Psycho-Oncology Psycho-Oncology 18: 11471155 (2009) Published online 12 January 2009 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/pon.

1512

Intrusive cognitions and their appraisal in anxious cancer patients

Katriina L. Whitaker1,2, Maggie Watson1,2 and Chris R. Brewin2
1 2

Psychological Medicine, Royal Marsden NHS Foundation Trust, UK Clinical, Educational and Health Psychology, University College London, London, UK

* Correspondence to: University College London, Gower Street, London, WC1E 7HN, UK. E-mail: k.whitaker@ucl.ac.uk

Abstract
Objective: Previous research found that anxious cancer patients experience uncontrollable negative intrusive cognitions that have an impact on coping and are associated with signicant psychological distress. This is the rst study to examine the appraisal of intrusive cognitions in an anxious group of cancer patients. Methods: A sample of 139 anxious cancer patients was assessed for evidence of intrusive phenomena, including memories, images and thoughts. Patients completed the Response to Intrusions Questionnaire and the Impact of Event Scale in relation to intrusive cognitions. Results: Forty-eight percent (67/139) reported frequent, uncontrollable intrusive cognitions. Intrusive thoughts and images were equally as common and images were associated with increased distress and uncontrollability. A signicant positive linear relationship was found between the number of intrusions and anxiety severity (Po0.05). Negative appraisal of intrusive cognitions was associated with anxiety (Po0.01) and depression severity (Po0.01), intrusion-specic distress (Po0.01), rumination (Po0.01) and cognitive avoidance (Po0.01), after controlling for intrusion frequency. Conclusion: Negative appraisal of intrusive cognitions plays a signicant role in psychological distress and intrusion-specic distress in anxious cancer patients. Finding similarities in the types of intrusive cognitions reported by cancer patients and other anxious populations highlights the potential applicability of psychological therapies developed to reduce the frequency and impact of intrusive cognitions. Copyright r 2009 John Wiley & Sons, Ltd.
Keywords: anxiety; cancer; intrusive cognitions; appraisal; coping

Received: 3 September 2008 Revised: 3 November 2008 Accepted: 5 November 2008

Introduction
Previous research showed that anxious cancer patients experience uncontrollable negative intrusive cognitions that are often associated with feelings of sadness, helplessnesshopelessness and interfere with daily life [1]. However, intrusions are less common than found in psychiatric populations (e.g. [2]) and associated with less distress [1]. Possible explanations for the nding that intrusive cognitions are less impacting for anxious cancer patients include the relatively mild levels of anxiety reported by cancer patients and the possibility that cancer patients appraise cognitions in a qualitatively dierent way from other groups. Negative appraisal of intrusive cognitions has been identied as a key cognitive mechanism that mediates the management of intrusive cognitions in PSTD [3] and depression [4]. Cancer patients have reported experiencing negative intrusive cognitions, often about illness, injury or death [1]. However, because cancer patients have the stressor of a physical

illness, intrusions may be appraised as a normal stress response reaction, rather than a sign of mental fragility. It is important to apply cognitive models [3] of intrusive symptoms to cancer patients, to ascertain whether intrusive cognitions are qualitatively dierent for physically ill patients, compared with patients with depression or anxiety disorders. As previously shown experiencing intrusive cognitions is related to anxiety in cancer patients and understanding the mechanisms involved is key to uncovering ways to alleviate intrusive cognitions and psychological distress. Following our previous work, which established the relationship between intrusive cognitions and anxiety in cancer patients by comparing anxious and non-anxious patients, this study investigates a new anxious sample of cancer patients to examine the role of negative appraisal. Although intrusive thoughts have been extensively investigated in cancer patients [5] and the presence of intrusive imagery has now also been considered [1], the appraisal of such cognitions has

Copyright r 2009 John Wiley & Sons, Ltd.

