- Email: [email protected]
Effects of neutralizing on intrusive thoughts: An experiment investigating the etiology of obsessive-compulsive disorder
Behm. Res. Thu.
Pergamon SOOO5-7%7(%)00112-X
Vol. 35. No. 3, pp. 21 I-219. 1997
0 1997 Elsevier Science Ltd Printed in Great Britain. All rights reserved
0005-7967/97$17.00+ 0.00
EFFECTS OF NEUTRALIZING ON INTRUSIVE THOUGHTS: AN EXPERIMENT INVESTIGATING THE ETIOLOGY OF OBSESSIVE-COMPULSIVE DISORDER PAUL M. SALKOVSKIS’, DAVID WESTBROOK*, JULIA DAVIS), ANNE JEAVONS* and ANN GLEDHILL4 ‘Department of Psychiatry, University of Oxford, Warneford Hospital, *Department of Clinical Psychology, Warneford Hospital, Oxford OX3 7JX, U.K., “Department of Psychology, UWIST, Card% CF3 7VX, U.K. and 4Department of Experimental Psychology, Oxford University and Corpus Christi College, Oxford, U.K. (Received 10 September 1996)
Summary-A large sample of non-clinical subjects were screened and those who reported experiencing relatively frequent intrusive thoughts with associated neutralizing were selected. These subjects were randomly allocated to one of two conditions: both groups listened to repeated recorded presentations of one of their intrusive thoughts and were then required either to (a) neutralize it, or (b) distract themselves for a similar period. Ratings of discomfort were taken during this procedure (first phase), and during identical presentations of the same thought without neutralizing or distracting (second phase). Results showed that the group who neutralized during the first phase experienced significantly more discomfort during the second phase and significantly stronger urges to neutralize and distract. There was also evidence that engaging in neutralizing responses during the first phase made it difficult to stop neutralizing during the second phase. The results are considered in the context of the cognitive-behavioural hypothesis that obsessional disorders develop as a consequence of neutralizing normal intrusive thoughts. 0 1997 Elsevier Science Ltd. All rights reserved
INTRODUCTION
Two types of theory have been proposed to account for obsessional disorders. Cognitive deficit theories (e.g. Reed, 1985) are based on the assumption that obsessional symptoms represent the expression of a generalized and involuntary disturbance of information processing. Other features, such as obsessional ritualizing, are regarded as secondary to the primary cognitive deficit. In contrast to the deficit view, behavioural and cognitive-behavioural theories (e.g. Salkovskis, 1985, 1989b, 1996c; Rachman & Hodgson, 1980) are based on the hypothesis that obsessional problems result from an exaggeration of the normal processes governing the specific interaction between thoughts, affective responses and behaviour. No generalized cognitive deficit is required within this framework; indeed, there are a range of reasons to doubt both the utility and the validity of psychological, neuropsychological and biological theories which attribute obsessional problems to generalized deficits of lesions (Salkovskis, 1996c). The cognitive-behavioural view proposes that in obsessional problems, the development of a learned and voluntary response (i.e. neutralizing or compulsive behaviour) is crucial to the persistence of intrusive thoughts which would otherwise be neither troublesome nor noticeably persistent. This emphasis on the role of neutralizing is an important common feature shared by the behavioural and cognitive-behavioural theories. One of the strengths of the behavioural approach is that it has led to the development of exposure and response prevention, which is currently the best established psychological treatment for obsessive-compulsive disorder (OCD) (Salkovskis & Kirk, 1989). However, exposure-based treatments have important limitations (Salkovskis, 1989a), particularly amongst patients with severe concurrent depression, patients in whom overvalued ideation is present, and in the high proportion of patients who refuse to comply with the rigorous requirements of strict exposure treatments (Marks, 1987). A better understanding of the mechanism by which exposure brings about fear reduction is needed for more effective and efficient therapeutic interventions to be 211
212
Paul M. Salkovskis et al.
developed (Salkovskis, 1996a). Recent cognitive accounts promise such understanding by suggesting that the effects of exposure are mediated by belief change, and that interventions can be tailored to producing maximal belief change (Salkovskis, 1996b). A clearer theoretical and clinical understanding of the cognitive mechanisms linking obsessional thinking and compulsive behaviour is a prerequisite if such an approach is to be successful in OCD. This link is crucial to the cognitive-behavioural hypothesis of the acquisition and maintenance of obsessional disorders which also incorporates and extends previous behavioural views (Salkovskis, 1985, 1989b, 1996~). This theory proposes that three factors interact in the development of obsessional problems. The first factor is the spontaneous occurrence of intrusive cognitions (thoughts, impulses or images); such intrusions are normal, and indeed almost universal (Salkovskis & Harrison, 1984; Rachman & de Silva, 1978). These intrusions form the basis for the subsequent development of obsessional problems in vulnerable individuals and that the transition to obsessional disorder will only happen if two other closely inter-related factors are present. The first of these occurs when, due to pre-existing attitudes and beliefs, sufferers interpret the occurrence and/or content of particular intrusions as personally relevant in a way which make them responsible for taking some further action to prevent harm to themselves or others. Thus, obsessions develop from intrusive thoughts which are appraised as both threatening and as indicating responsibility, making preventative action imperative. In terms of Beck’s (Beck, 1976) model of emotional disorders, intrusive cognitions result in ‘negative automatic thoughts’ concerning responsibility and hence induce unpleasant emotions. The other crucial factor is that, as a consequence of this perception of an obligation to react (‘responsibility’), the sufferer attempts to dissipate or remove such responsibility through efforts to neutralize the intrusion or effects (by motor or cognitive rituals, avoidance of provoking stimuli, seeking reassurance, or attempts to suppress the intrusive thoughts). ‘Responsibility’ is used in this context to mean that the person believes that he or she may be, or come to be, the cause of harm (to self or others) unless he or she takes some preventative or restorative action. A group of researchers working on obsessions recently defined the ‘responsibility’ appraisals characteristic of obsessional problems as: The belief that one has power which is pivotal to bring, about or prevent subjectively crucial negative outcomes. These outcomes are perceived as essential to prevent. They may be actual, that is, having consequences in the real world, and/or at a moral level (Salkovskis et al., 1996).
