- Email: [email protected]
Catastrophic worry in primary insomnia
Journal of Behavior Therapy and Experimental Psychiatry 34 (2003) 11–23
Catastrophic worry in primary insomnia Allison G. Harveya,b,*, Emmeline Greenallb a
Department of Experimental Psychology, University of Oxford, Oxford, UK b Department of Psychiatry, University of Oxford, UK
Received 1 February 2002; received in revised form 12 December 2002; accepted 24 December 2002
Abstract The present study aimed to provide an empirical test of the proposal that catastrophic worry about the consequences of not sleeping is common among patients with primary insomnia and serves to maintain the sleep disturbance. It was predicted that relative to good sleepers, patients with primary insomnia would catastrophize more and that catastrophizing would be associated with increased negative affect and increased perception of threat. A ‘catastrophizing interview’ was administered to 30 patients with primary insomnia and 30 good sleepers. Consistent with the predictions, the insomnia patients generated more catastrophes about the consequences of not sleeping and gave higher likelihood ratings than good sleepers. For the insomnia group, but not the good sleepers, the catastrophizing interview was associated with increased anxiety and discomfort. The limitations of the study, possibilities for future research, and clinical implications of these findings are discussed. r 2003 Elsevier Science Ltd. All rights reserved. Keywords: Insomnia; Worry; Catastrophizing; Anxiety; Sleep
1. Introduction In the past decade there has been a surge of interest in the role of cognitive processes in the maintenance of insomnia (Espie, 1991, 2002; Fichten et al., 2001; Harvey, 2002; Lundh, 1998; Perlis, Giles, Mendelson, Bootzin, & Wyatt, 1997; Morin, 1993). The most rigorously researched cognitive phenomenon in the context of insomnia, to date, is the role of dysfunctional beliefs about sleep. A measure of dysfunctional beliefs about sleep has been developed (Morin, 1993) and a short *Corresponding author. Department of Psychiatry, University of Oxford, Warneford Hospital, Oxford OX3 7JX, UK. Tel.: +44-1865-22-39-12; fax: +44-1865-310-447. E-mail address: [email protected] (A.G. Harvey). 0005-7916/03/$ - see front matter r 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0005-7916(03)00003-X
12
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
version validated (Espie, Inglis, Harvey, & Tessier, 2000). Using this measure, Morin and colleagues have found that older adults with insomnia are less realistic than good sleepers about how much sleep they require, they strongly endorse statements relating to the negative consequences of insomnia, and are more likely to attribute their insomnia to external and stable causes (Morin, Stone, Trinkle, Mercer, & Remsberg, 1993). This pattern of findings has been replicated by Fins et al. (1996). Further, Edinger and colleagues found that a reduction in dysfunctional beliefs about sleep, achieved as a result of cognitive behaviour therapy for insomnia, was associated with positive post-treatment outcome (Edinger, Wohlgemuth, Radtke, Marsh, & Quillian, 2001). Another cognitive phenomenon that has received much attention from insomnia researchers in the 1980s, but less in recent times, is the role of pre-sleep worry. One of the most replicated findings in the field is that people with insomnia complain that they cannot get to sleep because of unpleasant intrusive thoughts and excessive and uncontrollable worry during the pre-sleep period. In a sample of 296 patients with insomnia, cognitive arousal was ten times more likely than somatic arousal to be cited as the main determinant of the sleep disturbance (Lichstein & Rosenthal, 1980). Consistently, Espie and colleagues reported that the cognitive items of the sleep disturbance questionnaire (SDQ; e.g., ‘My minds keeps turning things over’, ‘I am unable to empty my mind’) were the most highly rated (Espie, Brooks, & Lindsay, 1989), a finding that has been recently replicated (Harvey, 2000). Also in support of a role for worry in the maintenance of insomnia are the studies delineating the personality of people with insomnia. In a review of this literature, Borkovec (1982) characterized patients with insomnia as anxious and obsessively worrisome. Furthermore, insomnia is frequently comorbid with one or more anxiety disorders, especially generalized anxiety disorder (GAD), a disorder characterized by pathological worry (Ford & Kamerow, 1989). Within the anxiety literature the tendency to catastrophize has been identified as exacerbating the adverse effects of pathological worry (Kendall & Hollon, 1989; Kendall & Ingram, 1987). Catastrophizing involves ‘dwelling on the worst possible outcomes of any situation in which there is a possibility for an unpleasant outcome. Theyperson overemphasizes the probability of this catastrophic outcome and usually exaggerates the possible consequences of its occurrence’ (p. 33, Beck & Emery, 1985). Catastrophizing often involves ‘the worrier persistently posing internal, automatic questions of the ‘what if?’ kind’ (p. 83, Startup & Davey, 2001. See also Kendall & Ingram, 1987). In the context of insomnia, both the clinical treatment literature (e.g., Morin, 1993; Perlis et al., 2000) and the self-help literature (e.g., Sharp, 2001) highlight the tendency for patients with insomnia to catastrophize the consequences of sleep disturbance: ‘insomniacs tend to catastrophize over temporary sleep loss and to amplify its negative impact on daytime functioning’ (Morin, p. 143). While these observations are intuitively and clinically appealing, the importance of catastrophizing to the maintenance of insomnia remains to be empirically established. Vasey and Borkovec (1992) developed an experimental paradigm to capture the catastrophizing process and its consequences. The paradigm is essentially the
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
13
opposite of the decatastrophizing method typically employed in cognitive therapy (Barlow, 1992; Beck & Emery, 1985). It requires participants to generate sequences of possible catastrophic consequences for worrisome topics using the ‘what if’ questioning style observed clinically (Kendall & Ingram, 1987) and empirically (Kendall & Hollon, 1989) to be characteristic of patients with anxiety disorders. Using their paradigm, Vasey and Borkovec (1992) found that chronic worriers generated more imagined catastrophes than non-worriers and the process of catastrophizing was associated with increased negative affect for chronic worriers but not the non-worriers. Further, the chronic worriers rated the likelihood that the catastrophe steps generated would actually happen higher than the non-worriers. Taken together, experts have observed catastrophic worry about the consequences of not sleeping to be characteristic of patients with insomnia. Within the anxiety literature an experimental paradigm for investigating the tendency to catastrophize and the consequences of catastrophizing has been developed. In the present study we adapted this experimental paradigm to provide the first empirical evaluation of the role of catastrophizing in primary insomnia. Three predictions were tested. First, based on the clinical observation that patients with insomnia are characterised by a tendency to catastrophize (Morin, 1993; Perlis et al., 2000; Sharp, 2001), it was predicted that patients with insomnia will generate a larger number of responses to ‘what if’ questions (i.e., they will catastrophize more) than good sleepers. Second, based on Vasey and Borkovec’s (1992) findings, it was predicted that patients with insomnia will report an increase in anxiety and discomfort over the course of the catastrophizing interview relative to good sleepers. Third, based on the proposal that a tendency to catastrophize is associated with a distorted perception of threat (Davey & Levy, 1998; Kendall & Ingram, 1987), we predicted that the insomnia group would rate the likelihood that the catastrophes generated might actually occur higher than the good sleeper group. Drawing hypotheses 2 and 3 from the anxiety literature is justified on the basis of the parallels noted between insomnia and anxiety, including the experience of worry and uncontrollable negative intrusive thoughts and images (Borkovec, Robinson, Pruzinsky, & De Pree, 1983).
