BEHAVIOUR RESEARCH AND THERAPY
PERGAMON
Behaviour Research and Therapy 36 (1998) 1131±1142
Intrusive memories and depression in cancer patients Chris R. Brewin a, *, Maggie Watson b, Siobhan McCarthy a, Philippa Hyman a, David Dayson c a
Cognition, Emotion, and Trauma Group, Department of Psychology, Royal Holloway, University of London, Egham, Surrey TW20 0EX, UK b Department of Psychological Medicine and Institute of Cancer Research, Royal Marsden Hospital, Sutton, Surrey, UK c Department of Psychiatry, University of Southampton, Southampton, UK Received 12 February 1998
Abstract Matched samples of depressed and nondepressed cancer patients were interviewed about past life events, particularly experiences of death and illness. They identi®ed and described any spontaneous intrusive visual memories they had experienced in the past week corresponding to these events. About one quarter reported such memories and, as predicted, the majority of intrusive memories concerned illness, injury and death. The mean levels of intrusion and avoidance were equivalent to patients with post-traumatic stress disorder. Consistent with prediction, depressed patients reported signi®cantly more intrusive memories than controls, and described the memories as typically beginning with or being exacerbated by the onset of depression. Greater numbers of intrusive memories were associated with more maladaptive coping, and greater avoidance with de®cits in autobiographical memory functioning. # 1998 Elsevier Science Ltd. All rights reserved.
1. Introduction Recent research has found that depressed patients, like patients with post-traumatic stress disorder, report high levels of intrusive visual memories of stressful life events. These intrusive memories may involve incidents of childhood abuse (Kuyken and Brewin, 1994a) and of more recent stressors such as the death of loved ones (Brewin et al., 1996c). Moreover, within depressed samples, increased levels of intrusive memories are associated with increased levels of depressive cognition (Kuyken and Brewin, 1995, in press). However, it is not known whether * Corresponding author. Tel.: +44-1784-443712; Fax: +44-1784-434347; E-mail:
[email protected] 0005-7967/98/$19.00 # 1998 Elsevier Science Ltd. All rights reserved. PII: S 0 0 0 5 - 7 9 6 7 ( 9 8 ) 0 0 0 8 4 - 9
1132
C.R. Brewin et al. / Behaviour Research and Therapy 36 (1998) 1131±1142
these repetitive memories are present at an equally high frequency before an individual becomes depressed, or whether the onset of depression is associated with an increase in their frequency. None of the studies so far carried out have included nondepressed controls, so it is also possible that intrusive memories are temporarily triggered by concurrent stimuli such as life stresses and not by depression itself. In this investigation we therefore compared intrusive memories in matched samples of depressed and nondepressed individuals with the common stressor of a cancer diagnosis. We also sought to replicate our previous ®nding that the avoidance of repetitive intrusive memories is associated with problems in autobiographical memory functioning that are a characteristic feature of depressive thinking (Kuyken and Brewin, 1995). Beck's (Beck et al., 1979) cognitive theory of depression proposed that negative life events activate latent schemes in memory that represent an accumulation of relevant past experience. The activation of schemes containing negative information about the self and the world, Beck suggested, in¯uences the style and content of depressive thinking, leading to more severe and prolonged depression. Research in social cognition has emphasized, however, that representations of the self in memory include speci®c autobiographical episodes as well as global trait judgements (Smith, 1990; Klein and Loftus, 1993). Studies of anxious and dysphoric individuals have con®rmed that self-concept discrepancies are linked to autobiographical episodes. Thus, priming these individuals with trait words relevant to areas in which they feel de®cient enhances the accessibility of speci®c negative memories (Strauman, 1992). Empirical con®rmation that depression is associated, not only with the activation of generalized negative beliefs, but also with the intrusion of highly speci®c autobiographical memories, is consistent with laboratory research and promises to yield important insights into the origins and formation of depressive thinking. Our previous work indicated that the avoidance of intrusive memories of childhood abuse episodes by depressed patients is related to a more global diculty in retrieving speci®c autobiographical memories to positive and negative cues (Kuyken and Brewin, 1995). This problem of overgeneral recall is frequently found in depressed and suicidal patients and may in¯uence the course of depression via its impact on problem-solving (Williams and Broadbent, 1986; Williams and Scott, 1988; Williams, 1992). The present study provided an opportunity to test whether intrusive memories related to dierent kinds of event might predict autobiographical memory functioning in depressed patients. We