The importance of the appraisal of intrusive thoughts as having implications for responsibility for harm to self or others therefore lies in the way that such appraisals link the occurrence and content of intrusive thought with both distress and the urge to engage in neutralizing behaviour. The cognitive hypothesis goes on to predict that this development of neutralizing (intended by the person to reduce threat arising from the idea of being responsible for preventable harm) will increase both the frequency of intrusive thoughts and the associated discomfort. It is also predicted that, in the longer term, responsibility-driven neutralizing will tend to increase perceived responsibility connected with particular thoughts and situations. This occurs because seeking to prevent oneself from causing harm carries an implicit acceptance of a duty to prevent such harm in the first place (and hence further inflates the sense of responsibility). These factors will tend to result in progressive intensification of the effects of obsessional thinking and neutralizing on the sufferer. In some instances, the occurrence of neutralizing can have the further effect of preventing the experience of disconfirmation of the patient’s fears (Salkovskis, 1991, 1996b). That is, the patients’ belief in their initial negative appraisal and the effectiveness of their neutralizing behaviour may often be maintained by the non-occurrence of the feared harm, which appears to have been prevented by the action taken, the value of which is thus ‘confirmed’. This model thus distinguishes between: (i) the initial occurrence of intrusive thoughts (which is said to be an automatic process); (ii) the way these thoughts are appraised or interpreted which is also automatic; and (iii) the resulting voluntary and strategic responses to the intrusions. Although there is a great deal of evidence for the value of preventing neutralizing when treating OCD (Salkovskis & Kirk, 1989, 1997), data on the proposed role of neutralizing activity in the development of obsessional problems is not currently available. Clinically, neutralizing is most
Effects of neutralizing on intrusive thoughts
213
obviously observed as overt compulsive behaviour, but it also often takes the form of cognitive neutralizing, including both attempts at restitution (‘putting right’) and mental checking (Salkovskis & Westbrook, 1989). Other neutralizing responses include avoidance of provoking stimuli, reassurance-seeking, and effort directed towards suppressing the intrusions. There is some recent evidence that direct attempts to suppress unwanted thoughts have a paradoxical effect, increasing the frequency of intrusions suppressed in non-clinical Ss. This effect is particularly evident in studies which used personally relevant unacceptable intrusive thoughts (Wegner, Schneider, Carter & White, 1987; Trinder & Salkovskis, 1994; Salkovskis & Campbell, 1994; Lavy & van den Hout, 1990). These data are consistent with the cognitive hypothesis of the importance of safety seeking behaviours, and are consistent with the proposal that motivated thought suppression may be important in the development and maintenance of obsessional problems. However, the hypothesized effects of neutralizing on intrusive cognitions has not yet been experimentally demonstrated. The present experiment was therefore designed to examine the effects of deliberately increasing (cognitive) neutralizing on the discomfort associated with naturally occurring and unacceptable intrusive thoughts in non-clinical Ss. METHOD
Overview
Naturally-occurring intrusive thoughts and associated neutralizing thoughts were elicited from non-clinical Ss, who were selected from a larger population because they reported a pre-existing tendency to neutralize strongly negative and relatively frequent intrusive thoughts. Whilst listening to repeated audiotaped presentations of their individual intrusions, Ss were asked either to neutralize or to distract for a comparable period. Discomfort associated with the intrusion was assessed at intervals throughout the procedure. After a pause, the tape was played again, but this time Ss in both groups were required to listen without any deliberate response other than further ratings of discomfort. Subjects
Non-clinical Ss were recruited from people attending a library facility, a general nursing course and evening classes at a college of further education. A total of 1370 Ss were screened for intrusive thoughts using a shortened version of the questionnaire used by Salkovskis and Harrison (1984). Subjects were selected if they reported that they had experienced 10 or more intrusive thoughts in the preceding week; if those thoughts caused discomfort with a rating of 30 or more on a loo-point scale; if they reported that they cognitively neutralized such intrusions ‘often’ or ‘always’; and if they were willing to participate in a psychology experiment. Subjects had to be in the age range 17-60. All Ss who fulfilled these criteria were invited for interview, but they were excluded from the experiment if they reported a past history of psychiatric disorder or treatment (4 Ss). Any Ss who usually neutralized by counting in any form were also eliminated from the experiment (1 S). Twelve Ss did not complete the experimental procedures satisfactorily, and were eliminated from the subsequent analysis (see below). The characteristics of the Ss in the two groups who satisfactorily completed the experiment (13 in the experimental condition, 15 in the control condition) are shown in Table 1. Subjects were randomly allocated to the experimental or control group prior to arrival for the experiment, using sampling without replacement. A running mean was kept of pre-experimental discomfort ratings on the screening questionnaire, and the final 3 Ss in the study were allocated on the basis of this measure to ensure that the initial discomfort levels associated with intrusions was equivalent between groups. No other information about these Ss was available to the experimenter at the time of these allocations. Procedure Pre-experimental. On arrival, Ss were told that the purpose of the experiment was to investigate intrusive, repetitive, unwanted and upsetting thoughts and images, and that the experiment would be fully explained at the end of the session. The first stage of the experiment involved an interview
214
Paul M. Salkovskis et al.
lasting between 1 and 14 hr, during which Ss were asked about their intrusive thoughts and completed a range of questionnaires and scales. Scales completed were a version of the Compulsive Activity Checklist, an assessment of the content of intrusive thoughts, a series of scales on the form of intrusive thoughts, and 8-point ratings of the severity and frequency of intrusive thoughts and associated impairment. (The data from this part of the experiment are reported separately.) Subjects were trained in the use of a loo-point visual analog scale rating of discomfort. Experiment. Towards the end of the interview, Ss were asked to describe in detail their most common unpleasant intrusive thought which was frequently associated with cognitive neutralizing, together with details of the neutralizing thought; both thoughts were noted verbatim. The intrusive thought was written on a sheet by the experimenter and then, when necessary, modified by the S so as to last for approximately 20 set when spoken out loud, whilst retaining its important content. The S was asked to record the written version of the intrusion verbatim onto a 60-set audiotape loop. A rating of discomfort evoked by the recorded version of the thought was then taken. The Ss’ normal neutralizing response (timed to take approximately 20 set) was prepared in the same way as had been done with the intrusion. Subjects in the experimental condition were told: What I’m going to do is ask you to listen carefully to the tape over the headphones. Each time you hear the thought, I would like you to immediately think the neutralizing thought. I will be asking for ratings of how uncomfortable the thought made you feel every now and then; when I show you the card, do the neutralizing then please just tell me the number which best described how you felt during the time you just heard the thought.
Subjects in the control condition were told: What I’m going to do is ask you to listen carefully to the tape over the headphones. Each time you hear the thought, I would like you to immediately count backwards mentally in the way we agreed. I will be asking for ratings of how uncomfortable the thought made you feel every now and then; when I show you the card, count backwards, then please just tell me the number which best described how you felt during the time you just heard the thought.
Once these instructions had been given, the experimenter ensured that Ss had fully understood what was required of them, repeating and clarifying instructions as necessary. When the experimenter was satisfied that Ss understood both the required response to the intrusion and the way ratings were to be done, they were asked to put on the headphones and the tape was started. Subjects heard 16 presentations of their intrusion in all. On presentations 1, 4, 8, 12 and 16, the Ss were shown the card with the VAS for discomfort, and ratings were recorded. The card was presented 20 set after the end of the intrusion, i.e. in the period just after neutralizing had occurred. The card was marked with the words “During the time you just heard the thought” at the top. After 16 presentations, the headphones were removed and Ss completed the post-tape questionnaire. At this stage, all Ss were asked how many times they had neutralized or counted as the thought occurred on the tape. Any Ss who reported complying with instructions on less than 12/16 (75%) of the presentations were excluded from the final analysis of the experimental data. During the 30-min gap before the next phase of the experiment, Ss completed the MOCI, BAI and BDI. After a full 30 min, all Ss were told: I’m going to ask you to listen to the same tape; this time, just listen carefully to the tape. In the same way as before, when I show you the card please just tell me the rating which best describes how you felt during the time you just heard the thought. We don’t want you to (neutralize/count) during this part of the experiment; just listen to the tape.
The experimenter ensured that the S fully understood the requirements of this phase, and the tape began as soon as the headphones were in place. Ratings of discomfort were again recorded at presentations 1, 4, 8, 12 and 16, with the same timing as in the previous phase. Once the taped sequence was over, Ss were asked whether they had counted or neutralized during it; any S doing so on more than 3/16 (18.75%) of occasions was excluded from the final analysis of this part of the experiment. At this stage, Ss also completed the post-tape questionnaire. Subjects were finally asked if they had any ideas about the purpose of the experiment or the expectations of the experimenter, and debriefed.
Effectsof neutralizing on intrusive thoughts
215
Measures
A modified version of the questionnaire used by Salkovskis and Dent (in preparation), 0 was used in the initial screening of Ss. The modifications involved deleting a number of items for brevity. Discomfort ratings during the experiment itself were based on a O-100 visual analog scale, marked with ‘No discomfort at all’ at one end, and ‘The most uncomfortable I have ever felt’ at the other. After the audiotaped presentation, Ss completed a questionnaire in which they made retrospective visual analog scale ratings (referenced to “During the time the tape was playing”) ofi (a) the strength of any urge to neutralize or put right; (b) the strength of any urge to distract themselves from the thoughts; and (c) overall discomfort during that period. Between the two experimental phases, Ss completed the Beck Depression Inventory (BDI: Beck, Ward, Mendelsohn, Mock & Erbaugh, 1961), the Maudsley Obsessive-Compulsive Inventory (MOCI: Hodgson & Rachman, 1978) and the Beck Anxiety Inventory (BAI: Beck, Epstein & Brown, 1988). Statistical analysis
Discomfort ratings taken during the audiotaped presentations were analyzed using a repeated measures analysis of covariance (ANCOVA) with trend analysis; the initial discomfort rating when the thought was first identified was used as covariate. Data from the questionnaire completed after the second phase were analysed using one way analysis of variance
RESULTS Group characteristics
The two groups did not differ [F,,,Z,,< 1 in each instance], in terms of measures of the discomfort evoked by intrusions prior to the experiment, clinical questionnaires, age, or any other characteristic measured prior to the experimental manipulation (see Table 1). Discomfort ratings
Mean discomfort ratings on first elicitation of the intrusion and during the taped presentations are shown in Fig. 1. Analysis of covariance using initial level of discomfort as covariate showed a significant main effect of group [neutralizing vs counting; F,,,*,, = 5.80, P < 0.0251 and a significant main effect of phase [response period vs listening period; F&) = 12.56, P < 0.005]. These main effects are modified by a significant phase x group interaction [Fc,,26j= 4.40, P < 0,051 and the crucial group x phase x time interaction [Ft4,,04j = 3.40, P < 0.025, Greenhouse-Geisser P < 0.051. Post hoc r-tests using the Bonferroni correction indicate that the last three points of the second phase are significantly different between groups; within-group planned comparisons show that, during phase 1, the first and last ratings were not significantly different for the counting group but were different for the neutralizing group, whilst in the second phase the first and last points are significantly different in both groups, in different directions. Orthogonal decomposition indicates that there was a significant group x phase linear interaction. These data indicate that the two groups show different trends during the second phase of the experiment.