2. Method 2.1. Participants The first recruitment strategy was to place posters and flyers around the city asking for those ‘‘interested in sleep research’’ to contact the experimenters. This strategy resulted in the recruitment of 36 good sleepers but only 8 patients with insomnia. Hence the original posters and flyers were replaced with ones that specifically advertised for people with insomnia. Overall, of the 78 people to respond, 15 cancelled prior to the appointment or did not attend and 3 were excluded as they were dissatisfied with their sleep but did not meet diagnostic criteria for insomnia. Note that the inclusion criteria for the good sleeper group were (1) not meeting diagnostic criteria for insomnia and (2) being satisfied with their sleep. The
14
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
remaining sixty individuals were included in the study; 30 were diagnosed with insomnia according to DSM-IV (American Psychiatric Association, 1994) and 30 were good sleepers. In the absence of a psychometrically validated alternative, a structured interview (Insomnia Diagnostic Interview; IDI) was administered in order to carefully assess for each of the DSM-IV criteria for insomnia. That is, questions were asked to establish (1) that the predominant complaint was a difficulty with initiating or maintaining sleep or non-restorative sleep for at least 1 month (Cluster A), (2) that the complaint causes distress or impairment (Cluster B), (3) that it does not occur exclusively as a result of another sleep or psychiatric disorder (Clusters C & D), and (4) that it is not due to the effects of a substance or illness (Cluster E). In the absence of DSM-IV including a severity criterion for insomnia, only those who complained of sleep problems for at least three nights a week were included (Morin, 1993; World Health Organization, 1992). In the insomnia group, nine participants (31%) reported having received treatment for insomnia in the past. In addition, three participants (10%) reported having, in the past, received psychological treatment for PTSD ðn ¼ 1Þ and anxiety ðn ¼ 2Þ: In the good sleeper group, three participants (10%) reported having, in the past, received psychological treatment for relationship problems ðn ¼ 2Þ and depression ðn ¼ 1Þ: 2.2. Procedure The IDI was administered by a qualified clinical psychologist and participants then completed the Beck Anxiety Inventory (BAI; Beck, Epstein, Brown, & Steer, 1988), Beck Depression Inventory (BDI; Beck, Ward, Mendelson, Mock, & Erbaugh, 1961), and the Penn State Worry Questionnaire (PSWQ; Meyer, Miller, Metzger, & Borkovec, 1990). Participants were asked to rate the severity of their insomnia in the past month (Response Scale: 0 ‘Not at all disturbing and/or disabling, 8 ‘Severely disturbing/disabling’) and how concerned they have been about their sleep problem in the past month (Response scale: 0 ‘Not at all’, 8 ‘Very much’). These variables will be known as ‘sleep disturbance severity’ and ‘sleep concern’, respectively. Participants were also asked to estimate the probability that they will sleep badly on any one night (Response Scale: 0 ‘It is very unlikely I will sleep badly’ to 10 ‘It is very likely I will sleep badly’) and to estimate the cost of sleeping badly on any one night’ (Response Scale: 0 ‘There is no cost, sleeping badly does not effect me at all’ to 10 ‘Sleeping badly is extremely costly and significantly disrupts my life). Theoretically, if catastrophizing is associated with increased perception of threat (Davey & Levy, 1998; Kendall & Ingram, 1987), it should also be associated with increased cost and probability estimates. The Catastrophizing Interview employed broadly followed the procedures of Vasey and Borkovec (1992) and Davey and Levy (1998), except that the lead question was specific to the consequences of sleep disturbance. In previous studies, all conducted in the context of chronic worry, the lead question was asked with reference to any current worry. In our procedure participants were asked ‘What is it that worries you about nights when you have problems getting to sleep?’. Where A
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
15
Table 1 An example of the catastrophizing steps generated by an insomnia patient and the likelihood rating for each response Catastrophizing step
Likelihood rating
I will feel bad the following morning I won’t cope at work The back log at work will build because I cannot concentrate I will get drowsy and irritable I won’t be nice company I’ll get into trouble at work—my boss will shout at me My job will be on the line I’ll be unemployed I will have a low self-esteem
10 9 7 8 9 6 7 9 9
Note: Likelihood rating ¼ 0 ‘not at all likely’, 10 ‘extremely likely’.
denotes the answer to the latter question the participant was asked ‘What is it that worries you about A?’. Where B denotes the answer the participant gave to this second question the participant was asked ‘What is it that worries you about B’. Following Davey and Levy (1998), this questioning process continued until either the participant said they had no more answers or until they repeated a similar answer three times. Participants were then asked to rate, for each response, the likelihood that it would actually occur (0 ‘not at all likely’, 10 ‘extremely likely’). This rating will be referred to as the ‘likelihood rating’. Note that immediately before and immediately after the procedure participants were asked to rate their anxiety (0 ‘Not at all anxious’ to 10 ‘Extremely anxious’) and discomfort (0 ‘No discomfort’ to 10 ‘Extreme discomfort’). An example of the catastrophizing steps generated by one insomnia patient and the likelihood rating for each response is presented in Table 1. At the end of the experiment the participants were debriefed and paid an honorarium.
3. Results 3.1. Participant characteristics Table 2 presents the mean scores for participant characteristics, sleep characteristics, and measures of psychopathology. The groups did not differ on sex composition [insomnia—18 females, 12 males; good sleepers—19 females, 11 males, w2 (1, N ¼ 60) ¼ 0:07; ns] or age. The average duration of the sleep disturbance for the insomnia group was 6.3 years. Sleep onset latency for the past month was longer and total sleep time shorter for the insomnia group compared to the good sleepers. The insomnia group scored higher on sleep disturbance severity, sleep concern, probability, cost, BAI, and the PSWQ. The insomnia group was marginally significantly higher on BDI compared with the good sleeper group ðp ¼ 0:07Þ:
16
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
Table 2 Mean values for participant characteristics, sleep characteristics and measures of psychopathology
Age Length of insomnia Sleep onset latency Total sleep time Sleep disturbance severity Sleep concern Probability Cost BAI BDI PSWQ
Insomnia
Good sleeper
tð58Þ
20.4 (4.7) 6.3 (4.7) 58.2 (45.9) 6.3 (1.2) 3.6 (1.3) 3.7 (1.8) 5.5 (1.7) 5.5 (1.8) 13.6 (9.6) 10.3 (7.3) 57.4 (15.1)
22.3 (8.9) N/A 15.3 (11.0) 7.5 (1.1) 0.2 (0.5) 0.1 (0.3) 1.5 (2.0) 3.9 (2.4) 7.3 (7.2) 7.0 (6.6) 41.2 (14.9)
1.03 N/A 4.98nn 4.15w 13.8w 10.9w 8.3w 2.8nn 2.90nn 1.83 4.17w
Note: Standard deviations appear in parentheses. Age and length of insomnia in years. Sleep onset latency¼for the past month, in minutes. Total sleep time ¼ for the past month, in hours. Sleep disturbance severity ¼ 0 ‘Not at all disturbing and/or disabling, 8 ‘Severely disturbing/disabling’. Sleep concern ¼ 0 ‘Not at all’, 8 ‘Very much’. Probability ¼ 0 ‘It is very unlikely I will sleep badly’ to 10 ‘It is very likely I will sleep badly’. Cost ¼ 0 ‘There is no cost, sleeping badly does not effect me at all’ to 10 ‘Sleeping badly is extremely costly and significantly disrupts my life. BAI ¼ Beck Anxiety Inventory. BDI ¼ Beck Depression Inventory. PSWQ ¼ Penn State Worry Questionnaire. nn po0:01: w po0:001:
3.2. Catastrophizing interview Immediately prior to the catastrophizing interview the insomnia group were more anxious, tð58Þ ¼ 7:64; po0:0001; and reported more discomfort, tð58Þ ¼ 5:01; po0:0001; compared to the good sleeper group (see left side of Figs. 1a and b for mean values). The mean number of catastrophizing steps was 6 ðSD ¼ 2:6Þ for the insomnia group and 1.8 ðSD ¼ 1:9Þ for the good sleeper group. The change in emotionality ratings from the beginning to the end of the catastrophizing interview was examined with two repeated measures ANOVAs. For the first ANOVA, Diagnosis (Insomnia vs. Good sleeper) was entered as a between subjects factor and Anxiety (pre-interview vs. post-interview) was entered as a within subjects factor. The Diagnosis main effect was significant, F ð1; 58Þ ¼ 142:51; po0:001; such that overall the insomnia group had higher anxiety scores than the good sleeper group. The Anxiety main effect was significant, F ð1; 58Þ ¼ 42:34; po0:001; such that overall ratings of anxiety increased after the catastrophizing interview relative to before it. The interaction was also significant, F ð1; 58Þ ¼ 42:34; po0:001: As depicted in Fig. 1a, the insomnia group rated their anxiety to be higher than the good sleeper group both before, tð58Þ ¼ 7:64; po0:001; and after, tð58Þ ¼ 13:10; po0:001; the interview. However, the insomnia group rated their anxiety to be significantly higher after the interview relative to before the interview, tð58Þ ¼ 7:31; po0:001: This same difference was not significant for the good sleeper group.
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
17
Fig. 1. (a) Mean values for anxiety ratings before and after the ‘catastrophizing interview’. (b) Mean values for discomfort ratings before and after the ‘catastrophizing interview’.