also found that higher levels of intrusive memories of childhood abuse were related to depressive cognitions such as a more negative attributional style and lower self-esteem, and to a more avoidant coping style (Kuyken and Brewin, in press). This latter association is of interest in light of a recent theory concerning the processing of traumatic memories. Brewin et al. (1996b) suggested that, following a trauma, memories of the experience may eventually cease to intrude either because they have been successfully processed, or because they have been prematurely inhibited. Premature inhibition, brought about perhaps as a result of cognitive avoidance, would leave the memories vulnerable to reactivation at a future time. Brewin et al. further proposed that an episode of depression might be a factor leading to the reactivation of traumatic memories that had been incompletely processed. In depressed patients, levels of intrusion and avoidance of these memories, as measured with standardized instruments such as the Impact of Event Scale (Horowitz et al., 1979) were
C.R. Brewin et al. / Behaviour Research and Therapy 36 (1998) 1131±1142
1133
equivalent to those found in patients with post-traumatic stress disorder (Kuyken and Brewin, 1994a; Brewin et al., 1996c). This permits an inference that the intrusive memories of our depressed samples were at an abnormal level, but without the proper control group this conclusion remains weak. A relatively stringent test of the hypothesis that depression is associated with an increase in the accessibility of speci®c autobiographical memories would involve comparing depressed and nondepressed samples who had both been exposed to similar types and degree of stress, so that the eects on memory of current environmental circumstances were matched as closely as possible. Two other methodological controls would give greater con®dence in these ®ndings. First, patients with comorbid post-traumatic stress disorder, an alternative source of intrusive memories, should be excluded from the depressed sample. Second, it would be necessary to ask the depressed group whether the onset of depression was associated with the start or with an exacerbation of intrusive memories, to guard against the possibility that they had always had high levels of such memories, even before they became depressed. Cancer patients appeared to be a particularly suitable group to study for a number of reasons. First, a cancer diagnosis is well known to be associated with depression (Spiegel, 1996). Also, since cancer is a major cause of mortality, many individuals will have relevant past experiences of the illness or death of close relatives or friends. Finally, cancer treatment is complicated, demanding, and timeconsuming, providing many opportunities for cueing recall of autobiographical memories. It would be of considerable interest to know if depressed cancer patients are repeatedly exposed to memories of speci®c prior experiences involving cancer. Access to information in memory involving pain, illness and death of their loved ones might, for example, contribute to negative expectations concerning their current treatment and quality of life, exacerbating helplessness and hopelessness, and leading to more anxious preoccupation with their illness, and more avoidance coping. Similarly, constantly replaying their own experiences of cancer diagnosis and treatment might in some instances worsen depression. Intrusive memories are not exclusively associated with depression, but may occur in nonclinical samples (Brewin et al., 1996a). Thus we expected a certain number of nondepressed cancer patients to report intrusive memories. However, the major prediction of this study was that depressed patients would report signi®cantly more intrusive memories than controls, even when dierences in cancer variables, other unrelated life events and childhood adversity were controlled. Patients were expected to report that onset of depression had been associated with an onset of intrusive memories or with an exacerbation of existing intrusive memories. Following cognitive theory (Beck et al., 1979), it was further expected that the majority of these intrusive memories would be related to relevant past experiences, such as witnessing illness and death. Following previous work (Watson et al., 1994; Kuyken and Brewin, in press), intrusive memories were expected to be associated with poor coping, speci®cally anxious preoccupation, and helpless and hopelessness. Since depression has been found to be associated with avoidance coping in depressed psychiatric patients (Kuyken and Brewin, 1994b) but not in depressed cancer patients (Watson et al., 1994), no speci®c predictions were made concerning this variable. It was further predicted that, in line with the ®ndings from studies of autobiographical memory in depressed psychiatric patients, overgeneral memories would be
1134
C.R. Brewin et al. / Behaviour Research and Therapy 36 (1998) 1131±1142
more common among depressed cancer patients, and that overgeneral memories would be associated with more attempts to avoid intrusive memories of upsetting events.