Table Variable Sex ratio (m:r) Age BAI BDI MOCI: Total Doubting Slowness Washing Checking
I. Characteristics of the two groutx Counting n = I.5 Mean (SD)
Neutralizing n = 13 Mean (SD) 5.8
6:9 22.1 13.9 9.0
(8.55) (7.6) (7.9)
22.0 16.8 10.8
(5.3) (7.3) (8.0)
7.3 3.2 2.3 1.9 2.1
(5.4) (1.7) (0.9) (1.9) (2.4)
8.1 3.4 2.4 I .9 2.8
(3.9) (1.5) (1.0) (1.3) (2.4)
Paul M. Salkovskis et al.
216 60
50
e
/ ,/2 ‘S 40 F? t P E S .z 30 n
/
9 \
20
-*-
neutralise (n = 13)
- --n -
count (n = 15)
b-q \ ‘b
10
I
I
I
I
I
I
I
I
I
I
J
bl
fl
f4
f8
f12
f16
sl
s4
58
s12
~16
rest
phase 1: “respond”
phase 2: “listen”
Fig. 1. Discomfort ratings on O-100 scale taken before tape (bl) and during 16 presentations of the thought during which the subject responds (flLfl6) and 16 subsequent presentations which are not responded to (sl-~16).
Post-tape questionnaire
Group means for the variables measured at the end of the second phase are shown in Table 2. Consistent with the discomfort ratings for each presentation, scores for overall discomfort ratings as measured at the end of the second taped sequence were significantly different between groups PW5, = 7.81, P < 0.011. There were also significant differences between groups during the second phase for ratings of the urge to neutralize [F,,,,, = 14.09, P < O,OOl] and the urge to distract [F,,,,, = 4.88, P < 0.051. Behavioural tendency to neutralize
The tendency to neutralize can be measured both by rating and by the degree to which neutralizing was actually carried out. In the second phase, it was required that Ss not neutralize. Those Ss who did neutralize in the second phase were not included in the main analysis; however, such a response can be regarded as indicating a particularly strong tendency to neutralize with a clear behavioural consequence. A number of Ss failed to meet experimental criteria during the second phase because they continued to either neutralize or count, despite instructions to the contrary. Nine Ss from the experimental group exceeded criteria for neutralizing during this phase, while 3 Ss were excluded from the control condition (2 because they continued to count, 1 because of neutralizing). A Fisher exact test indicated that the experimental condition was significantly associated with neutralizing during the second phase (P = 0.025, two tailed). These results may indicate that the Ss carrying out their naturally occurring neutralizing strengthened the behavioural tendency to neutralize. Table 2. Questionnaire
results for the two groups completed presentation of the intrusive thought
after the second
Variable
Counting n = 15 Mean (SD)
Neutralizing n = 13 Mean (SD)
Overall discomfort Urge to neutralize Urge to distract
26.00 22.70 32.70
46.54 55.80 55.80
(17.7) (18.7) (27.1)
(21.3) (27.7) (28. I)
Effects of neutralizing on intrusive thoughts
211
DISCUSSION
The results of this study indicate that when non-clinical Ss were asked to neutralize intrusive thoughts, more discomfort was associated with subsequent presentations of the same thought than in a comparable group of Ss asked to use a distraction strategy. This difference in discomfort was also accompanied by a significantly greater urge to neutralize and distract during the second phase, and by an increased rate of actual neutralizing (contrary to experimental instructions). This latter finding suggests that neutralizing activity may be associated with a self-perpetuating mechanism; in obsessional patients, such a mechanism might be a key factor in the maintenance of obsessional behaviours, both overt and covert. The demonstration that increasing neutralizing may increase both discomfort and the urge to neutralize is an important complement to data from response-prevention treatment in clinical cases (Foa & Steketee, 1979), which shows that preventing neutralizing results in a longer-term decline in both factors. Unexpectedly, the overall mean rating of discomfort during the time Ss were neutralizing or counting was (non-significantly) higher for the neutralizing group. Additionally, only the neutralizing group showed a significant decrease in discomfort over repeated presentations during this phase, from its initial relatively raised level to levels similar to those observed in the control condition. Previous work has suggested that neutralizing in clinical Ss is accompanied by a rapid decrease in discomfort each time ritualizing occurred (Rachman, de Silva & Roper, 1976). It might be that, in the present experiment, such substantial decreases were occurring each time Ss neutralized, but were not picked up because of the timing of our repeated ratings: the rating was made for discomfort “during the time you just heard the thought” and neutralizing occurred after that point. Thus we could be picking up the peaks in a sawtooth pattern of discomfort. However, the overall discomfort ratings for this phase are higher for the neutralizing group, suggesting that if such decreases are occurring, they are neither large nor persistent. Direct evaluation of this point is required. Alternatively, neutralizing may have an immediate generalized anxiety elevating effect which is countered by a subsequent decline in discomfort with repeated occurrence, as observed in the present experiment. A decline in discomfort such as that associated with repeated neutralizing in the first phase of the present experiment may be perceived by the S as evidence for the effectiveness of such a strategy and hence increase the likelihood of subsequent neutralizing. The person would not be aware of the fact that, overall, the strategy was having the effect of increasing their discomfort and anxiety. If this is so, it is not entirely unexpected; safety seeking behaviour generally follows this type of pattern, so that the person believes that a particular strategy (such as avoidance) is helpful whilst in fact a more enduring reduction in anxiety would accompany reduced avoidance. The present experiment relies heavily on self-report both for assessment of the manipulation and for measurement of the dependent variables. However, the nature of the phenomena studied in obsessional thinking tends to dictate the use of self-report, and it is not necessarily a problem to do so. Psychophysiological measures may be useful in future work (Eves & Tata, 1989) and, if overt neutralizing were used instead of cognitive neutralizing, observer ratings of actual neutralizing would be possible. This latter possibility would probably be most appropriate in clinical Ss. More problematic in the present experiment was the fact that a number of Ss had to be eliminated from the analyses because they failed to complete the experiment properly; it is difficult to see how this problem could be overcome. These Ss did not significantly differ from those who were included on any of the variables measured, nor was there any sign of any tendency to differ except on the urge to neutralize, as would be expected. The experiment demonstrates that increasing the rate of Ss’ previously established neutralizing responses may have a large and relatively enduring impact on discomfort and can also increase the urge to continue such neutralizing. It is important to the cognitive-behavioural hypothesis to know whether a logically consistent neutralizing response which has not previously been used by the S would have the same effects, since this would clarify the parameters of the hypothesis proposed for the initial development of obsessional thoughts. It could be predicted that similar results would be obtained where there was a match between the perceived safety-enhancing effects of the chosen neutralizing behaviour and the type of threat perceived to be associated with the
218
Paul M.