For the second ANOVA, Diagnosis (Insomnia vs. Good sleeper) was entered as a between subjects factor and Discomfort (pre-interview vs. post-interview) was entered as a within subjects factor. The Diagnosis main effect was significant, F ð1; 58Þ ¼ 53:17; po0:001; such that overall the insomnia group had higher discomfort scores than the good sleeper group. The Discomfort main effect was significant, F ð1; 58Þ ¼ 32:36; po0:001; such that ratings of discomfort increased after the catastrophizing interview relative to before it. The interaction was also significant, F ð1; 58Þ ¼ 19:95; po0:001: As depicted in Fig. 1b, the insomnia group rated their discomfort to be higher than the good sleeper group both before,
18
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
tð58Þ ¼ 5:01; po0:001; and after, tð58Þ ¼ 7:89; po0:001; the interview. However, the insomnia group rated their discomfort to be significantly higher after the interview relative to before the interview, tð58Þ ¼ 5:31; po0:001: This same difference was not significant for the good sleeper group. Eleven people in the good sleeper group did not generate any catastrophizing steps. Not including these participants, the average likelihood rating for the insomnia group was higher (M ¼ 6:52; SD ¼ 1:9) than the good sleeper group (M ¼ 5:0; SD ¼ 2:6), tð47Þ ¼ 2:32; po0:05: 3.3. Correlational analyses Table 3 presents correlational analyses between the total number of catastrophe steps generated and the average likelihood rating with sleep and psychopathology variables. As the relationship between variables was expected to differ according to diagnosis, these analyses were carried out separately for the insomnia and good sleeper groups. For the insomnia group, the number of catastrophe steps generated was significantly positively correlated with sleep concern and PSWQ score. For the insomnia group, the average likelihood rating was significantly negatively correlated with length of insomnia and significantly positively correlated with cost and
Table 3 Correlational analysis between number of catastrophizing steps generated, average likelihood ratings, sleep and psychopathology measures Good sleeper group
Insomnia group
Number of steps Average likelihood Number of steps Average likelihood ðn ¼ 30Þ ðn ¼ 19Þa ðn ¼ 30Þ ðn ¼ 30Þ Length of insomnia N/A Sleep disturbance severity 0.17 Sleep concern 0.01 Probability 0.12 Cost 0.25 BAI 0.08 BDI 0.30 PSWQ 0.16 Anxiety rating Discomfort rating a
N/A 0.17 0.12 0.31 0.31 0.43 0.13 0.14
0.01 0.07 0.42n 0.21 0.18 0.30 0.33 0.40n
0.50nn 0.39 0.45 0.38n 0.54n 0.12 0.01 0.19
n ¼ 11 did not generate any catastrophizing steps and so did not make likelihood ratings. Length of insomnia in years. Sleep disturbance severity ¼ 0 ‘Not at all disturbing and/or disabling, 8 ‘Severely disturbing/disabling’. Sleep concern ¼ 0 ‘Not at all’, 8 ‘Very much’. Probability ¼ 0 ‘It is very unlikely I will sleep badly’ to 10 ‘It is very likely I will sleep badly’. Cost ¼ 0 ‘There is no cost, sleeping badly does not effect me at all’ to 10 ‘Sleeping badly is extremely costly and significantly disrupts my life. BAI ¼ Beck Anxiety Inventory. BDI ¼ Beck Depression Inventory. PSWQ ¼ Penn State Worry Questionnaire. n po0:05: nn po0:01:
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
19
probability estimates. There were no significant correlations for the good sleeper group.
4. Discussion The aim of the present study was to borrow an experimental paradigm from the anxiety literature to investigate the presence and role of catastrophizing in the maintenance of insomnia. Three hypotheses were tested. The first hypothesis was that patients with insomnia would generate more catastrophizing steps than good sleepers to the question ‘‘What is it that worries you about nights when you have problems getting to sleep?’’. This hypothesis was supported as patients with insomnia produced significantly more catastrophizing steps than good sleepers. This finding is consistent with previous studies implicating excessive worry about sleep as characteristic of insomnia (Lichstein & Rosenthal, 1980; Espie et al., 1989; Harvey, 2000), and with Borkovec’s (1982) finding that patients with insomnia tend to be anxious and obsessively worrisome. Further, the findings empirically validate observations in both the clinical (Morin, 1993; Perlis et al., 2000) and self-help (Sharp, 2001) literatures that patients with insomnia have a tendency to catastrophize about their sleep disturbance. Finally, Borkovec et al. (1983) have previously noted several parallels between insomnia and the anxiety disorders. The current study adds to this list by highlighting that like individuals with anxiety (Kendall & Hollon, 1989), individuals with insomnia exhibit a tendency to catastrophize. The second hypothesis was that patients with insomnia would report an increase in anxiety and discomfort over the course of the catastrophizing interview. In support, when anxiety and discomfort ratings were examined, ratings made by the insomnia group increased significantly from immediately before to immediately following the catastrophizing interview. This same difference was not significant for the good sleeper group. The increase in anxiety and discomfort following catastrophizing for the patients with insomnia suggests that the procedure leads to activation of threatrelated material (Vasey & Borkovec, 1992). If this paradigm captures something of the internal dialogue during the pre-sleep period when difficulty falling asleep is experienced, one could speculate that the observed increase in affect associated with catastrophizing could contribute to the maintenance of insomnia because increasing anxiety and discomfort are unlikely to be conditions conducive to optimal sleeponset (Harvey, 2002). Espie (2002) highlighted the importance of a state of minimal affective load for optimal sleep-onset. This paradigm may also capture something of the internal dialogue during the day. If so, the observed increase in affect associated with catastrophizing would create cognitive load (Sarason, 1984; Sarason, Sarason, Keefe, Hayes, & Shearin, 1986) that could contribute to deficits in memory, attention, and concentration, which if attributed to lack of sleep are likely to increase pre-occupation with and anxiety about getting enough sleep (Harvey, 2002). This may be one of the pathways by which daytime processes serve to exacerbate nighttime sleep problems.
20
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
The third hypothesis tested was that the insomnia group would rate the likelihood that the catastrophizing steps generated might actually occur higher than the good sleeper group. While patients with insomnia perceived the catastrophizing steps they generated to be more likely to occur than the good sleepers. That is, although the evidence was consistent with the third hypothesis, baseline differences in anxiety and discomfort appear to have contributed significantly to the inflated likelihood ratings that characterized the insomnia group. This latter finding makes intuitive sense; the likelihood estimations are highly likely to be influenced by the extent to which the person is feeling comfortable and secure at that time point. However, it should be noted that a caveat on this finding is that the 11 good sleepers who were not able to generate any catastrophe steps, and did not generate a likelihood rating, were not included in this analysis. It is therefore likely that this may have skewed the good sleeper group. A number of interesting results emerged from the correlational analyses. For the insomnia group, the more catastrophizing steps generated the greater the amount of sleep concern and the higher the PSWQ score. These results suggest that the tendency to catastrophize among patents with insomnia is associated with a specific sleep outcome, concern about sleep, and with a non-specific outcome, the general tendency to worry. The shorter the duration of the insomnia the higher the ratings of the likelihood that the catastrophes generated might actually occur. There are two possibilities here. First, this finding suggests that over time, while individuals with insomnia continue to catastrophize, they become more realistic about the likelihood of the catastrophe actually occurring. Second, perhaps catastrophizing plays a more important role during the early stages and becomes less relevant when the insomnia becomes chronic. Suggestive of a mechanism by which catastrophizing may contribute to the maintenance of insomnia, the higher the average likelihood rating the higher the estimated probability of sleeping badly and the higher the estimated cost of sleeping badly. As we would expect, no correlations reached significance for the good sleeper group. It is interesting to note that the number of catastrophizing steps generated in this study by the insomnia group ðM ¼ 6:0Þ was lower than the number reported by Vasey and Borkovec ðM ¼ 14:2Þ and more akin to the results of Davey and Levy (M ¼ 8:1—experiment 1, M ¼ 6:5—experiment 2). There are a number of possible accounts for this observation. First, before starting the catastrophizing interview Vasey and Borkovec (1992) administered a practice trial that involved demonstrating the stepwise procedure. This component was not part of the procedure in the present study or in the Davey and Levy (1998) study. Prior exposure to the procedure may have increased the fluency of Vasey and Borkovec’s participants. Second, Vasey and Borkovec (1992) recruited chronic worriers. It is possible that such participants catastrophize more than insomnia patients. With a view to furthering research on catastrophizing in insomnia, several methodological points require consideration. It should be noted that the sample was based on a university population. As such, it is not necessarily representative of a general practitioner-referred insomnia population. It should be noted, however, that all participants met strict DSM-IV criteria for primary insomnia and that the
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
21
average duration of the insomnia was 6.3 years. Second, although the sample was not objectively assessed for comorbid sleep disorders (e.g., obstructive sleep apnea and narcolepsy) the IDI included a section designed to carefully screen for the key symptoms of comorbid sleep disorders. None of the participants responded positively to any of these screening questions. Considerations directly relating to the catastrophizing interview are threefold. First, it should be noted that the experimenter who administered the catastrophizing interview (AH) was not blind to diagnosis. While the questioning method and stop rules were clearly specified prior to testing and carefully adhered to during testing, this potential limitation should be corrected in future research. Second, careful consideration should be given to the opening phrase of the catastrophizing interview. Subtle changes in wording such as ‘What is it that worries you on nights that you cannot sleep?’ may impact responses. In addition, variations in the opening phrase may help illuminate different aspects of catastrophizing. As the nature of daytime processes in insomnia are increasingly of interest (Harvey, 2002), it may be illuminating to adopt an opening phrase such as: ‘What is it that worries you about your daytime activities following a night of poor sleep’. Finally, prior to the catastrophizing interview the insomnia group reported higher anxiety and distress relative to the good sleeper group. Miller and Chapman (2001) have drawn attention to the error introduced by correcting for baseline differences between groups by conducting analysis of covariance when the design does not involve random allocation to groups. To ensure that the greater number of catastrophizing steps generated by the insomnia group, relative to the good sleeper group, are not solely accounted for by baseline differences, future research should compare an insomnia group to a high anxious good sleeper group. Another important issue is that the association between catastrophizing and insomnia does not necessarily imply a causal relationship. Freedman and Sattler (1982) have argued that excessive cognitive activity is epiphenomenal to sleeplessness. An assumption of this study is that responses to the catastrophizing interview conducted in a laboratory setting reflect or parallel processes that operate in the ‘real world’. Naturalistic observational studies of catastrophizing are necessary to shed light on the validity of this assumption. Of course, this study was heavily based on theory and research relating to GAD. While GAD is commonly comorbid with insomnia and the two disorders have many common features (Borkovec et al., 1983), researchers should remain ‘on the look-out’ for differences across the two disorders. In conclusion, the present study constitutes an initial empirical attempt to capture the catastrophizing process in the context of insomnia. The method employed appears to have been useful for observing one aspect of the thought processes/ internal dialogue characteristic of patients with insomnia. As predicted, we found that the insomnia patients generated more potential catastrophes about the consequences of not sleeping relative to good sleepers. For the insomnia group, but not the good sleepers, catastrophizing was associated with increased anxiety and discomfort. In a large case series, Perlis et al. (2000), as part of a multi-component treatment package, employed Barlow’s (1992) decastrophization method. This involved exploring, with the patient, the discrepancy between their certainty of
22
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
negative outcome and the real probability of negative outcome. The results of the present study provides endorsement for the inclusion of such interventions in treatments for insomnia.