2. Method 2.1. Patients A consecutive series of 740 outpatients diagnosed with cancer attending the Royal Marsden Hospital, London and Sutton branches, was screened for depression using the Hospital Anxiety and Depression Scale. Patients who had not yet received a diagnosis of cancer or who were unsure of their diagnosis were excluded. Following previous research (Greer et al., 1992), patients who scored 8 or above on the depression sub-scale (8.78%) were categorized as depressed and selected for the second, interview stage of the study. For each depressed patient interviewed, a control patient was selected from among other attenders at the same clinic. Controls were individually matched on age, sex, type of cancer, and stage of disease with the depressed group, but scored 4 or below on the depression subscale. Where more than one patient was suitable as a control, one patient was selected at random. 65 depressed patients and 65 controls were interviewed. 96 were female (73.8%) and 34 male (26.2%). Patients ranged in age from 24 to 81 years, with a mean age of 54 years (s.d. 13.3 years). 74 patients had breast cancer (56.92%); 23 had gastro-intestinal cancer (17.69%); 23 had lymphomas (17.69%); 4 had gynaecological cancer (3%) and 6 had lung cancer (4.62%). Patients were categorized into three groups depending upon the size and spread of their cancer. 38 patients had localized disease con®ned to one primary site (29.23%), 20 patients had locoregional disease (15.38%), and 72 patients had metastatic disease (55.38%). Patients had been diagnosed an average of 46.55 months earlier. Of 20 patients who declined to participate at the screening stage (2.7%), 7 did not have enough time, 2 felt too ill, 5 were too anxious, or wanted to forget they had cancer, and 6 did not specify a reason for their refusal. Of the 8 patients who declined to participate at the interview stage, 3 did not want to talk about their illness, 3 did not have the time to participate, 1 felt too ill to do an interview and 1 patient's husband objected to the interview. 7 patients died between screening and interview. 2.2. Measures 2.2.1. Hospital Anxiety and Depression Scale (HADS: Zigmond and Snaith, 1983) The HADS is a 14-item self-report scale which was developed speci®cally for the measurement of depression and anxiety in physically ill populations. The HADS has been validated for a range of medical diagnoses (Mayou and Hawton, 1986; Moorey et al., 1991), and the subscale scores of depression and anxiety have been validated in cancer patients (Razavi et al., 1990).
C.R. Brewin et al. / Behaviour Research and Therapy 36 (1998) 1131±1142
1135
2.2.2. Structured clinical interview for DSMIII-R (SCID: American Psychiatric Association, 1987) The SCID is a widely used measure for assessing patients' level of depression. The interview was adapted to assess whether patients met DSMIV criteria for a diagnosis of major depressive episode. 2.2.3. Posttraumatic Stress Scale-interview version (Foa et al., 1993) This is a 17-item measure for the diagnosis of PTSD according to DSM-IIIR criteria. It has good reliability and validity, and excellent convergent validity with the SCID (Foa et al., 1993). 2.2.4. Mini-Mental Adjustment of Cancer Scale (Mini-MAC: Watson et al., 1994) The Mini-MAC is a 29-item self report scale based on the original 40-item version and is used to assess patients' coping responses to a cancer diagnosis. It has 5 subscales: helplessness± hopelessness, cognitive avoidance, ®ghting spirit, anxious preoccupation and fatalism. External validation has been established (Watson et al., 1994). The Mini-MAC was administered at screening and only for the ®rst part of the study, so that complete data are only available on a subset of interviewed patients. 2.2.5. Life events and memories interview Patients were administered a semi-structured interview covering a series of life events and stressors. They were ®rst asked whether any of their family or friends had died and how long ago this was. In the next section of the interview they were asked about the occurrence of 10 major life events consistently linked with the onset of depression (List of Threatening Experiences: Brugha et al., 1985). Patients were asked if any of these had happened to them at any time in the past or (for the depressed sample) in the 12 months preceding the current episode. An additional open-ended question about possible triggers of the current episode was included for the depressed sample. In the third section of the interview they were asked about childhood, including harsh discipline and unwanted sexual experiences, using questions extensively validated in previous research (e.g. Andrews et al., 1990). In the ®nal section of the interview patients were asked if they had noticed visual memories of any of these deaths, life events, or childhood experiences, or of any other negative event, spontaneously coming into their minds during the past week. To qualify, memories had