Salkovskiser al.
target intrusive thoughts. It is also not clear whether the results observed here are confined to those Ss who neutralize frequently, as in the sample used here, or whether this is a more generalized effect. As a further point, work on positive intrusions (Edwards & Dickerson, 1987) suggest that similar effects to those obtained in the present experiment might be obtained with positive thoughts and positive affect; that is, the mechanism could be a result of salience and not specific to negatively toned material. However, it is important to note that Salkovskis (1985) and Salkovskis, Richards and Forrester (1995) have emphasized that both the occurrence and the content of intrusions contribute to the negative appraisal. Thus, an apparently positive intrusive thought which inappropriately (to the person) occurs during a solemn or sad occasion may result in discomfort, self-criticism, attempts to suppress and neutralize and so on. Again, the cognitive-behavioural hypothesis of OCD suggests that appraisal in terms of responsibility for harm is crucial. Other factors may also have mediated the differences observed in discomfort in the present experiment; for example, neutralizing may have had an impact on vividness, evaluative dimensions (e.g. acceptability), intrusiveness and so on. Future experiments might attempt to evaluate these possible mechanisms, perhaps particularly focusing on evaluative mechanisms (i.e. the way in which the intrusion is appraised; Salkovskis, 1989b, 1996c; Salkovskis et al., 1995), since there is already some evidence of an association between attitudes of responsibility and the occurrence of neutralizing. The experiment reported here demonstrated the way in which neutralizing by non-clinical Ss can increase the discomfort subsequently associated with naturally occurring negative intrusive cognitions. The apparent similarity to obsessional phenomena is compelling, and is consistent with predictions arising from cognitive-behavioural hypothesis. More specifically, the data provide support for the hypothesis that a key factor in obsessional disorders is the occurrence of increased threat-focused neutralizing behaviour in response to normal intrusions, rather than as a result of generalized cognitive deficits. Cognitive-behavioural theories of the maintenance of anxiety disorders have previously stressed the importance of such safety seeking behaviours (Salkovskis, 1991, 1996b). The present experiment suggests that such behaviours may also be involved in the etiology of obsessional problems. Future work may need to consider the importance of responsibility appraisals in the development of neutralizing behaviours in obsessional problems. If this can be established, then each of the key components of the cognitive-behavioural hypothesis will have been empirically demonstrated. Similar work is also needed on the related phenomenon of deliberate thought suppression (Wegner et al., 1987; Wegner, 1989; Trinder & Salkovskis, 1994; Salkovskis & Campbell, 1994), which almost certainly shares many of the characteristics of neutralizing as a coping strategy. It has also been suggested that the basis for the priority of processing of obsessional thoughts may lie in the interaction between automatic vs controlled processing (in this instance, intrusive thoughts vs attempted coping actions) and that the link between the two can be found in the motivating properties of appraisal of the originating intrusive thoughts (Salkovskis et al., 1995). Future investigations could thus focus on factors which may be involved in the origin and maintenance of obsessional problems, including neutralizing, thought suppression, adverse mood and beliefs concerning responsibility and interactions between these. Acknowledgements-Paul Salkovskis is Wellcome Trust Senior Research Fellow. Rowena Cook of Oxford College of Further Education provided invaluable assistance in the recruitment of Ss for this experiment. Martina Mueller and Hilary Warwick gave helpful comments on an earlier version of this manuscript, and Pavlos Anastasiades gave invaluable statistical advice.
REFERENCES Beck, A. T. (1976). Cog&ire therapy and the emotionnl disorders. New York: International Universities Press. Beck, A. T., Epstein, N., & Brown, G. (1988). An inventory for measuring clinical anxiety. Journal of Consulting, and Clinical Psychology, 56, 893-897.
Beck, A. T., Ward, C. H., Mendelsohn, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 18, 561-571.
Edwards, S., & Dickerson, M. (1987). On the similarity of positive and negative intrusions. Behoviour Research and Therapy, 25, 207-211.
Eves, F., & Tata, P. (1989). Phasic cardiac and electrodermal reactions to idiographic stimuli in obsessional subjects. Behavioural Psvchotherapy, 17, 71-82.
Effects of neutralizing on intrusive thoughts
219
Foa, E. B., & Steketee, G. S. (1979). Obsessive-compulsives: conceptual issues and treatment interventions. Progress in Behavior ModiJication, 8, l-53. Hodgson, R. J., & Rachman, S. J. (1978). Obsessional-compulsive complaints. Behaviour Research and Therapy, 15, 385-389.
Lavy, E., & van den Hout, M. (1990). Thought suppression induces intrusion. Behaviouraf Psychotherapy, 18, 251-258. Marks, 1. M. (1987). Fears, phobias and rituals. New York: Oxford University Press. Rachman. S. J., & Hodgson, R. L. (1980). Obsessions and compulsions. Englewood Cliffs, NJ: Prentice Hall. Rachman, S. J., & de Silva, P. (1978). Abnormal and normal obsessions. Behaviour Research and Therapy, 16, 233-248. Rachman, S. J., de Silva, P., & Roper, G. (1976). The spontaneous decay of compulsive urges. Behaviour Research and Therapy, 14, 445453.
Reed, G. F. (1985). Obsessional experience and compulsive behaviour: A cognitive-structural approach. London: Academic Press. Salkovskis, P. M. (1985). Obsessional-compulsive problems: A cognitive-behavioural analysis. Behaoiour Research and Therapy, 23, 571-583.
Salkovskis, P. M. (1989a). Obsessions and compulsions. In J. Scott, J. M. G. Williams & A. T. Beck (Eds.), Cognitive therapy: A clinical casebook. London: Croom Helm. Salkovskis, P. M. (1989b). Cognitive-behavioural factors and the persistence of intrusive thoughts in obsessional problems, Behaviour Research and Therapy, 27, 677-682.
Salkovskis, P. M. (1991). The importance of behaviour in the maintenance of anxiety and panic: A cognitive account. Behavioural Psychotherapy,
19, 619.
P. M. (1996a). Resolving the cognition-behaviour debate. In P. M. Salkovskis (Ed.), Trends in cognitive-behaviour therapy. Chichester: Wiley. Salkovskis. P. M. (1996b). The cognitive approach to anxiety: Threat beliefs, safety seeking behaviour, and the special case of health anxiety and obsessions. In P. M. Salkovskis (Ed.), Frontiers of cognitive therapy (pp. 48-74). New York: Guilford. Salkovskis, P. M. (1996~). Cognitive-behavioural approaches to the understanding of obsessional problems. In R. Rapee (Ed.), Current controversies in the anxiety disorders. New York: Guilford. Salkovskis, P. M., &Campbell, P. (1994). Thought suppression in naturally occurring negative intrusive thoughts. Behaviour Salkovskis,
Research and Therapy, 32, 1-8.
Salkovskis, P. M., & Dent, H. (in preparation). Intrusive thoughts and negative beliefs. Salkovskis, P. M., & Harrison, J. (1984). Abnormal and normal obsessions-a replication. Behauiour Research and Therapy, 22, 5499552.
Salkovskis, P. M., & Kirk, J. (1989). Obsessional disorders. In K. Hawton, P. M. Salkovskis, J. Kirk & D. M. Clark (Eds.), Cognitive-behavioural treatment for psychiatric disorders: A practical guide. Oxford: Oxford University Press. Salkovskis. P. M., & Kirk, J. (1997). Obsessivecompulsive disorder. In D. M. Clark & C. G. Fairburn (Eds.), Science and practice of cognitive behaviour therapy. Oxford: Oxford University Press. Salkovskis, P. M., Rachman, S. L., Ladouceur, R., Freeston, M., Taylor, S., Kyrios, M., & Sica, C. (1996). Defining responsibility in obsessional problems: Proceedings of the Smith College Women’s Room-after
the Toronto Cafeteria.
Salkovskis, P. M., Richards, C., & Forrester, E. (1995). The relationship between obsessional problems and intrusive thoughts. Behavioural and Cognitive Psychotherapy, 23, 281-299. Salkovskis. P. M., & Westbrook, D. (1989). Behaviour therapy and obsessional rummations: Can failure be turned into success?. Behaviour Research and Therapy, 27, 149-160. Trinder. H., & Salkovskis, P. M. (1994). Personally relevant intrusions outside the laboratory: Long-term suppression increases intrusion. Behaviour Research and Therapy, 32, 833-842. Wegner. D. M. (1989). White bears and other um~anted thoughts: Suppression, obsession and thepsychology of mental control. New York: Viking. Wegner, D. M., Schneider, D. J., Carter, S. R., & White, T. L. (1987). Paradoxical effects of thought suppression. Journal of Personality and Social Psychology, 53, S-13.