Acknowledgements The authors are grateful to Dr. Lars-Gunnar Lundh for helpful comments on the manuscript. This research was supported by the Wellcome Trust (Grant reference number—065913).
References American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Barlow, D. H. (1992). Cognitive-behavioral approaches to panic disorder and social phobia. Bulletin of the Menninger Clinic, 56. Beck, A. T., & Emery, G. D. (1985). Anxiety disorders and phobias: A cognitive perspective. New York: Basic Books. Beck, A. R., Epstein, N., Brown, G., & Steer, R. A. (1988). An inventory for measuring clinical anxiety: Psychometric properties. Journal of Consulting and Clinical Psychology, 56, 893–897. Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory measuring depression. Archives of General Psychiatry, 4, 561–571. Borkovec, T. D. (1982). Insomnia. Journal of Consulting and Clinical Psychology, 50, 880–895. Borkovec, T. D., Robinson, E., Pruzinsky, T., & De Pree, J. A. (1983). Preliminary exploration of worry: Some characteristics and processes. Behaviour Research and Therapy, 21, 9–16. Davey, G. C. L., & Levy, S. (1998). Catastrophic worrying: Personal inadequacy and a perseverative iterative style as features of the catastrophizing process. Journal of Abnormal Psychology, 107, 576–586. Edinger, J. D., Wohlgemuth, W. K., Radtke, R. A., Marsh, G. R., & Quillian (2001). Does cognitivebehavioural insomnia therapy alter dysfunctional beliefs about sleep? Sleep, 24, 591–599. Espie, C. A. (1991). The psychological treatment of insomnia. England: Wiley. Espie, C. A. (2002). Insomnia: Conceptual issues in the development, persistence, and treatment of sleep disorder in adults. Annual Review of Psychology, 53, 215–243. Espie, C. A., Brooks, D. N., & Lindsay, W. R. (1989). An evaluation of tailored psychological treatment of insomnia. Journal of Behaviour Therapy and Experimental Psychiatry, 20, 143–153. Espie, C. A., Inglis, S. J., Harvey, L., & Tessier, S. (2000). ‘Insomniacs’ attributions. Psychometric properties of the dysfunctional beliefs and attitudes about sleep scale and the sleep disturbance questionnaire. Journal of Psychosomatic Research, 48, 141–148. Fichten, C. S., Libman, E., Creti, L., Amsel, R., Sabourin, S., Brender, W., & Bailes, S. (2001). Role of thoughts during nocturnal awake times in the insomnia experience of older adults. Cognitive Therapy and Research, 25, 665–692. Fins, A. I., Edinger, J. D., Sullivan, R. J., Marsh, G. R., Dailey, D., Hope, T. V., Young, M., Shaw, E., & Vailas, D. (1996). Dysfunctional cognitions about sleep among older adults and their relationship to objective sleep findings. Sleep Research, 25, 242. Ford, D., & Kamerow, D. (1989). Epidemiologic study of sleep disturbances and psychiatric disorders. Journal of the American Medical Association, 262, 1479–1484. Freedman, R. R., & Sattler, H. L. (1982). Physiological and psychological factors in sleep-onset insomnia. Journal of Abnormal Psychology, 91, 380–389. Harvey, A. G. (2000). Pre-sleep cognitive activity in insomnia: A comparison of sleep-onset insomniacs and good sleepers. British Journal of Clinical Psychology, 39, 275–286.
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
23
Harvey, A. G. (2002). A cognitive model of insomnia. Behaviour Research and Therapy, 40, 869–893. Kendall, P. C., & Hollon, S. D. (1989). Anxious self-talk: Development of the anxious self-statements questionnaire (ASSQ). Cognitive Therapy and Research, 13, 81–93. Kendall, P. C., & Ingram, R. E. (1987). The future for cognitive assessment of anxiety: Let’s get specific. In L. Michaelson, & L. M. Ascher (Eds.), Anxiety and stress disorders: Cognitive-behavioural assessment and treatment. New York: Guildford Press. Lichstein, K. L., & Rosenthal, T. L. (1980). Insomniacs’ perceptions of cognitive versus somatic determinants of sleep disturbance. Journal of Abnormal Psychology, 89, 105–107. Lundh, L.-G. (1998). Cognitive-behavioural analysis and treatment of insomnia. Scandinavian Journal of Behaviour Therapy, 27, 10–29. Meyer, T. J., Miller, M. L., Metzger, R. L., & Borkovec, T. D. (1990). Development and validation of the Penn State Worry Questionnaire. Behaviour Research and Therapy, 28, 487–495. Miller, G. A., & Chapman, J. P. (2001). Misunderstanding analysis of covariance. Journal of Abnormal Psychology, 110, 40–48. Morin, C. M. (1993). Insomnia: Psychological assessment and management. New York: Guildford Press. Morin, C. M., Stone, J., Trinkle, D., Mercer, J., & Remsberg, S. (1993). Dysfunctional beliefs and attitudes about sleep among older adults with and without insomnia complaints. Psychology and Aging, 8, 463–467. Perlis, M., Aloia, M., Millikan, A., Boehmler, J., Smith, M., Greenblatt, & Giles, D. (2000). Behavioral treatment of insomnia: A clinical case series study. Journal of Behavioral Medicine, 23, 149–161. Perlis, M. L., Giles, D. E., Mendelson, W. B., Bootzin, R. R., & Wyatt, J. K. (1997). Psychophysiological insomnia: The behavioural model and a neurocognitive perspective. Journal of Sleep Research, 6, 179–188. Sarason, I. G. (1984). Test anxiety, stress, and social support. Journal of Personality, 49, 101–114. Sarason, I. G., Sarason, B. R., Keefe, D. E., Hayes, B. E., & Shearin, E. N. (1986). Cognitive interference: Situational determinants and traitlike characteristics. Journal of Personality and Social Psychology, 51, 215–226. Sharp, T. (2001). The good sleep guide. Australia: Penguin. Startup, H. M., & Davey, G. C. L. (2001). Mood as input and catastrophic worrying. Journal of Abnormal Psychology, 110, 83–96. Vasey, M. W., & Borkovec, T. D. (1992). Catastrophizing assessment of worrisome thoughts. Cognitive Therapy and Research, 16, 505–519. World Health Organization. (1992). The ICD-10 classification of mental and behavioural disorder: Diagnostic criteria for research (10th revision). Geneva: Author.