to consist of a visual image of a speci®c scene that had actually taken place. General thoughts or worries were not included. Patients were asked to identify up to two intrusive memories, selecting the most intrusive if there were more than two. For each memory they were asked a series of questions designed to yield descriptive information about memory characteristics, including: (a) how long the memories lasted. Answers were coded using ®ve response categories (1 = seconds, 2 = minutes, 3 = up to an hour, 4 = several hours, 5 = constantly preoccupied); (b) how frequent the memories were. Answers were coded using two response categories (1 = once a week or less, 2 = several times a week or more); (c) how clear and vivid the memories were. Three response options were provided (1 = unclear, 2 = some detail, 3 = vivid); (d) whether strong physical sensations accompanied the memories, such as heart racing, sweating, trembling, nausea, headache, chills/¯ushes, and `butter¯ies in the stomach'. Answers were coded using two response categories (1 = no physical sensations, 2 = physical
1136
C.R. Brewin et al. / Behaviour Research and Therapy 36 (1998) 1131±1142
sensations present); (e) feeling of reliving versus experiencing the memory as having happened in the past. Two response options were provided (1 = reliving the experience, 2 = looking back at the past); (f) how distressing the memories were. Patients rated distress on a 10-point scale (0 = not distressing at all, 10 = extremely distressing). Vividness, physical sensations, feelings of reliving, and distress are all characteristic of traumatic memories (Brewin et al., 1996b). In addition patients in the depressed group were asked whether the intrusive memory was more frequent, less frequent, or unchanged in frequency, compared to the period before they became depressed. Patients also rated each memory on the Impact of Event Scale. 2.2.6. Impact of Event Scale (IES: Horowitz et al., 1979) The IES is a 15-item self-report scale which examines subjective distress as a result of a speci®c event. The scale has been used in past studies to assess the impact of life-threatening illness such as cancer (Kaasa et al., 1993; Kelly et al., 1995). The IES is divided into 2 subscales: intrusion and avoidance. The intrusion subscale measures the intrusiveness of the memories, and is related to the loss of voluntary control over the regulation of thoughts. The avoidance scale measures the extent memories are consciously suppressed from memory. Each item is rated on a four point scale, from not at all, rarely, sometimes to often. The IES intrusion subscale has an alpha coecient of 0.78 and avoidance subscale has an alpha of 0.82 (Corcoran and Fischer, 1987). 2.2.7. Autobiographical Memory Test (Williams and Broadbent, 1986) Subjects were given one minute to retrieve a speci®c personal memory to cue words describing positive and negative emotions. Five positive (happy, surprised, interested, successful and safe) and ®ve negative cue words (clumsy, angry, sorry, hurt and lonely) were printed on ®ve by three inch cards. Words were presented in a pseudo-random order: the ®ve negative and ®ve positive cue words were shued and then presented with positive and negative words alternating. If subjects did not give a speci®c memory they were prompted to do so (Can you think of a speci®c time Ð one particular event?), and the cumulative time taken to produce a speci®c memory was recorded. Following Williams and Broadbent (1986), subjects not producing a speci®c memory were allocated a time of 60s. Comparison with a second rater showed that speci®c and general memories could be quite reliably distinguished (inter-rater agreement on a sample of 200 memories was 91%, similar to that reported by Williams and Broadbent, 1986). 2.3. Procedure Outpatients attending follow-up clinics were asked if they would like to participate in a research study aimed at understanding why some patients experience depression. They were then asked if they were aware of their (cancer) diagnosis, and those who were unaware were excluded from the study. This ensured that new patients without a con®rmed diagnosis, or patients who may have been in denial of their illness would not be distressed by the questions which directly related to cancer. The aims of the study were then fully explained and patients completed the HADS.
C.R. Brewin et al. / Behaviour Research and Therapy 36 (1998) 1131±1142
1137
Patients who were identi®ed as being in the depressed group were then contacted and invited to participate in a later interview, either at the hospital or at their home. Interviews were audiotaped. The HADS questionnaire was repeated before the interview to ensure they still met the criteria for inclusion in the depressed group. Control patients were then selected for interview, matched on age, sex, type and stage of cancer with the depressed patients, using information from the hospital computer systems and medical ®les.