Pergamon SOOO5-7%7(%)00112-X
Vol. 35. No. 3, pp. 21 I-219. 1997
0 1997 Elsevier Science Ltd Printed in Great Britain. All rights reserved
0005-7967/97$17.00+ 0.00
EFFECTS OF NEUTRALIZING ON INTRUSIVE THOUGHTS: AN EXPERIMENT INVESTIGATING THE ETIOLOGY OF OBSESSIVE-COMPULSIVE DISORDER PAUL M. SALKOVSKIS’, DAVID WESTBROOK*, JULIA DAVIS), ANNE JEAVONS* and ANN GLEDHILL4 ‘Department of Psychiatry, University of Oxford, Warneford Hospital, *Department of Clinical Psychology, Warneford Hospital, Oxford OX3 7JX, U.K., “Department of Psychology, UWIST, Card% CF3 7VX, U.K. and 4Department of Experimental Psychology, Oxford University and Corpus Christi College, Oxford, U.K. (Received 10 September 1996)
Summary-A large sample of non-clinical subjects were screened and those who reported experiencing relatively frequent intrusive thoughts with associated neutralizing were selected. These subjects were randomly allocated to one of two conditions: both groups listened to repeated recorded presentations of one of their intrusive thoughts and were then required either to (a) neutralize it, or (b) distract themselves for a similar period. Ratings of discomfort were taken during this procedure (first phase), and during identical presentations of the same thought without neutralizing or distracting (second phase). Results showed that the group who neutralized during the first phase experienced significantly more discomfort during the second phase and significantly stronger urges to neutralize and distract. There was also evidence that engaging in neutralizing responses during the first phase made it difficult to stop neutralizing during the second phase. The results are considered in the context of the cognitive-behavioural hypothesis that obsessional disorders develop as a consequence of neutralizing normal intrusive thoughts. 0 1997 Elsevier Science Ltd. All rights reserved
INTRODUCTION
Two types of theory have been proposed to account for obsessional disorders. Cognitive deficit theories (e.g. Reed, 1985) are based on the assumption that obsessional symptoms represent the expression of a generalized and involuntary disturbance of information processing. Other features, such as obsessional ritualizing, are regarded as secondary to the primary cognitive deficit. In contrast to the deficit view, behavioural and cognitive-behavioural theories (e.g. Salkovskis, 1985, 1989b, 1996c; Rachman & Hodgson, 1980) are based on the hypothesis that obsessional problems result from an exaggeration of the normal processes governing the specific interaction between thoughts, affective responses and behaviour. No generalized cognitive deficit is required within this framework; indeed, there are a range of reasons to doubt both the utility and the validity of psychological, neuropsychological and biological theories which attribute obsessional problems to generalized deficits of lesions (Salkovskis, 1996c). The cognitive-behavioural view proposes that in obsessional problems, the development of a learned and voluntary response (i.e. neutralizing or compulsive behaviour) is crucial to the persistence of intrusive thoughts which would otherwise be neither troublesome nor noticeably persistent. This emphasis on the role of neutralizing is an important common feature shared by the behavioural and cognitive-behavioural theories. One of the strengths of the behavioural approach is that it has led to the development of exposure and response prevention, which is currently the best established psychological treatment for obsessive-compulsive disorder (OCD) (Salkovskis & Kirk, 1989). However, exposure-based treatments have important limitations (Salkovskis, 1989a), particularly amongst patients with severe concurrent depression, patients in whom overvalued ideation is present, and in the high proportion of patients who refuse to comply with the rigorous requirements of strict exposure treatments (Marks, 1987). A better understanding of the mechanism by which exposure brings about fear reduction is needed for more effective and efficient therapeutic interventions to be 211
212
Paul M. Salkovskis et al.
developed (Salkovskis, 1996a). Recent cognitive accounts promise such understanding by suggesting that the effects of exposure are mediated by belief change, and that interventions can be tailored to producing maximal belief change (Salkovskis, 1996b). A clearer theoretical and clinical understanding of the cognitive mechanisms linking obsessional thinking and compulsive behaviour is a prerequisite if such an approach is to be successful in OCD. This link is crucial to the cognitive-behavioural hypothesis of the acquisition and maintenance of obsessional disorders which also incorporates and extends previous behavioural views (Salkovskis, 1985, 1989b, 1996~). This theory proposes that three factors interact in the development of obsessional problems. The first factor is the spontaneous occurrence of intrusive cognitions (thoughts, impulses or images); such intrusions are normal, and indeed almost universal (Salkovskis & Harrison, 1984; Rachman & de Silva, 1978). These intrusions form the basis for the subsequent development of obsessional problems in vulnerable individuals and that the transition to obsessional disorder will only happen if two other closely inter-related factors are present. The first of these occurs when, due to pre-existing attitudes and beliefs, sufferers interpret the occurrence and/or content of particular intrusions as personally relevant in a way which make them responsible for taking some further action to prevent harm to themselves or others. Thus, obsessions develop from intrusive thoughts which are appraised as both threatening and as indicating responsibility, making preventative action imperative. In terms of Beck’s (Beck, 1976) model of emotional disorders, intrusive cognitions result in ‘negative automatic thoughts’ concerning responsibility and hence induce unpleasant emotions. The other crucial factor is that, as a consequence of this perception of an obligation to react (‘responsibility’), the sufferer attempts to dissipate or remove such responsibility through efforts to neutralize the intrusion or effects (by motor or cognitive rituals, avoidance of provoking stimuli, seeking reassurance, or attempts to suppress the intrusive thoughts). ‘Responsibility’ is used in this context to mean that the person believes that he or she may be, or come to be, the cause of harm (to self or others) unless he or she takes some preventative or restorative action. A group of researchers working on obsessions recently defined the ‘responsibility’ appraisals characteristic of obsessional problems as: The belief that one has power which is pivotal to bring, about or prevent subjectively crucial negative outcomes. These outcomes are perceived as essential to prevent. They may be actual, that is, having consequences in the real world, and/or at a moral level (Salkovskis et al., 1996).