Catastrophic worry in primary insomnia Allison G. Harveya,b,*, Emmeline Greenallb a
Department of Experimental Psychology, University of Oxford, Oxford, UK b Department of Psychiatry, University of Oxford, UK
Received 1 February 2002; received in revised form 12 December 2002; accepted 24 December 2002
Abstract The present study aimed to provide an empirical test of the proposal that catastrophic worry about the consequences of not sleeping is common among patients with primary insomnia and serves to maintain the sleep disturbance. It was predicted that relative to good sleepers, patients with primary insomnia would catastrophize more and that catastrophizing would be associated with increased negative affect and increased perception of threat. A ‘catastrophizing interview’ was administered to 30 patients with primary insomnia and 30 good sleepers. Consistent with the predictions, the insomnia patients generated more catastrophes about the consequences of not sleeping and gave higher likelihood ratings than good sleepers. For the insomnia group, but not the good sleepers, the catastrophizing interview was associated with increased anxiety and discomfort. The limitations of the study, possibilities for future research, and clinical implications of these findings are discussed. r 2003 Elsevier Science Ltd. All rights reserved. Keywords: Insomnia; Worry; Catastrophizing; Anxiety; Sleep
1. Introduction In the past decade there has been a surge of interest in the role of cognitive processes in the maintenance of insomnia (Espie, 1991, 2002; Fichten et al., 2001; Harvey, 2002; Lundh, 1998; Perlis, Giles, Mendelson, Bootzin, & Wyatt, 1997; Morin, 1993). The most rigorously researched cognitive phenomenon in the context of insomnia, to date, is the role of dysfunctional beliefs about sleep. A measure of dysfunctional beliefs about sleep has been developed (Morin, 1993) and a short *Corresponding author. Department of Psychiatry, University of Oxford, Warneford Hospital, Oxford OX3 7JX, UK. Tel.: +44-1865-22-39-12; fax: +44-1865-310-447. E-mail address: [email protected] (A.G. Harvey). 0005-7916/03/$ - see front matter r 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0005-7916(03)00003-X
12
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
version validated (Espie, Inglis, Harvey, & Tessier, 2000). Using this measure, Morin and colleagues have found that older adults with insomnia are less realistic than good sleepers about how much sleep they require, they strongly endorse statements relating to the negative consequences of insomnia, and are more likely to attribute their insomnia to external and stable causes (Morin, Stone, Trinkle, Mercer, & Remsberg, 1993). This pattern of findings has been replicated by Fins et al. (1996). Further, Edinger and colleagues found that a reduction in dysfunctional beliefs about sleep, achieved as a result of cognitive behaviour therapy for insomnia, was associated with positive post-treatment outcome (Edinger, Wohlgemuth, Radtke, Marsh, & Quillian, 2001). Another cognitive phenomenon that has received much attention from insomnia researchers in the 1980s, but less in recent times, is the role of pre-sleep worry. One of the most replicated findings in the field is that people with insomnia complain that they cannot get to sleep because of unpleasant intrusive thoughts and excessive and uncontrollable worry during the pre-sleep period. In a sample of 296 patients with insomnia, cognitive arousal was ten times more likely than somatic arousal to be cited as the main determinant of the sleep disturbance (Lichstein & Rosenthal, 1980). Consistently, Espie and colleagues reported that the cognitive items of the sleep disturbance questionnaire (SDQ; e.g., ‘My minds keeps turning things over’, ‘I am unable to empty my mind’) were the most highly rated (Espie, Brooks, & Lindsay, 1989), a finding that has been recently replicated (Harvey, 2000). Also in support of a role for worry in the maintenance of insomnia are the studies delineating the personality of people with insomnia. In a review of this literature, Borkovec (1982) characterized patients with insomnia as anxious and obsessively worrisome. Furthermore, insomnia is frequently comorbid with one or more anxiety disorders, especially generalized anxiety disorder (GAD), a disorder characterized by pathological worry (Ford & Kamerow, 1989). Within the anxiety literature the tendency to catastrophize has been identified as exacerbating the adverse effects of pathological worry (Kendall & Hollon, 1989; Kendall & Ingram, 1987). Catastrophizing involves ‘dwelling on the worst possible outcomes of any situation in which there is a possibility for an unpleasant outcome. Theyperson overemphasizes the probability of this catastrophic outcome and usually exaggerates the possible consequences of its occurrence’ (p. 33, Beck & Emery, 1985). Catastrophizing often involves ‘the worrier persistently posing internal, automatic questions of the ‘what if?’ kind’ (p. 83, Startup & Davey, 2001. See also Kendall & Ingram, 1987). In the context of insomnia, both the clinical treatment literature (e.g., Morin, 1993; Perlis et al., 2000) and the self-help literature (e.g., Sharp, 2001) highlight the tendency for patients with insomnia to catastrophize the consequences of sleep disturbance: ‘insomniacs tend to catastrophize over temporary sleep loss and to amplify its negative impact on daytime functioning’ (Morin, p. 143). While these observations are intuitively and clinically appealing, the importance of catastrophizing to the maintenance of insomnia remains to be empirically established. Vasey and Borkovec (1992) developed an experimental paradigm to capture the catastrophizing process and its consequences. The paradigm is essentially the
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
13
opposite of the decatastrophizing method typically employed in cognitive therapy (Barlow, 1992; Beck & Emery, 1985). It requires participants to generate sequences of possible catastrophic consequences for worrisome topics using the ‘what if’ questioning style observed clinically (Kendall & Ingram, 1987) and empirically (Kendall & Hollon, 1989) to be characteristic of patients with anxiety disorders. Using their paradigm, Vasey and Borkovec (1992) found that chronic worriers generated more imagined catastrophes than non-worriers and the process of catastrophizing was associated with increased negative affect for chronic worriers but not the non-worriers. Further, the chronic worriers rated the likelihood that the catastrophe steps generated would actually happen higher than the non-worriers. Taken together, experts have observed catastrophic worry about the consequences of not sleeping to be characteristic of patients with insomnia. Within the anxiety literature an experimental paradigm for investigating the tendency to catastrophize and the consequences of catastrophizing has been developed. In the present study we adapted this experimental paradigm to provide the first empirical evaluation of the role of catastrophizing in primary insomnia. Three predictions were tested. First, based on the clinical observation that patients with insomnia are characterised by a tendency to catastrophize (Morin, 1993; Perlis et al., 2000; Sharp, 2001), it was predicted that patients with insomnia will generate a larger number of responses to ‘what if’ questions (i.e., they will catastrophize more) than good sleepers. Second, based on Vasey and Borkovec’s (1992) findings, it was predicted that patients with insomnia will report an increase in anxiety and discomfort over the course of the catastrophizing interview relative to good sleepers. Third, based on the proposal that a tendency to catastrophize is associated with a distorted perception of threat (Davey & Levy, 1998; Kendall & Ingram, 1987), we predicted that the insomnia group would rate the likelihood that the catastrophes generated might actually occur higher than the good sleeper group. Drawing hypotheses 2 and 3 from the anxiety literature is justified on the basis of the parallels noted between insomnia and anxiety, including the experience of worry and uncontrollable negative intrusive thoughts and images (Borkovec, Robinson, Pruzinsky, & De Pree, 1983).