3. Results 3.1. Characteristics of the groups Of the depressed sample, only 28 met DSM-IV criteria for a major depressive episode. It was therefore decided to compare the nondepressed controls both with this group (the severely depressed) and with patients scoring more than 8 on the HAD depression scale but not meeting DSM-IV criteria (the mildly depressed). Prior to any analysis six patients who met diagnostic criteria for post-traumatic stress disorder were excluded from the sample (5 severely depressed and 1 control patient). The three groups did not dier signi®cantly in sex (w2(2) < 1), age (F(2,121) < 1), time since diagnosis (F(2,120) < 1), stage of illness (F(2,121) = 1.13, p>0.10), or type of cancer (w2(8) < 1). Nor did the groups dier in number of deaths experienced (F(2,121) = 1.86, p>0.10), number of other life events experienced (F(2,121) = 1.74, p>0.10), or experience of any childhood adversity (F(2,121) < 1). The groups did dier on HAD depression (control x=1.81, mildly depressed x=9.84, severely depressed x=10.70, F(2,121) = 357.4, p < 0.001). Post hoc tests (least signi®cant dierence) indicated that the controls diered from both depressed groups ( p < 0.05), who did not dier from each other. The groups did not dier on number of previous episodes of depression (F(2,93) = 1.62, p>0.10), and the two depressed groups did not dier on the length of the current depressive episode (F(1,34) < 1. The groups did dier on age at leaving full-time education (control x=15.7, mildly depressed x=16.3, severely depressed x=17.6, F(2,109) = 3.58, p < 0.05. Post hoc tests indicated that the severely depressed diered from the controls ( p < 0.05). However, owing to historical changes in the school leaving age, age at leaving full-time education is confounded with patient age, with a signi®cant negative correlation of r(112) = ÿ 0.36, p < 0.001. All the variables on which groups were to be compared were therefore checked for possibly confounding associations with age at leaving full-time education (excluding the separate eect of age). No such confounding associations were evident. 3.2. Number and characteristics of intrusive memories Twenty-nine patients (23%) reported an intrusive memory, of whom 12 reported at least one additional memory. In 8 of these 12 cases, the second memory was related to the same incident as the ®rst memory, and in the remaining 4 cases the second memory referred to a completely separate incident. Detailed information was collected on a total of 41 memories.
1138
C.R. Brewin et al. / Behaviour Research and Therapy 36 (1998) 1131±1142
The duration of the memories was as follows: 3/35 lasted seconds, 18/35 lasted minutes, 7/35 lasted up to an hour, 6/35 lasted several hours, and 1/35 was constantly present (6 patients were unable to answer or were not asked the question). 27/41 memories intruded once a week or less, and 14/41 intruded several times a week or more. Overwhelmingly the memories were vivid (38/41) rather than unclear or with some detail only (3/41 for these combined categories). 17/40 memories were accompanied by physical sensations (one patient was not asked the question). Patients experienced themselves as reliving 17/41 memories, and as looking back at the remaining 24/41 memories in the past. The mean level of distress was 6.72 (SD 2.7). The mean IES score of all intrusive memories was 42.03 (range 18±67).
3.3. Content of intrusive memories Twenty-four of the memories (59%) referred to the illness, injury, or death of a relative or friend. Of these 24, 11 (46%) referred to an episode involving death from cancer. This proportion approximated patients' experience of deaths from cancer among their relatives and friends expressed as a proportion of all deaths they had experienced (50%). A further 7 (17%) intrusive memories involved the patient's own experiences of having cancer, such as being given the diagnosis. Thus 76% of the intrusive memories were clearly related to illness, injury and death, and 44% speci®cally to cancer.
3.4. Intrusive memories and coping with cancer Table 1 shows the correlations between mean scores on the 5 subscales of the Mini-Mac measure and the total number of intrusive memories patients reported. Intrusive memories were signi®cantly positively associated with anxious preoccupation and helplessness/ hopelessness, and showed marginally signi®cant associations with cognitive avoidance and fatalism. Table 1 also shows that these associations were still signi®cant or increased in signi®cance after controlling for levels of depression. Table 1 Correlations between intrusive memories and coping with cancer Mini-MAC subscales
Number of intrusive memories
Number of intrusive memories (controlling for depression)
Anxious preoccupation Cognitive avoidance Fatalism Fighting spirit Helplessness and hopelessness
0.22* 0.19 0.21 0.00 0.34**
0.23* 0.24* 0.23* 0.07 0.24*
**
p < 0.01; *p < 0.05 (1-tailed).