The importance of the appraisal of intrusive thoughts as having implications for responsibility for harm to self or others therefore lies in the way that such appraisals link the occurrence and content of intrusive thought with both distress and the urge to engage in neutralizing behaviour. The cognitive hypothesis goes on to predict that this development of neutralizing (intended by the person to reduce threat arising from the idea of being responsible for preventable harm) will increase both the frequency of intrusive thoughts and the associated discomfort. It is also predicted that, in the longer term, responsibility-driven neutralizing will tend to increase perceived responsibility connected with particular thoughts and situations. This occurs because seeking to prevent oneself from causing harm carries an implicit acceptance of a duty to prevent such harm in the first place (and hence further inflates the sense of responsibility). These factors will tend to result in progressive intensification of the effects of obsessional thinking and neutralizing on the sufferer. In some instances, the occurrence of neutralizing can have the further effect of preventing the experience of disconfirmation of the patient’s fears (Salkovskis, 1991, 1996b). That is, the patients’ belief in their initial negative appraisal and the effectiveness of their neutralizing behaviour may often be maintained by the non-occurrence of the feared harm, which appears to have been prevented by the action taken, the value of which is thus ‘confirmed’. This model thus distinguishes between: (i) the initial occurrence of intrusive thoughts (which is said to be an automatic process); (ii) the way these thoughts are appraised or interpreted which is also automatic; and (iii) the resulting voluntary and strategic responses to the intrusions. Although there is a great deal of evidence for the value of preventing neutralizing when treating OCD (Salkovskis & Kirk, 1989, 1997), data on the proposed role of neutralizing activity in the development of obsessional problems is not currently available. Clinically, neutralizing is most
Effects of neutralizing on intrusive thoughts
213
obviously observed as overt compulsive behaviour, but it also often takes the form of cognitive neutralizing, including both attempts at restitution (‘putting right’) and mental checking (Salkovskis & Westbrook, 1989). Other neutralizing responses include avoidance of provoking stimuli, reassurance-seeking, and effort directed towards suppressing the intrusions. There is some recent evidence that direct attempts to suppress unwanted thoughts have a paradoxical effect, increasing the frequency of intrusions suppressed in non-clinical Ss. This effect is particularly evident in studies which used personally relevant unacceptable intrusive thoughts (Wegner, Schneider, Carter & White, 1987; Trinder & Salkovskis, 1994; Salkovskis & Campbell, 1994; Lavy & van den Hout, 1990). These data are consistent with the cognitive hypothesis of the importance of safety seeking behaviours, and are consistent with the proposal that motivated thought suppression may be important in the development and maintenance of obsessional problems. However, the hypothesized effects of neutralizing on intrusive cognitions has not yet been experimentally demonstrated. The present experiment was therefore designed to examine the effects of deliberately increasing (cognitive) neutralizing on the discomfort associated with naturally occurring and unacceptable intrusive thoughts in non-clinical Ss. METHOD
Overview
Naturally-occurring intrusive thoughts and associated neutralizing thoughts were elicited from non-clinical Ss, who were selected from a larger population because they reported a pre-existing tendency to neutralize strongly negative and relatively frequent intrusive thoughts. Whilst listening to repeated audiotaped presentations of their individual intrusions, Ss were asked either to neutralize or to distract for a comparable period. Discomfort associated with the intrusion was assessed at intervals throughout the procedure. After a pause, the tape was played again, but this time Ss in both groups were required to listen without any deliberate response other than further ratings of discomfort. Subjects
Non-clinical Ss were recruited from people attending a library facility, a general nursing course and evening classes at a college of further education. A total of 1370 Ss were screened for intrusive thoughts using a shortened version of the questionnaire used by Salkovskis and Harrison (1984). Subjects were selected if they reported that they had experienced 10 or more intrusive thoughts in the preceding week; if those thoughts caused discomfort with a rating of 30 or more on a loo-point scale; if they reported that they cognitively neutralized such intrusions ‘often’ or ‘always’; and if they were willing to participate in a psychology experiment. Subjects had to be in the age range 17-60. All Ss who fulfilled these criteria were invited for interview, but they were excluded from the experiment if they reported a past history of psychiatric disorder or treatment (4 Ss). Any Ss who usually neutralized by counting in any form were also eliminated from the experiment (1 S). Twelve Ss did not complete the experimental procedures satisfactorily, and were eliminated from the subsequent analysis (see below). The characteristics of the Ss in the two groups who satisfactorily completed the experiment (13 in the experimental condition, 15 in the control condition) are shown in Table 1. Subjects were randomly allocated to the experimental or control group prior to arrival for the experiment, using sampling without replacement. A running mean was kept of pre-experimental discomfort ratings on the screening questionnaire, and the final 3 Ss in the study were allocated on the basis of this measure to ensure that the initial discomfort levels associated with intrusions was equivalent between groups. No other information about these Ss was available to the experimenter at the time of these allocations. Procedure Pre-experimental. On arrival, Ss were told that the purpose of the experiment was to investigate intrusive, repetitive, unwanted and upsetting thoughts and images, and that the experiment would be fully explained at the end of the session. The first stage of the experiment involved an interview
214
Paul M. Salkovskis et al.
lasting between 1 and 14 hr, during which Ss were asked about their intrusive thoughts and completed a range of questionnaires and scales. Scales completed were a version of the Compulsive Activity Checklist, an assessment of the content of intrusive thoughts, a series of scales on the form of intrusive thoughts, and 8-point ratings of the severity and frequency of intrusive thoughts and associated impairment. (The data from this part of the experiment are reported separately.) Subjects were trained in the use of a loo-point visual analog scale rating of discomfort. Experiment. Towards the end of the interview, Ss were asked to describe in detail their most common unpleasant intrusive thought which was frequently associated with cognitive neutralizing, together with details of the neutralizing thought; both thoughts were noted verbatim. The intrusive thought was written on a sheet by the experimenter and then, when necessary, modified by the S so as to last for approximately 20 set when spoken out loud, whilst retaining its important content. The S was asked to record the written version of the intrusion verbatim onto a 60-set audiotape loop. A rating of discomfort evoked by the recorded version of the thought was then taken. The Ss’ normal neutralizing response (timed to take approximately 20 set) was prepared in the same way as had been done with the intrusion. Subjects in the experimental condition were told: What I’m going to do is ask you to listen carefully to the tape over the headphones. Each time you hear the thought, I would like you to immediately think the neutralizing thought. I will be asking for ratings of how uncomfortable the thought made you feel every now and then; when I show you the card, do the neutralizing then please just tell me the number which best described how you felt during the time you just heard the thought.
Subjects in the control condition were told: What I’m going to do is ask you to listen carefully to the tape over the headphones. Each time you hear the thought, I would like you to immediately count backwards mentally in the way we agreed. I will be asking for ratings of how uncomfortable the thought made you feel every now and then; when I show you the card, count backwards, then please just tell me the number which best described how you felt during the time you just heard the thought.
Once these instructions had been given, the experimenter ensured that Ss had fully understood what was required of them, repeating and clarifying instructions as necessary. When the experimenter was satisfied that Ss understood both the required response to the intrusion and the way ratings were to be done, they were asked to put on the headphones and the tape was started. Subjects heard 16 presentations of their intrusion in all. On presentations 1, 4, 8, 12 and 16, the Ss were shown the card with the VAS for discomfort, and ratings were recorded. The card was presented 20 set after the end of the intrusion, i.e. in the period just after neutralizing had occurred. The card was marked with the words “During the time you just heard the thought” at the top. After 16 presentations, the headphones were removed and Ss completed the post-tape questionnaire. At this stage, all Ss were asked how many times they had neutralized or counted as the thought occurred on the tape. Any Ss who reported complying with instructions on less than 12/16 (75%) of the presentations were excluded from the final analysis of the experimental data. During the 30-min gap before the next phase of the experiment, Ss completed the MOCI, BAI and BDI. After a full 30 min, all Ss were told: I’m going to ask you to listen to the same tape; this time, just listen carefully to the tape. In the same way as before, when I show you the card please just tell me the rating which best describes how you felt during the time you just heard the thought. We don’t want you to (neutralize/count) during this part of the experiment; just listen to the tape.
The experimenter ensured that the S fully understood the requirements of this phase, and the tape began as soon as the headphones were in place. Ratings of discomfort were again recorded at presentations 1, 4, 8, 12 and 16, with the same timing as in the previous phase. Once the taped sequence was over, Ss were asked whether they had counted or neutralized during it; any S doing so on more than 3/16 (18.75%) of occasions was excluded from the final analysis of this part of the experiment. At this stage, Ss also completed the post-tape questionnaire. Subjects were finally asked if they had any ideas about the purpose of the experiment or the expectations of the experimenter, and debriefed.
Effectsof neutralizing on intrusive thoughts
215
Measures
A modified version of the questionnaire used by Salkovskis and Dent (in preparation), 0 was used in the initial screening of Ss. The modifications involved deleting a number of items for brevity. Discomfort ratings during the experiment itself were based on a O-100 visual analog scale, marked with ‘No discomfort at all’ at one end, and ‘The most uncomfortable I have ever felt’ at the other. After the audiotaped presentation, Ss completed a questionnaire in which they made retrospective visual analog scale ratings (referenced to “During the time the tape was playing”) ofi (a) the strength of any urge to neutralize or put right; (b) the strength of any urge to distract themselves from the thoughts; and (c) overall discomfort during that period. Between the two experimental phases, Ss completed the Beck Depression Inventory (BDI: Beck, Ward, Mendelsohn, Mock & Erbaugh, 1961), the Maudsley Obsessive-Compulsive Inventory (MOCI: Hodgson & Rachman, 1978) and the Beck Anxiety Inventory (BAI: Beck, Epstein & Brown, 1988). Statistical analysis
Discomfort ratings taken during the audiotaped presentations were analyzed using a repeated measures analysis of covariance (ANCOVA) with trend analysis; the initial discomfort rating when the thought was first identified was used as covariate. Data from the questionnaire completed after the second phase were analysed using one way analysis of variance
RESULTS Group characteristics
The two groups did not differ [F,,,Z,,< 1 in each instance], in terms of measures of the discomfort evoked by intrusions prior to the experiment, clinical questionnaires, age, or any other characteristic measured prior to the experimental manipulation (see Table 1). Discomfort ratings
Mean discomfort ratings on first elicitation of the intrusion and during the taped presentations are shown in Fig. 1. Analysis of covariance using initial level of discomfort as covariate showed a significant main effect of group [neutralizing vs counting; F,,,*,, = 5.80, P < 0.0251 and a significant main effect of phase [response period vs listening period; F&) = 12.56, P < 0.005]. These main effects are modified by a significant phase x group interaction [Fc,,26j= 4.40, P < 0,051 and the crucial group x phase x time interaction [Ft4,,04j = 3.40, P < 0.025, Greenhouse-Geisser P < 0.051. Post hoc r-tests using the Bonferroni correction indicate that the last three points of the second phase are significantly different between groups; within-group planned comparisons show that, during phase 1, the first and last ratings were not significantly different for the counting group but were different for the neutralizing group, whilst in the second phase the first and last points are significantly different in both groups, in different directions. Orthogonal decomposition indicates that there was a significant group x phase linear interaction. These data indicate that the two groups show different trends during the second phase of the experiment.