2. Method 2.1. Participants The first recruitment strategy was to place posters and flyers around the city asking for those ‘‘interested in sleep research’’ to contact the experimenters. This strategy resulted in the recruitment of 36 good sleepers but only 8 patients with insomnia. Hence the original posters and flyers were replaced with ones that specifically advertised for people with insomnia. Overall, of the 78 people to respond, 15 cancelled prior to the appointment or did not attend and 3 were excluded as they were dissatisfied with their sleep but did not meet diagnostic criteria for insomnia. Note that the inclusion criteria for the good sleeper group were (1) not meeting diagnostic criteria for insomnia and (2) being satisfied with their sleep. The
14
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
remaining sixty individuals were included in the study; 30 were diagnosed with insomnia according to DSM-IV (American Psychiatric Association, 1994) and 30 were good sleepers. In the absence of a psychometrically validated alternative, a structured interview (Insomnia Diagnostic Interview; IDI) was administered in order to carefully assess for each of the DSM-IV criteria for insomnia. That is, questions were asked to establish (1) that the predominant complaint was a difficulty with initiating or maintaining sleep or non-restorative sleep for at least 1 month (Cluster A), (2) that the complaint causes distress or impairment (Cluster B), (3) that it does not occur exclusively as a result of another sleep or psychiatric disorder (Clusters C & D), and (4) that it is not due to the effects of a substance or illness (Cluster E). In the absence of DSM-IV including a severity criterion for insomnia, only those who complained of sleep problems for at least three nights a week were included (Morin, 1993; World Health Organization, 1992). In the insomnia group, nine participants (31%) reported having received treatment for insomnia in the past. In addition, three participants (10%) reported having, in the past, received psychological treatment for PTSD ðn ¼ 1Þ and anxiety ðn ¼ 2Þ: In the good sleeper group, three participants (10%) reported having, in the past, received psychological treatment for relationship problems ðn ¼ 2Þ and depression ðn ¼ 1Þ: 2.2. Procedure The IDI was administered by a qualified clinical psychologist and participants then completed the Beck Anxiety Inventory (BAI; Beck, Epstein, Brown, & Steer, 1988), Beck Depression Inventory (BDI; Beck, Ward, Mendelson, Mock, & Erbaugh, 1961), and the Penn State Worry Questionnaire (PSWQ; Meyer, Miller, Metzger, & Borkovec, 1990). Participants were asked to rate the severity of their insomnia in the past month (Response Scale: 0 ‘Not at all disturbing and/or disabling, 8 ‘Severely disturbing/disabling’) and how concerned they have been about their sleep problem in the past month (Response scale: 0 ‘Not at all’, 8 ‘Very much’). These variables will be known as ‘sleep disturbance severity’ and ‘sleep concern’, respectively. Participants were also asked to estimate the probability that they will sleep badly on any one night (Response Scale: 0 ‘It is very unlikely I will sleep badly’ to 10 ‘It is very likely I will sleep badly’) and to estimate the cost of sleeping badly on any one night’ (Response Scale: 0 ‘There is no cost, sleeping badly does not effect me at all’ to 10 ‘Sleeping badly is extremely costly and significantly disrupts my life). Theoretically, if catastrophizing is associated with increased perception of threat (Davey & Levy, 1998; Kendall & Ingram, 1987), it should also be associated with increased cost and probability estimates. The Catastrophizing Interview employed broadly followed the procedures of Vasey and Borkovec (1992) and Davey and Levy (1998), except that the lead question was specific to the consequences of sleep disturbance. In previous studies, all conducted in the context of chronic worry, the lead question was asked with reference to any current worry. In our procedure participants were asked ‘What is it that worries you about nights when you have problems getting to sleep?’. Where A
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
15
Table 1 An example of the catastrophizing steps generated by an insomnia patient and the likelihood rating for each response Catastrophizing step
Likelihood rating
I will feel bad the following morning I won’t cope at work The back log at work will build because I cannot concentrate I will get drowsy and irritable I won’t be nice company I’ll get into trouble at work—my boss will shout at me My job will be on the line I’ll be unemployed I will have a low self-esteem
10 9 7 8 9 6 7 9 9
Note: Likelihood rating ¼ 0 ‘not at all likely’, 10 ‘extremely likely’.
denotes the answer to the latter question the participant was asked ‘What is it that worries you about A?’. Where B denotes the answer the participant gave to this second question the participant was asked ‘What is it that worries you about B’. Following Davey and Levy (1998), this questioning process continued until either the participant said they had no more answers or until they repeated a similar answer three times. Participants were then asked to rate, for each response, the likelihood that it would actually occur (0 ‘not at all likely’, 10 ‘extremely likely’). This rating will be referred to as the ‘likelihood rating’. Note that immediately before and immediately after the procedure participants were asked to rate their anxiety (0 ‘Not at all anxious’ to 10 ‘Extremely anxious’) and discomfort (0 ‘No discomfort’ to 10 ‘Extreme discomfort’). An example of the catastrophizing steps generated by one insomnia patient and the likelihood rating for each response is presented in Table 1. At the end of the experiment the participants were debriefed and paid an honorarium.
3. Results 3.1. Participant characteristics Table 2 presents the mean scores for participant characteristics, sleep characteristics, and measures of psychopathology. The groups did not differ on sex composition [insomnia—18 females, 12 males; good sleepers—19 females, 11 males, w2 (1, N ¼ 60) ¼ 0:07; ns] or age. The average duration of the sleep disturbance for the insomnia group was 6.3 years. Sleep onset latency for the past month was longer and total sleep time shorter for the insomnia group compared to the good sleepers. The insomnia group scored higher on sleep disturbance severity, sleep concern, probability, cost, BAI, and the PSWQ. The insomnia group was marginally significantly higher on BDI compared with the good sleeper group ðp ¼ 0:07Þ:
16
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
Table 2 Mean values for participant characteristics, sleep characteristics and measures of psychopathology
Age Length of insomnia Sleep onset latency Total sleep time Sleep disturbance severity Sleep concern Probability Cost BAI BDI PSWQ
Insomnia
Good sleeper
tð58Þ
20.4 (4.7) 6.3 (4.7) 58.2 (45.9) 6.3 (1.2) 3.6 (1.3) 3.7 (1.8) 5.5 (1.7) 5.5 (1.8) 13.6 (9.6) 10.3 (7.3) 57.4 (15.1)
22.3 (8.9) N/A 15.3 (11.0) 7.5 (1.1) 0.2 (0.5) 0.1 (0.3) 1.5 (2.0) 3.9 (2.4) 7.3 (7.2) 7.0 (6.6) 41.2 (14.9)
1.03 N/A 4.98nn 4.15w 13.8w 10.9w 8.3w 2.8nn 2.90nn 1.83 4.17w
Note: Standard deviations appear in parentheses. Age and length of insomnia in years. Sleep onset latency¼for the past month, in minutes. Total sleep time ¼ for the past month, in hours. Sleep disturbance severity ¼ 0 ‘Not at all disturbing and/or disabling, 8 ‘Severely disturbing/disabling’. Sleep concern ¼ 0 ‘Not at all’, 8 ‘Very much’. Probability ¼ 0 ‘It is very unlikely I will sleep badly’ to 10 ‘It is very likely I will sleep badly’. Cost ¼ 0 ‘There is no cost, sleeping badly does not effect me at all’ to 10 ‘Sleeping badly is extremely costly and significantly disrupts my life. BAI ¼ Beck Anxiety Inventory. BDI ¼ Beck Depression Inventory. PSWQ ¼ Penn State Worry Questionnaire. nn po0:01: w po0:001:
3.2. Catastrophizing interview Immediately prior to the catastrophizing interview the insomnia group were more anxious, tð58Þ ¼ 7:64; po0:0001; and reported more discomfort, tð58Þ ¼ 5:01; po0:0001; compared to the good sleeper group (see left side of Figs. 1a and b for mean values). The mean number of catastrophizing steps was 6 ðSD ¼ 2:6Þ for the insomnia group and 1.8 ðSD ¼ 1:9Þ for the good sleeper group. The change in emotionality ratings from the beginning to the end of the catastrophizing interview was examined with two repeated measures ANOVAs. For the first ANOVA, Diagnosis (Insomnia vs. Good sleeper) was entered as a between subjects factor and Anxiety (pre-interview vs. post-interview) was entered as a within subjects factor. The Diagnosis main effect was significant, F ð1; 58Þ ¼ 142:51; po0:001; such that overall the insomnia group had higher anxiety scores than the good sleeper group. The Anxiety main effect was significant, F ð1; 58Þ ¼ 42:34; po0:001; such that overall ratings of anxiety increased after the catastrophizing interview relative to before it. The interaction was also significant, F ð1; 58Þ ¼ 42:34; po0:001: As depicted in Fig. 1a, the insomnia group rated their anxiety to be higher than the good sleeper group both before, tð58Þ ¼ 7:64; po0:001; and after, tð58Þ ¼ 13:10; po0:001; the interview. However, the insomnia group rated their anxiety to be significantly higher after the interview relative to before the interview, tð58Þ ¼ 7:31; po0:001: This same difference was not significant for the good sleeper group.
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
17
Fig. 1. (a) Mean values for anxiety ratings before and after the ‘catastrophizing interview’. (b) Mean values for discomfort ratings before and after the ‘catastrophizing interview’.