C.R. Brewin et al. / Behaviour Research and Therapy 36 (1998) 1131±1142
1139
3.5. Intrusive memories and depression As predicted, despite the similarities in deaths experienced, other life events, and childhood adversity, the three groups diered markedly in the likelihood of reporting intrusive memories. Seven controls (11%), 12 mildly depressed (32%), and 10 severely depressed (43%) reported at least one intrusive memory. There was a highly signi®cant group dierence in the mean number of memories reported (control x=0.23, mildly depressed x=0.54, severely depressed x=0.78, F(2,121) = 5.57, p < 0.01). Post hoc tests indicated that the controls diered signi®cantly from the other two groups ( p < 0.05), who did not dier from each other. Participants in the depressed groups were more likely to report that their ®rst intrusive memory had begun with or been exacerbated by the onset of depression (13/20 instances) than that the memory had stayed the same (5/20 instances) or lessened in intensity with the onset of depression (2/20 instances). This distribution diered signi®cantly from chance, w2(2) = 9.71, p < 0.01. 3.6. Depression and autobiographical memory The three groups were compared on indices of autobiographical memory functioning, as shown in Table 2. The groups did not dier on latency to retrieve either positive memories (F(2,112) = 1.44, p>0.10) or negative memories (F(2,112) = 1.52, p>0.10). When the latency data were reanalyzed excluding trials on which no speci®c memory had been retrieved, these nonsigni®cant ®ndings remained essentially unchanged. Table 2 also shows the mean number of general memories retrieved to each kind of cue. There was a signi®cant group dierence in the total number of overgeneral memories (F(2,111) = 3.85, p < 0.05), post hoc tests showing the severely depressed to have more overgeneral memories than the controls ( p < 0.05). The groups did not dier signi®cantly in their response to positive cues (F(2,111) = 1.47, p>0.10), although the means were in the expected direction. There was however a signi®cant group eect for negative cues (F(2,111) = 4.28, p < 0.02). Post hoc tests once again showed that the severely depressed retrieved signi®cantly more overgeneral memories than the controls ( p < 0.05). Table 2 Autobiographical memory functioning and depression
Group Positive cue Controls Mildly depressed Severely depressed Negative cue Controls Mildly depressed Severely depressed
Latency to speci®c memory
Number of general ®rst memories
n
M
SD
M
SD
62 32 21
29.5 34.1 30.9
11.5 14.3 11.3
1.37 1.71 1.86
1.23 1.24 1.42
62 32 21
32.2 36.1 34.9
11.6 10.3 10.4
1.50 1.94 2.33
1.25 1.03 1.20
1140
C.R. Brewin et al. / Behaviour Research and Therapy 36 (1998) 1131±1142
Finally, within the sample of patients with intrusive memories, the number of overgeneral memories produced by each patient was correlated with their IES scores for the ®rst memory. Overgeneral memories were correlated r(24) = 0.05, 0.46, and 0.35 with intrusion, avoidance, and total IES score respectively. Avoidance was still signi®cantly associated with number of overgeneral memories after controlling for HAD depression scores, partial r = 0.47, p < 0.05 (2-tailed).
4. Discussion Speci®c, highly intrusive autobiographical memories were found in about one quarter of this selected sample of depressed patients and controls with a cancer diagnosis. The average amount of intrusion and avoidance of these memories reported on the IES was similar to that found by Kuyken and Brewin (1994a) and Brewin et al. (1996c) with depressed samples. As predicted, the likelihood of having intrusive memories was strongly related to being depressed. This was also the ®rst study to demonstrate that the depressed have more intrusive memories when compared to matched controls. The present study provided a particularly stringent test of the hypothesis, in that all patients had a cancer diagnosis, a highly stressful event that would in its own right be expected to be a potent trigger for recall of relevant prior episodes. The link between depression and intrusive memories was supported by retrospective report data, which indicated that the majority of patients associated the onset of depression with the beginning or with an exacerbation of intrusions. However, it is noteworthy that depression was not invariably accompanied by an exacerbation of these memories. Moreover, although we know of no speci®c data that challenge the ability of depressed people to make accurate self-reports of this kind, the reliability of these reports is unknown. The presence of intrusive memories was systematically related to indices of coping with cancer. In particular, after controlling for depression, the number of intrusive memories was associated with more anxious preoccupation, more cognitive avoidance, more fatalism, and more helplessness and hopelessness. This is of particular signi®cance because anxious preoccupation and helplessness±hopelessness are important responses to cancer and are strongly related to psychopathology (Watson et al., 1991, 1994). Past work with psychiatric samples has also found that other negative cognitions, such as low self-esteem, are related to intrusive memories (Kuyken and Brewin, in press). This study was also the ®rst to our knowledge to investigate autobiographical memory functioning in a sample of physically ill depressed patients. Like previous studies, we found that severely depressed patients retrieved more overgeneral memories to a series of positive and negative cue words than did controls. Of greater interest, perhaps, was the replication of the previous ®nding of an association between avoidance of intrusive memories and failure to retrieve speci®c memories on this task. Kuyken and Brewin (1995) were the ®rst to show this eect in a sample of depressed patients, restricting their investigation to intrusive memories of childhood abuse. The present study extends their ®ndings by suggesting that avoidance of any intrusive memory can lead to overgeneral recall. Conclusions about the direction of causality would however be premature given that the data were cross-sectional.