Table Variable Sex ratio (m:r) Age BAI BDI MOCI: Total Doubting Slowness Washing Checking
I. Characteristics of the two groutx Counting n = I.5 Mean (SD)
Neutralizing n = 13 Mean (SD) 5.8
6:9 22.1 13.9 9.0
(8.55) (7.6) (7.9)
22.0 16.8 10.8
(5.3) (7.3) (8.0)
7.3 3.2 2.3 1.9 2.1
(5.4) (1.7) (0.9) (1.9) (2.4)
8.1 3.4 2.4 I .9 2.8
(3.9) (1.5) (1.0) (1.3) (2.4)
Paul M. Salkovskis et al.
216 60
50
e
/ ,/2 ‘S 40 F? t P E S .z 30 n
/
9 \
20
-*-
neutralise (n = 13)
- --n -
count (n = 15)
b-q \ ‘b
10
I
I
I
I
I
I
I
I
I
I
J
bl
fl
f4
f8
f12
f16
sl
s4
58
s12
~16
rest
phase 1: “respond”
phase 2: “listen”
Fig. 1. Discomfort ratings on O-100 scale taken before tape (bl) and during 16 presentations of the thought during which the subject responds (flLfl6) and 16 subsequent presentations which are not responded to (sl-~16).
Post-tape questionnaire
Group means for the variables measured at the end of the second phase are shown in Table 2. Consistent with the discomfort ratings for each presentation, scores for overall discomfort ratings as measured at the end of the second taped sequence were significantly different between groups PW5, = 7.81, P < 0.011. There were also significant differences between groups during the second phase for ratings of the urge to neutralize [F,,,,, = 14.09, P < O,OOl] and the urge to distract [F,,,,, = 4.88, P < 0.051. Behavioural tendency to neutralize
The tendency to neutralize can be measured both by rating and by the degree to which neutralizing was actually carried out. In the second phase, it was required that Ss not neutralize. Those Ss who did neutralize in the second phase were not included in the main analysis; however, such a response can be regarded as indicating a particularly strong tendency to neutralize with a clear behavioural consequence. A number of Ss failed to meet experimental criteria during the second phase because they continued to either neutralize or count, despite instructions to the contrary. Nine Ss from the experimental group exceeded criteria for neutralizing during this phase, while 3 Ss were excluded from the control condition (2 because they continued to count, 1 because of neutralizing). A Fisher exact test indicated that the experimental condition was significantly associated with neutralizing during the second phase (P = 0.025, two tailed). These results may indicate that the Ss carrying out their naturally occurring neutralizing strengthened the behavioural tendency to neutralize. Table 2. Questionnaire
results for the two groups completed presentation of the intrusive thought
after the second
Variable
Counting n = 15 Mean (SD)
Neutralizing n = 13 Mean (SD)
Overall discomfort Urge to neutralize Urge to distract
26.00 22.70 32.70
46.54 55.80 55.80
(17.7) (18.7) (27.1)
(21.3) (27.7) (28. I)
Effects of neutralizing on intrusive thoughts
211
DISCUSSION
The results of this study indicate that when non-clinical Ss were asked to neutralize intrusive thoughts, more discomfort was associated with subsequent presentations of the same thought than in a comparable group of Ss asked to use a distraction strategy. This difference in discomfort was also accompanied by a significantly greater urge to neutralize and distract during the second phase, and by an increased rate of actual neutralizing (contrary to experimental instructions). This latter finding suggests that neutralizing activity may be associated with a self-perpetuating mechanism; in obsessional patients, such a mechanism might be a key factor in the maintenance of obsessional behaviours, both overt and covert. The demonstration that increasing neutralizing may increase both discomfort and the urge to neutralize is an important complement to data from response-prevention treatment in clinical cases (Foa & Steketee, 1979), which shows that preventing neutralizing results in a longer-term decline in both factors. Unexpectedly, the overall mean rating of discomfort during the time Ss were neutralizing or counting was (non-significantly) higher for the neutralizing group. Additionally, only the neutralizing group showed a significant decrease in discomfort over repeated presentations during this phase, from its initial relatively raised level to levels similar to those observed in the control condition. Previous work has suggested that neutralizing in clinical Ss is accompanied by a rapid decrease in discomfort each time ritualizing occurred (Rachman, de Silva & Roper, 1976). It might be that, in the present experiment, such substantial decreases were occurring each time Ss neutralized, but were not picked up because of the timing of our repeated ratings: the rating was made for discomfort “during the time you just heard the thought” and neutralizing occurred after that point. Thus we could be picking up the peaks in a sawtooth pattern of discomfort. However, the overall discomfort ratings for this phase are higher for the neutralizing group, suggesting that if such decreases are occurring, they are neither large nor persistent. Direct evaluation of this point is required. Alternatively, neutralizing may have an immediate generalized anxiety elevating effect which is countered by a subsequent decline in discomfort with repeated occurrence, as observed in the present experiment. A decline in discomfort such as that associated with repeated neutralizing in the first phase of the present experiment may be perceived by the S as evidence for the effectiveness of such a strategy and hence increase the likelihood of subsequent neutralizing. The person would not be aware of the fact that, overall, the strategy was having the effect of increasing their discomfort and anxiety. If this is so, it is not entirely unexpected; safety seeking behaviour generally follows this type of pattern, so that the person believes that a particular strategy (such as avoidance) is helpful whilst in fact a more enduring reduction in anxiety would accompany reduced avoidance. The present experiment relies heavily on self-report both for assessment of the manipulation and for measurement of the dependent variables. However, the nature of the phenomena studied in obsessional thinking tends to dictate the use of self-report, and it is not necessarily a problem to do so. Psychophysiological measures may be useful in future work (Eves & Tata, 1989) and, if overt neutralizing were used instead of cognitive neutralizing, observer ratings of actual neutralizing would be possible. This latter possibility would probably be most appropriate in clinical Ss. More problematic in the present experiment was the fact that a number of Ss had to be eliminated from the analyses because they failed to complete the experiment properly; it is difficult to see how this problem could be overcome. These Ss did not significantly differ from those who were included on any of the variables measured, nor was there any sign of any tendency to differ except on the urge to neutralize, as would be expected. The experiment demonstrates that increasing the rate of Ss’ previously established neutralizing responses may have a large and relatively enduring impact on discomfort and can also increase the urge to continue such neutralizing. It is important to the cognitive-behavioural hypothesis to know whether a logically consistent neutralizing response which has not previously been used by the S would have the same effects, since this would clarify the parameters of the hypothesis proposed for the initial development of obsessional thoughts. It could be predicted that similar results would be obtained where there was a match between the perceived safety-enhancing effects of the chosen neutralizing behaviour and the type of threat perceived to be associated with the
218
Paul M.