For the second ANOVA, Diagnosis (Insomnia vs. Good sleeper) was entered as a between subjects factor and Discomfort (pre-interview vs. post-interview) was entered as a within subjects factor. The Diagnosis main effect was significant, F ð1; 58Þ ¼ 53:17; po0:001; such that overall the insomnia group had higher discomfort scores than the good sleeper group. The Discomfort main effect was significant, F ð1; 58Þ ¼ 32:36; po0:001; such that ratings of discomfort increased after the catastrophizing interview relative to before it. The interaction was also significant, F ð1; 58Þ ¼ 19:95; po0:001: As depicted in Fig. 1b, the insomnia group rated their discomfort to be higher than the good sleeper group both before,
18
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
tð58Þ ¼ 5:01; po0:001; and after, tð58Þ ¼ 7:89; po0:001; the interview. However, the insomnia group rated their discomfort to be significantly higher after the interview relative to before the interview, tð58Þ ¼ 5:31; po0:001: This same difference was not significant for the good sleeper group. Eleven people in the good sleeper group did not generate any catastrophizing steps. Not including these participants, the average likelihood rating for the insomnia group was higher (M ¼ 6:52; SD ¼ 1:9) than the good sleeper group (M ¼ 5:0; SD ¼ 2:6), tð47Þ ¼ 2:32; po0:05: 3.3. Correlational analyses Table 3 presents correlational analyses between the total number of catastrophe steps generated and the average likelihood rating with sleep and psychopathology variables. As the relationship between variables was expected to differ according to diagnosis, these analyses were carried out separately for the insomnia and good sleeper groups. For the insomnia group, the number of catastrophe steps generated was significantly positively correlated with sleep concern and PSWQ score. For the insomnia group, the average likelihood rating was significantly negatively correlated with length of insomnia and significantly positively correlated with cost and
Table 3 Correlational analysis between number of catastrophizing steps generated, average likelihood ratings, sleep and psychopathology measures Good sleeper group
Insomnia group
Number of steps Average likelihood Number of steps Average likelihood ðn ¼ 30Þ ðn ¼ 19Þa ðn ¼ 30Þ ðn ¼ 30Þ Length of insomnia N/A Sleep disturbance severity 0.17 Sleep concern 0.01 Probability 0.12 Cost 0.25 BAI 0.08 BDI 0.30 PSWQ 0.16 Anxiety rating Discomfort rating a
N/A 0.17 0.12 0.31 0.31 0.43 0.13 0.14
0.01 0.07 0.42n 0.21 0.18 0.30 0.33 0.40n
0.50nn 0.39 0.45 0.38n 0.54n 0.12 0.01 0.19
n ¼ 11 did not generate any catastrophizing steps and so did not make likelihood ratings. Length of insomnia in years. Sleep disturbance severity ¼ 0 ‘Not at all disturbing and/or disabling, 8 ‘Severely disturbing/disabling’. Sleep concern ¼ 0 ‘Not at all’, 8 ‘Very much’. Probability ¼ 0 ‘It is very unlikely I will sleep badly’ to 10 ‘It is very likely I will sleep badly’. Cost ¼ 0 ‘There is no cost, sleeping badly does not effect me at all’ to 10 ‘Sleeping badly is extremely costly and significantly disrupts my life. BAI ¼ Beck Anxiety Inventory. BDI ¼ Beck Depression Inventory. PSWQ ¼ Penn State Worry Questionnaire. n po0:05: nn po0:01:
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
19
probability estimates. There were no significant correlations for the good sleeper group.
4. Discussion The aim of the present study was to borrow an experimental paradigm from the anxiety literature to investigate the presence and role of catastrophizing in the maintenance of insomnia. Three hypotheses were tested. The first hypothesis was that patients with insomnia would generate more catastrophizing steps than good sleepers to the question ‘‘What is it that worries you about nights when you have problems getting to sleep?’’. This hypothesis was supported as patients with insomnia produced significantly more catastrophizing steps than good sleepers. This finding is consistent with previous studies implicating excessive worry about sleep as characteristic of insomnia (Lichstein & Rosenthal, 1980; Espie et al., 1989; Harvey, 2000), and with Borkovec’s (1982) finding that patients with insomnia tend to be anxious and obsessively worrisome. Further, the findings empirically validate observations in both the clinical (Morin, 1993; Perlis et al., 2000) and self-help (Sharp, 2001) literatures that patients with insomnia have a tendency to catastrophize about their sleep disturbance. Finally, Borkovec et al. (1983) have previously noted several parallels between insomnia and the anxiety disorders. The current study adds to this list by highlighting that like individuals with anxiety (Kendall & Hollon, 1989), individuals with insomnia exhibit a tendency to catastrophize. The second hypothesis was that patients with insomnia would report an increase in anxiety and discomfort over the course of the catastrophizing interview. In support, when anxiety and discomfort ratings were examined, ratings made by the insomnia group increased significantly from immediately before to immediately following the catastrophizing interview. This same difference was not significant for the good sleeper group. The increase in anxiety and discomfort following catastrophizing for the patients with insomnia suggests that the procedure leads to activation of threatrelated material (Vasey & Borkovec, 1992). If this paradigm captures something of the internal dialogue during the pre-sleep period when difficulty falling asleep is experienced, one could speculate that the observed increase in affect associated with catastrophizing could contribute to the maintenance of insomnia because increasing anxiety and discomfort are unlikely to be conditions conducive to optimal sleeponset (Harvey, 2002). Espie (2002) highlighted the importance of a state of minimal affective load for optimal sleep-onset. This paradigm may also capture something of the internal dialogue during the day. If so, the observed increase in affect associated with catastrophizing would create cognitive load (Sarason, 1984; Sarason, Sarason, Keefe, Hayes, & Shearin, 1986) that could contribute to deficits in memory, attention, and concentration, which if attributed to lack of sleep are likely to increase pre-occupation with and anxiety about getting enough sleep (Harvey, 2002). This may be one of the pathways by which daytime processes serve to exacerbate nighttime sleep problems.
20
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
The third hypothesis tested was that the insomnia group would rate the likelihood that the catastrophizing steps generated might actually occur higher than the good sleeper group. While patients with insomnia perceived the catastrophizing steps they generated to be more likely to occur than the good sleepers. That is, although the evidence was consistent with the third hypothesis, baseline differences in anxiety and discomfort appear to have contributed significantly to the inflated likelihood ratings that characterized the insomnia group. This latter finding makes intuitive sense; the likelihood estimations are highly likely to be influenced by the extent to which the person is feeling comfortable and secure at that time point. However, it should be noted that a caveat on this finding is that the 11 good sleepers who were not able to generate any catastrophe steps, and did not generate a likelihood rating, were not included in this analysis. It is therefore likely that this may have skewed the good sleeper group. A number of interesting results emerged from the correlational analyses. For the insomnia group, the more catastrophizing steps generated the greater the amount of sleep concern and the higher the PSWQ score. These results suggest that the tendency to catastrophize among patents with insomnia is associated with a specific sleep outcome, concern about sleep, and with a non-specific outcome, the general tendency to worry. The shorter the duration of the insomnia the higher the ratings of the likelihood that the catastrophes generated might actually occur. There are two possibilities here. First, this finding suggests that over time, while individuals with insomnia continue to catastrophize, they become more realistic about the likelihood of the catastrophe actually occurring. Second, perhaps catastrophizing plays a more important role during the early stages and becomes less relevant when the insomnia becomes chronic. Suggestive of a mechanism by which catastrophizing may contribute to the maintenance of insomnia, the higher the average likelihood rating the higher the estimated probability of sleeping badly and the higher the estimated cost of sleeping badly. As we would expect, no correlations reached significance for the good sleeper group. It is interesting to note that the number of catastrophizing steps generated in this study by the insomnia group ðM ¼ 6:0Þ was lower than the number reported by Vasey and Borkovec ðM ¼ 14:2Þ and more akin to the results of Davey and Levy (M ¼ 8:1—experiment 1, M ¼ 6:5—experiment 2). There are a number of possible accounts for this observation. First, before starting the catastrophizing interview Vasey and Borkovec (1992) administered a practice trial that involved demonstrating the stepwise procedure. This component was not part of the procedure in the present study or in the Davey and Levy (1998) study. Prior exposure to the procedure may have increased the fluency of Vasey and Borkovec’s participants. Second, Vasey and Borkovec (1992) recruited chronic worriers. It is possible that such participants catastrophize more than insomnia patients. With a view to furthering research on catastrophizing in insomnia, several methodological points require consideration. It should be noted that the sample was based on a university population. As such, it is not necessarily representative of a general practitioner-referred insomnia population. It should be noted, however, that all participants met strict DSM-IV criteria for primary insomnia and that the
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
21
average duration of the insomnia was 6.3 years. Second, although the sample was not objectively assessed for comorbid sleep disorders (e.g., obstructive sleep apnea and narcolepsy) the IDI included a section designed to carefully screen for the key symptoms of comorbid sleep disorders. None of the participants responded positively to any of these screening questions. Considerations directly relating to the catastrophizing interview are threefold. First, it should be noted that the experimenter who administered the catastrophizing interview (AH) was not blind to diagnosis. While the questioning method and stop rules were clearly specified prior to testing and carefully adhered to during testing, this potential limitation should be corrected in future research. Second, careful consideration should be given to the opening phrase of the catastrophizing interview. Subtle changes in wording such as ‘What is it that worries you on nights that you cannot sleep?’ may impact responses. In addition, variations in the opening phrase may help illuminate different aspects of catastrophizing. As the nature of daytime processes in insomnia are increasingly of interest (Harvey, 2002), it may be illuminating to adopt an opening phrase such as: ‘What is it that worries you about your daytime activities following a night of poor sleep’. Finally, prior to the catastrophizing interview the insomnia group reported higher anxiety and distress relative to the good sleeper group. Miller and Chapman (2001) have drawn attention to the error introduced by correcting for baseline differences between groups by conducting analysis of covariance when the design does not involve random allocation to groups. To ensure that the greater number of catastrophizing steps generated by the insomnia group, relative to the good sleeper group, are not solely accounted for by baseline differences, future research should compare an insomnia group to a high anxious good sleeper group. Another important issue is that the association between catastrophizing and insomnia does not necessarily imply a causal relationship. Freedman and Sattler (1982) have argued that excessive cognitive activity is epiphenomenal to sleeplessness. An assumption of this study is that responses to the catastrophizing interview conducted in a laboratory setting reflect or parallel processes that operate in the ‘real world’. Naturalistic observational studies of catastrophizing are necessary to shed light on the validity of this assumption. Of course, this study was heavily based on theory and research relating to GAD. While GAD is commonly comorbid with insomnia and the two disorders have many common features (Borkovec et al., 1983), researchers should remain ‘on the look-out’ for differences across the two disorders. In conclusion, the present study constitutes an initial empirical attempt to capture the catastrophizing process in the context of insomnia. The method employed appears to have been useful for observing one aspect of the thought processes/ internal dialogue characteristic of patients with insomnia. As predicted, we found that the insomnia patients generated more potential catastrophes about the consequences of not sleeping relative to good sleepers. For the insomnia group, but not the good sleepers, catastrophizing was associated with increased anxiety and discomfort. In a large case series, Perlis et al. (2000), as part of a multi-component treatment package, employed Barlow’s (1992) decastrophization method. This involved exploring, with the patient, the discrepancy between their certainty of
22
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
negative outcome and the real probability of negative outcome. The results of the present study provides endorsement for the inclusion of such interventions in treatments for insomnia.