C.R. Brewin et al. / Behaviour Research and Therapy 36 (1998) 1131±1142
1141
Previous studies of stress-related processes in cancer patients have focused on intrusions and avoidance of thoughts about the illness, showing that they have high levels of preoccupation and distress with general illness-related issues and events (Kaasa et al., 1993; Kelly et al., 1995; Tjemsland et al., 1996). Our data go beyond these studies in identifying the intrusion, predominantly among depressed patients, of speci®c autobiographical memories of stressful events from the recent or distant past. Theoretically, the greater accessibility of these memories may be causally involved in producing more severe or chronic depressive episodes (Teasdale, 1988). Consistent with this view, there is evidence that the subsequent course of psychopathology is indeed worse if patients have memories of stressful events intruding frequently into their minds (Brewin et al., 1997, 1998). We therefore suggest that enquiring about such events may be of assistance to clinicians in order to develop hypotheses about the information in memory that may be supporting depressive thinking and maladaptive coping. According to Brewin et al. (1996b), these memories are likely to correspond to events that have received insucient and prematurely inhibited emotional processing. By identifying signi®cant personal episodes, therapists may generate additional opportunities for reinterpreting and cognitively reframing distressing experiences in a more positive way, for example by using ideas and techniques developed for treating post-traumatic stress disorder (e.g., Resick and Schnicke, 1993; Rothbaum and Foa, 1996). The centrality of these experiences in patients' mental lives suggests that this may be a particularly fruitful focus for cognitive work. In conclusion, we have demonstrated that depression among cancer patients is associated with high frequency intrusions of speci®c autobiographical memories that revolve around themes of illness and death and are often cancer-related. These memories, some concerning the patient's own illness but many involving experiences with family and friends, may be a persistent source of information that feeds into and maintains maladaptive coping and negative cognitions about the future. The eort to avoid such memories also appears to be associated with a more global diculty in retrieving speci®c autobiographical memories, another factor believed to undermine constructive coping attempts (Williams, 1992). We suggest that the study of these memory processes may oer new opportunities for therapeutic work with depressed cancer patients.
Acknowledgements This research was supported by Project Grant no. CP1033/0101 awarded by the Cancer Research Campaign.
References American Psychiatric Association (1987). Diagnostic and statistical manual of mental disorders (3rd ed., revised). Washington, D.C.: APA. Andrews, B., Brown, G. W. & Creasey, L. (1990) Intergenerational links between psychiatric disorder in mothers and daughters: The role of parenting experiences. Journal of Child Psychology and Psychiatry, 31, 1115±1129. Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive therapy of depression. New York: Wiley.
1142
C.R. Brewin et al. / Behaviour Research and Therapy 36 (1998) 1131±1142
Brewin, C. R., Christodoulides, J., & Hutchinson, G. (1996a). Intrusive thoughts and intrusive memories in a nonclinical sample. Cognition and Emotion, 10, 107±112. Brewin, C. R., Dalgleish, T., & Joseph, S. (1996b). A dual representation theory of post-traumatic stress disorder. Psychological Review, 103, 670±686. Brewin, C. R., Hunter, E., Carroll, F., & Tata, P. (1996c). Intrusive memories in depression. Psychological Medicine, 26, 1271±1276. Brewin, C. R., Reynolds, M., & Tata, P. (1997). Autobiographical memory processes and the course of depression. Manuscript submitted for publication. Brewin, C. R., Watson, M., McCarthy, S., Hyman, P., & Dayson, D. (1998). Memory processes and the course of psychopathology in cancer patients. Psychological Medicine, 28, 219±224. Brugha, T. S., Bebbington, P., Tennant, C., & Hurry, J. (1985). The list of threatening experiences: A subset of 12 life event categories with considerable long-term contextual threat. Psychological Medicine, 15, 189±194. Corcoran, K., & Fischer, J. (1987). Measures of clinical practice. New York: Free