Salkovskiser al.
target intrusive thoughts. It is also not clear whether the results observed here are confined to those Ss who neutralize frequently, as in the sample used here, or whether this is a more generalized effect. As a further point, work on positive intrusions (Edwards & Dickerson, 1987) suggest that similar effects to those obtained in the present experiment might be obtained with positive thoughts and positive affect; that is, the mechanism could be a result of salience and not specific to negatively toned material. However, it is important to note that Salkovskis (1985) and Salkovskis, Richards and Forrester (1995) have emphasized that both the occurrence and the content of intrusions contribute to the negative appraisal. Thus, an apparently positive intrusive thought which inappropriately (to the person) occurs during a solemn or sad occasion may result in discomfort, self-criticism, attempts to suppress and neutralize and so on. Again, the cognitive-behavioural hypothesis of OCD suggests that appraisal in terms of responsibility for harm is crucial. Other factors may also have mediated the differences observed in discomfort in the present experiment; for example, neutralizing may have had an impact on vividness, evaluative dimensions (e.g. acceptability), intrusiveness and so on. Future experiments might attempt to evaluate these possible mechanisms, perhaps particularly focusing on evaluative mechanisms (i.e. the way in which the intrusion is appraised; Salkovskis, 1989b, 1996c; Salkovskis et al., 1995), since there is already some evidence of an association between attitudes of responsibility and the occurrence of neutralizing. The experiment reported here demonstrated the way in which neutralizing by non-clinical Ss can increase the discomfort subsequently associated with naturally occurring negative intrusive cognitions. The apparent similarity to obsessional phenomena is compelling, and is consistent with predictions arising from cognitive-behavioural hypothesis. More specifically, the data provide support for the hypothesis that a key factor in obsessional disorders is the occurrence of increased threat-focused neutralizing behaviour in response to normal intrusions, rather than as a result of generalized cognitive deficits. Cognitive-behavioural theories of the maintenance of anxiety disorders have previously stressed the importance of such safety seeking behaviours (Salkovskis, 1991, 1996b). The present experiment suggests that such behaviours may also be involved in the etiology of obsessional problems. Future work may need to consider the importance of responsibility appraisals in the development of neutralizing behaviours in obsessional problems. If this can be established, then each of the key components of the cognitive-behavioural hypothesis will have been empirically demonstrated. Similar work is also needed on the related phenomenon of deliberate thought suppression (Wegner et al., 1987; Wegner, 1989; Trinder & Salkovskis, 1994; Salkovskis & Campbell, 1994), which almost certainly shares many of the characteristics of neutralizing as a coping strategy. It has also been suggested that the basis for the priority of processing of obsessional thoughts may lie in the interaction between automatic vs controlled processing (in this instance, intrusive thoughts vs attempted coping actions) and that the link between the two can be found in the motivating properties of appraisal of the originating intrusive thoughts (Salkovskis et al., 1995). Future investigations could thus focus on factors which may be involved in the origin and maintenance of obsessional problems, including neutralizing, thought suppression, adverse mood and beliefs concerning responsibility and interactions between these. Acknowledgements-Paul Salkovskis is Wellcome Trust Senior Research Fellow. Rowena Cook of Oxford College of Further Education provided invaluable assistance in the recruitment of Ss for this experiment. Martina Mueller and Hilary Warwick gave helpful comments on an earlier version of this manuscript, and Pavlos Anastasiades gave invaluable statistical advice.
REFERENCES Beck, A. T. (1976). Cog&ire therapy and the emotionnl disorders. New York: International Universities Press. Beck, A. T., Epstein, N., & Brown, G. (1988). An inventory for measuring clinical anxiety. Journal of Consulting, and Clinical Psychology, 56, 893-897.
Beck, A. T., Ward, C. H., Mendelsohn, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 18, 561-571.
Edwards, S., & Dickerson, M. (1987). On the similarity of positive and negative intrusions. Behoviour Research and Therapy, 25, 207-211.
Eves, F., & Tata, P. (1989). Phasic cardiac and electrodermal reactions to idiographic stimuli in obsessional subjects. Behavioural Psvchotherapy, 17, 71-82.
Effects of neutralizing on intrusive thoughts
219
Foa, E. B., & Steketee, G. S. (1979). Obsessive-compulsives: conceptual issues and treatment interventions. Progress in Behavior ModiJication, 8, l-53. Hodgson, R. J., & Rachman, S. J. (1978). Obsessional-compulsive complaints. Behaviour Research and Therapy, 15, 385-389.
Lavy, E., & van den Hout, M. (1990). Thought suppression induces intrusion. Behaviouraf Psychotherapy, 18, 251-258. Marks, 1. M. (1987). Fears, phobias and rituals. New York: Oxford University Press. Rachman. S. J., & Hodgson, R. L. (1980). Obsessions and compulsions. Englewood Cliffs, NJ: Prentice Hall. Rachman, S. J., & de Silva, P. (1978). Abnormal and normal obsessions. Behaviour Research and Therapy, 16, 233-248. Rachman, S. J., de Silva, P., & Roper, G. (1976). The spontaneous decay of compulsive urges. Behaviour Research and Therapy, 14, 445453.
Reed, G. F. (1985). Obsessional experience and compulsive behaviour: A cognitive-structural approach. London: Academic Press. Salkovskis, P. M. (1985). Obsessional-compulsive problems: A cognitive-behavioural analysis. Behaoiour Research and Therapy, 23, 571-583.
Salkovskis, P. M. (1989a). Obsessions and compulsions. In J. Scott, J. M. G. Williams & A. T. Beck (Eds.), Cognitive therapy: A clinical casebook. London: Croom Helm. Salkovskis, P. M. (1989b). Cognitive-behavioural factors and the persistence of intrusive thoughts in obsessional problems, Behaviour Research and Therapy, 27, 677-682.
Salkovskis, P. M. (1991). The importance of behaviour in the maintenance of anxiety and panic: A cognitive account. Behavioural Psychotherapy,
19, 619.
P. M. (1996a). Resolving the cognition-behaviour debate. In P. M. Salkovskis (Ed.), Trends in cognitive-behaviour therapy. Chichester: Wiley. Salkovskis. P. M. (1996b). The cognitive approach to anxiety: Threat beliefs, safety seeking behaviour, and the special case of health anxiety and obsessions. In P. M. Salkovskis (Ed.), Frontiers of cognitive therapy (pp. 48-74). New York: Guilford. Salkovskis, P. M. (1996~). Cognitive-behavioural approaches to the understanding of obsessional problems. In R. Rapee (Ed.), Current controversies in the anxiety disorders. New York: Guilford. Salkovskis, P. M., &Campbell, P. (1994). Thought suppression in naturally occurring negative intrusive thoughts. Behaviour Salkovskis,
Research and Therapy, 32, 1-8.
Salkovskis, P. M., & Dent, H. (in preparation). Intrusive thoughts and negative beliefs. Salkovskis, P. M., & Harrison, J. (1984). Abnormal and normal obsessions-a replication. Behauiour Research and Therapy, 22, 5499552.
Salkovskis, P. M., & Kirk, J. (1989). Obsessional disorders. In K. Hawton, P. M. Salkovskis, J. Kirk & D. M. Clark (Eds.), Cognitive-behavioural treatment for psychiatric disorders: A practical guide. Oxford: Oxford University Press. Salkovskis. P. M., & Kirk, J. (1997). Obsessivecompulsive disorder. In D. M. Clark & C. G. Fairburn (Eds.), Science and practice of cognitive behaviour therapy. Oxford: Oxford University Press. Salkovskis, P. M., Rachman, S. L., Ladouceur, R., Freeston, M., Taylor, S., Kyrios, M., & Sica, C. (1996). Defining responsibility in obsessional problems: Proceedings of the Smith College Women’s Room-after
the Toronto Cafeteria.
Salkovskis, P. M., Richards, C., & Forrester, E. (1995). The relationship between obsessional problems and intrusive thoughts. Behavioural and Cognitive Psychotherapy, 23, 281-299. Salkovskis. P. M., & Westbrook, D. (1989). Behaviour therapy and obsessional rummations: Can failure be turned into success?. Behaviour Research and Therapy, 27, 149-160. Trinder. H., & Salkovskis, P. M. (1994). Personally relevant intrusions outside the laboratory: Long-term suppression increases intrusion. Behaviour Research and Therapy, 32, 833-842. Wegner. D. M. (1989). White bears and other um~anted thoughts: Suppression, obsession and thepsychology of mental control. New York: Viking. Wegner, D. M., Schneider, D. J., Carter, S. R., & White, T. L. (1987). Paradoxical effects of thought suppression. Journal of Personality and Social Psychology, 53, S-13.