations (pp. 191L205). New York: Plenum Press.
Beck, A. T. (1952). Successful outpatient psychotherapy of a chronic schizophrenic with a delusion based on borrowed guilt. Psychiafry, 15, 305-312.
Beck, A. T., Rush, A. J., Shaw, B. F. & Emery, G. (1979). Cognitive therapJb of depression. New York: Guilford Press. Beckham, E. E. (1990). Psychotherapy of depression research at a crossroads: directions for the 1990s. Clinical Psycho1og.y
Review, IO, 207-228. Bentall, R. P. (1994). Cognitive biases and abnormal beliefs: towards a model of persecutory delusions. In David, A. S.
& Cutting, J. (Eds), The neuropsychology of schizophrenia (~p.337~360). London: Erlbaum. Bentall, R. P., Kinderman, P. & Kaney, S. (1994). The self attributional processes and abnormal beliefs: towards a model
of persecutory delusions. Behaviour Research and Therapy, 32, MS 027. Brehm, J. W. (1962). A theory of psychological reactance. New York: Academic Press. Brett-Jones, J., Garety, P. A. & Hemsley, D. R. (1987). Measuring delusional experiences: a method and its application.
British Journal of Clinical Psychology, 26, 257-265. Brockington, I. F. (1991). Factors involved in delusion formation. British Journal qf Psychiatry, 159(suppl.), 4246. Chadwick, P. D. J. (in press). Examining specific cognitive change in cognitive therapy with depression: controlled case
experiment. Journal of Cognitive Psychotherapy: an Iniernational Quarterly. Chadwick, P. D. J. & Birchwood, M. J. (1994a). The omnipotence of voices: a cognitive approach to auditory hallucinations.
British Journal of Psychiatry. In press. Chadwick, P. D. J. & Birchwood, M. J. (1994b). Cognitive therapy for voices. In Haddock, G. & Slade, P. (Eds),
Cognirive~behavioral interventions with ps,ychotic disorders. London: Routledge. In preparation. Chadwick, P. D. J. & Lowe, C. F. (1990). Measurement and modification of delusional beliefs. Journal of Consulting and
Clinical Psychology, 58, 225-232. Chadwick, P. D. J. & Lowe, C. F. (in press). Modifying delusions: the role of empirical testing. Behavior Therapy. DeLisi, L. E. (1990). Depression in schizophrenia. Washington, DC APA. Fowler, D. & Morley, S. (1989). The cognitive behavioural treatment of hallucinations and delusions: a preliminary study.
Behavioural Psychotherapy, 17, 267-282. Garety, P. A. (1985). Delusions: problems in definition and measurement. Brirish Journa/ qfMedica1 Psychology, 58, 25-34.
Cognitive approach to delusions 367
Garety, P. A. (1991). Reasoning and delusions. British Journal of Psychiatry, 159(suppl.), 14--18. Hartman, L. M. & Cashman, F. E. (1983). Cognitive behavioural and psychopharmacological treatment of delusional
symptoms: a preliminary report. Behavioural Psychotherapy, I I, 506 1. Hole, R. W., Rush, A. J. & Beck, A. T, (1973), A cognitive investigation of schizophrenic delusions. Psychiatry, 42, 312-319. Hollon, S. D., Shelton, R. C. & Loosen, P. T, (1991). Cognitive therapy and pharmacotherapy for depression. Journal of
Consulting and Clinical Psychology, 59, 88-99. Horne, P. J. & Lowe, C. F. (1993). Determinants of human performance on concurrent schedules, Journal of the
Experimenral Analysis of Behaviour, 59, 2960. Huq, S. F., Garety, P. A. & Hemsley, D. R. (1988). Probabalistic judgements in deluded and non-deluded subjects. Quarterly
Journal of Experimental Psychology, 4OA, 801L812. Johnson, W. G., Ross, J. M. & Mastria, M. A. (1977). Delusional behaviour: an attributional analysis, Journal of Abnormal
Psychology, 86, 421426. Kazdin, A. E. (1982). Single case research designs. New York: Oxford University Press. Kendler, K. S., Glazer, W. M. & Morgenstern, H. (1983). Dimensions of delusional experience. American Journal of
Psychiatry, 140, 466469. Lowe, C. F. (1979). Determinants of human operant behaviour. In Zeiler, M. & Harzem, P. (Eds), Advances in analysis
of behaviour: Vol. I, Reinforcement and organization of behaviour. Chichester, U.K. Wiley. Lowe, C. F. (1983). Radical behaviourism and human psychology. In Davey D. L. (Ed.), Animal models of human hehaviour
(pp. 71-94). Chichester, U.K.: Wiley. Lowe, C. F. & Chadwick, P. D. J. (1990). Verbal control of delusions. Behaviour Therapy, 21, 461479. Lowe, C. F., Home, P. J. & Higson, P. J. (1987). Operant conditioning: the hiatus between theory and practice in clinical
psychology. In Eysenck, H. J. & Martin, 1. (Eds), Theoreficalfoundarions ofbehaviour therapy (pp. 153-165). New York: Plenum Press.
Luria, A. (1961). The role of speech in ihe regulation of normal and abnormal behaviour. New York: Liveright. Maher. B. A. (1974). Delusional thinking and perceptual disorder. Journal of Individual Psychology, 30, 98-113. Maher, B. A. (i988): Anomalous experience and-delusional thinking: the logic of explanationIn Oltmanns, T. F. & Maher,
B. A. (Eds), Delusional beliefs (pp. 15-33). New York: Wiley. Milton, F., Patwa, K. & Hafner, R. J. (1978). Confrontation vs. belief modification in persistently deluded patients. Brifish
Journal of Medical Psychology, 51, 127-130. Muran, J. C. (1991). A reformulation of the ABC model in cognitive psychotherapies. Clinical P.sychology Review. If,
399418. Popper, K. R. (1977). On hypotheses. In Johnson-Laird, P. N. & Wason, P. C. (Eds), Thinking: readings in cognitive science
(pp. 264-273). Bath, U.K.: Cambridge University Press. Strauss, J. S. (1969). Hallucinations and delusions as points on continua functions, Archives of General Psychiarry, 21,
581-586. Strauss, J. S. (1991). The person with delusions. British Journal of Psychiatry, 159(suppl.), 5762. Trower, P., Casey, A. & Dryden, W. (1988). Cognitive behavioural counselling in action. Bristol, U.K.: Sage. Vygotsky, L. S. (1962). Thought and language. Cambridge, MA: MIT Press. Watts, F. N., Powell, E. G. & Austin, S. V. (1973). The modification of abnormal beliefs. British Journal of Medical
Psychology, 46, 3599363. Zigler, & Click, (1988). Is paranoid schizophrenia really camouflaged depression?. American Psychologist, 43, 284-290.