Press. Foa, E. B., Riggs, D. S., Dancu, C. V., & Rothbaum, B. O. (1993). Reliability and validity of a brief instument for assessing posttraumatic stress disorder. Journal of Traumatic Stress, 6, 459±473. Greer, S., Moorey, S., Baruch, J., Watson, M., Robertson, B., Mason, A., Rowden, L., Law, M., & Bliss, J. (1992). Adjuvant psychological therapy for patients with cancer: A prospective randomised trial. British Medical Journal, 304, 675±680. Horowitz, M., Wilner, N., & Alvarez, W. (1979). Impact of Event Scale: A measure of subjective stress. Psychosomatic Medicine, 41, 209±218. Kaasa, S., Malt, U., Hagen, S., Wist, E., Moum, T., & Kvikstad, A. (1993). Psychological distress in cancer patients with advanced disease. Radiotherapy and Oncology, 27, 193±197. Kelly, B., Smithers, M., Swanson, C., McLeod, R., Thompson, D., & Walpole, E. (1995). Psychological responses to malignant melanoma. General Hospital Psychiatry, 17, 126±134. Klein, S. B., & Loftus, J. (1993). The mental representation of trait and autobiographical knowledge about the self. In T. K. Srull & R. S. Wyer (eds.), Advances in social cognition (vol. V, pp. 1±49). Hillsdale, NJ: Erlbaum. Kuyken, W., & Brewin, C. R. (1994a). Intrusive memories of childhood abuse during depressive episodes. Behaviour Research and Therapy, 32, 525±528. Kuyken, W., & Brewin, C. R. (1994b). Stress and coping in depressed women. Cognitive Therapy and Research, 18, 403±412. Kuyken, W., & Brewin, C. R. (1995). Autobiographical memory functioning in depression and reports of early abuse. Journal of Abnormal Psychology, 104, 585±591. Kuyken, W. & Brewin, C. R. (in press). The relation of early abuse to cognition and coping in depression. Cognitive Therapy and Research. Mayou, R., & Hawton, K. (1986). Psychiatric disorder in the general hospital. British Journal of Psychiatry, 149, 172±190. Moorey, S., Greer, S., Watson, M., Gorman, C., Rowden, L., Tunmore, R., Robertson, B., & Bliss, J. (1991). The factor structure and factor stability of the hospital anxiety and depression scale in patients with cancer. British Journal of Psychiatry, 158, 255±259. Razavi, D., Delvaux, N., Farvacques, C., & Robaye, E. (1990). Screening for adjustment disorders and major depressive disorders in cancer in-patients. British Journal of Psychiatry, 156, 79±83. Resick, P. A., & Schnicke, M. K. (1993). Cognitive processing therapy for rape victims. Newbury Park, CA: Sage. Rothbaum, B. O., & Foa, E. B. (1996). Cognitive-behavioral therapy for posttraumatic stress disorder. In B. A. van der Kolk, A. C. McFarlane, & L. Weisaeth (eds.), Traumatic stress (pp. 491±509). New York: Guilford Press. Smith, E. R. (1990). Content and process speci®city in the eects of prior experiences. In T. K. Srull & R. S. Wyer (eds), Advances in social cognition (vol. III, pp. 1±59). Hillsdale, NJ: Erlbaum. Spiegel, D. (1996). Cancer and depression. British Journal of Psychiatry, 168 (Suppl. 30), 109±116. Strauman, T. J. (1992). Self-guides, autobiographical memory, and anxiety and dysphoria: Toward a cognitive model of vulnerability to emotional distress. Journal of Abnormal Psychology, 101, 87±95. Teasdale, J. D. (1988). Cognitive vulnerability to persistent depression. Cognition and Emotion, 2, 247±274. Tjemsland, L., Sùreide, J. A., & Malt, U. F. (1996). Traumatic distress symptoms in early breast cancer I: Acute response to diagnosis. Psycho-oncology, 5, 1±8. Watson, M., Greer, S., Rowden, L., Gorman, C., Robertson, B., Bliss, J. M., & Tunmore, R. (1991). Relationships between emotional control, adjustment to cancer, and depression and anxiety in breast cancer patients. Psychological Medicine, 21, 51±57. Watson, M., Law, M., dos Santos, M., Greer, S., Baruch, J., & Bliss, J. (1994). The Mini-MAC: Further development of the mental adjustment to cancer scale. Journal of Psychosocial Oncology, 12, 33±46. Williams, J. M. G. (1992). Autobiographical memory and emotional disorders. In S.-A. Christianson (ed.), Handbook of Emotion and Memory (pp. 451±477). Hillsdale, NJ: Lawrence Erlbaum. Williams, J. M. G., & Broadbent, K. (1986). Autobiographical memory in attempted suicide patients. Journal of Abnormal Psychology, 95, 144±149. Williams, J. M. G., & Scott, J. (1988). Autobiographical memory in depression. Psychological Medicine, 18, 689±695. Zigmond, A., & Snaith, R. P. (1983). The hospital anxiety and depression scale. Acta Psychiatrica Scandinavica, 67, 361±370.