Acknowledgements The authors are grateful to Dr. Lars-Gunnar Lundh for helpful comments on the manuscript. This research was supported by the Wellcome Trust (Grant reference number—065913).
References American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Barlow, D. H. (1992). Cognitive-behavioral approaches to panic disorder and social phobia. Bulletin of the Menninger Clinic, 56. Beck, A. T., & Emery, G. D. (1985). Anxiety disorders and phobias: A cognitive perspective. New York: Basic Books. Beck, A. R., Epstein, N., Brown, G., & Steer, R. A. (1988). An inventory for measuring clinical anxiety: Psychometric properties. Journal of Consulting and Clinical Psychology, 56, 893–897. Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory measuring depression. Archives of General Psychiatry, 4, 561–571. Borkovec, T. D. (1982). Insomnia. Journal of Consulting and Clinical Psychology, 50, 880–895. Borkovec, T. D., Robinson, E., Pruzinsky, T., & De Pree, J. A. (1983). Preliminary exploration of worry: Some characteristics and processes. Behaviour Research and Therapy, 21, 9–16. Davey, G. C. L., & Levy, S. (1998). Catastrophic worrying: Personal inadequacy and a perseverative iterative style as features of the catastrophizing process. Journal of Abnormal Psychology, 107, 576–586. Edinger, J. D., Wohlgemuth, W. K., Radtke, R. A., Marsh, G. R., & Quillian (2001). Does cognitivebehavioural insomnia therapy alter dysfunctional beliefs about sleep? Sleep, 24, 591–599. Espie, C. A. (1991). The psychological treatment of insomnia. England: Wiley. Espie, C. A. (2002). Insomnia: Conceptual issues in the development, persistence, and treatment of sleep disorder in adults. Annual Review of Psychology, 53, 215–243. Espie, C. A., Brooks, D. N., & Lindsay, W. R. (1989). An evaluation of tailored psychological treatment of insomnia. Journal of Behaviour Therapy and Experimental Psychiatry, 20, 143–153. Espie, C. A., Inglis, S. J., Harvey, L., & Tessier, S. (2000). ‘Insomniacs’ attributions. Psychometric properties of the dysfunctional beliefs and attitudes about sleep scale and the sleep disturbance questionnaire. Journal of Psychosomatic Research, 48, 141–148. Fichten, C. S., Libman, E., Creti, L., Amsel, R., Sabourin, S., Brender, W., & Bailes, S. (2001). Role of thoughts during nocturnal awake times in the insomnia experience of older adults. Cognitive Therapy and Research, 25, 665–692. Fins, A. I., Edinger, J. D., Sullivan, R. J., Marsh, G. R., Dailey, D., Hope, T. V., Young, M., Shaw, E., & Vailas, D. (1996). Dysfunctional cognitions about sleep among older adults and their relationship to objective sleep findings. Sleep Research, 25, 242. Ford, D., & Kamerow, D. (1989). Epidemiologic study of sleep disturbances and psychiatric disorders. Journal of the American Medical Association, 262, 1479–1484. Freedman, R. R., & Sattler, H. L. (1982). Physiological and psychological factors in sleep-onset insomnia. Journal of Abnormal Psychology, 91, 380–389. Harvey, A. G. (2000). Pre-sleep cognitive activity in insomnia: A comparison of sleep-onset insomniacs and good sleepers. British Journal of Clinical Psychology, 39, 275–286.
A.G. Harvey, E. Greenall / J. Behav. Ther. & Exp. Psychiat. 34 (2003) 11–23
23
Harvey, A. G. (2002). A cognitive model of insomnia. Behaviour Research and Therapy, 40, 869–893. Kendall, P. C., & Hollon, S. D. (1989). Anxious self-talk: Development of the anxious self-statements questionnaire (ASSQ). Cognitive Therapy and Research, 13, 81–93. Kendall, P. C., & Ingram, R. E. (1987). The future for cognitive assessment of anxiety: Let’s get specific. In L. Michaelson, & L. M. Ascher (Eds.), Anxiety and stress disorders: Cognitive-behavioural assessment and treatment. New York: Guildford Press. Lichstein, K. L., & Rosenthal, T. L. (1980). Insomniacs’ perceptions of cognitive versus somatic determinants of sleep disturbance. Journal of Abnormal Psychology, 89, 105–107. Lundh, L.-G. (1998). Cognitive-behavioural analysis and treatment of insomnia. Scandinavian Journal of Behaviour Therapy, 27, 10–29. Meyer, T. J., Miller, M. L., Metzger, R. L., & Borkovec, T. D. (1990). Development and validation of the Penn State Worry Questionnaire. Behaviour Research and Therapy, 28, 487–495. Miller, G. A., & Chapman, J. P. (2001). Misunderstanding analysis of covariance. Journal of Abnormal Psychology, 110, 40–48. Morin, C. M. (1993). Insomnia: Psychological assessment and management. New York: Guildford Press. Morin, C. M., Stone, J., Trinkle, D., Mercer, J., & Remsberg, S. (1993). Dysfunctional beliefs and attitudes about sleep among older adults with and without insomnia complaints. Psychology and Aging, 8, 463–467. Perlis, M., Aloia, M., Millikan, A., Boehmler, J., Smith, M., Greenblatt, & Giles, D. (2000). Behavioral treatment of insomnia: A clinical case series study. Journal of Behavioral Medicine, 23, 149–161. Perlis, M. L., Giles, D. E., Mendelson, W. B., Bootzin, R. R., & Wyatt, J. K. (1997). Psychophysiological insomnia: The behavioural model and a neurocognitive perspective. Journal of Sleep Research, 6, 179–188. Sarason, I. G. (1984). Test anxiety, stress, and social support. Journal of Personality, 49, 101–114. Sarason, I. G., Sarason, B. R., Keefe, D. E., Hayes, B. E., & Shearin, E. N. (1986). Cognitive interference: Situational determinants and traitlike characteristics. Journal of Personality and Social Psychology, 51, 215–226. Sharp, T. (2001). The good sleep guide. Australia: Penguin. Startup, H. M., & Davey, G. C. L. (2001). Mood as input and catastrophic worrying. Journal of Abnormal Psychology, 110, 83–96. Vasey, M. W., & Borkovec, T. D. (1992). Catastrophizing assessment of worrisome thoughts. Cognitive Therapy and Research, 16, 505–519. World Health Organization. (1992). The ICD-10 classification of mental and behavioural disorder: Diagnostic criteria for research (10th revision). Geneva: Author.