1148

K. L. Whitaker et al.

yet to be investigated. Understanding appraisal of intrusions is important for therapeutic reasons, because modifying negative appraisals may be a successful treatment approach for reducing anxiety. The concept of intrusion appraisal was rst discussed by Ehlers and Steil [3] and later elaborated by Ehlers and Clark [6] in their cognitive approach to posttraumatic stress disorder. They suggested that the meaning attributed to the experience of intrusive recollections (e.g. I am going mad) determines how distressing the intrusions are and the extent to which patients engage in coping strategies to control the intrusions. Coping strategies are implicated in symptom maintenance, as they prevent a change in the meaning of the trauma and subsequent recollections [3,6]. In support of this model, retrospective studies have shown that negative appraisal of intrusive cognitions was associated with PTSD severity and intrusion-specic distress, after controlling for intrusion frequency in ambulance workers [7], victims of physical or sexual assault [8], and motor vehicle accident survivors [9]. These ndings have been replicated in prospective research [10,11]. Negative appraisal of intrusions has also been related to overall traumatic grief and depression in the bereaved [12] and negative appraisal of intrusive memories was positively associated with intrusion-related distress and level of depression in student samples [4] and a dysphoric sample [13]. Although intrusion appraisal has not yet been considered in cancer patients, the cognitive model of adjustment [14] predicts that negative appraisal (i.e. negative meanings that individuals assign to the cancer experience) may inuence emotional responses, which include experiences such as intrusive cognitions. In line with this, the meaning of events to an individual with cancer has been highlighted as key to understanding anxiety [15] and previous research has found a link between the negative appraisal of disease-threat and severity of stress response scores [16,17]. The cognitive approach to the maintenance of intrusive cognitions [3] also predicted that the negative meaning assigned to intrusive cognitions would lead to maladaptive coping strategies. In support of this, negative meaning of intrusive cognitions was signicantly related to coping strategies used to control intrusive cognitions such as avoidance, rumination, suppression and distraction in motor vehicle accident survivors [9] and negative interpretation of intrusive memories was signicantly related to cognitive avoidance in a student sample [4]. Assessing the presence, appraisal of, and coping with intrusive cognitions in a large anxious cancer sample also allows further exploration of the characteristics of intrusive cognitions in anxious cancer patients. For example, previous research showed that intrusive verbal thoughts were more common than intrusive visual imagery [1]. However,
Copyright r 2009 John Wiley & Sons, Ltd.

it has been suggested that imagery is dierent from verbal representations of the same material because it is associated with increased emotional responses [18]. The present study investigated negative appraisal of intrusive cognitions in anxious cancer patients and assessed how cancer patients cope with intrusive cognitions. The following hypotheses were proposed: (1) there would be a positive linear relationship between the number of intrusions reported and anxiety severity, (2) verbal intrusions (i.e. thoughts) would be more common than visual intrusions (i.e. images and memories) (3) visual intrusions would be more distressing than verbal intrusions, (4) there would be a positive association between negative appraisal of intrusive cognitions and anxiety/ depression severity, after controlling for intrusion frequency, (5) there would be a positive association between negative appraisal of intrusive cognitions and intrusion associated distress after controlling for intrusion frequency and (6) negative appraisal of intrusive cognitions would be associated with the extent to which patients engaged in maladaptive coping strategies to control their intrusions.

Method Patients and procedure

The study was approved by the Royal Marsden Hospital Local Research Ethics Committee. Two groups of patients were approached: (1) A consecutive series of patients attending outpatient clinics were screened using the HADS, completed on site or at home and returned by mail. Patients scoring X8 were invited to participate in the study interview, either on the telephone or in person. (2) Patients referred to the Psychological Medicine Service at the hospital were sent an invitation pack inviting them to take part. The invitation pack included the HADS to send back with the reply form expressing an interest in taking part. Patients identied as anxious (X8) were contacted by telephone to arrange an interview, either on the telephone or in person. Patients who expressed an interest in taking part, but did not meet criteria for anxiety, were contacted by telephone to inform them that they were ineligible. For all participants, if a time lapse of more than 21 days had passed, they were asked to repeat the HADS questionnaire, to ensure eligibility. Patients interviewed also completed the IES, RIQ items and coping scales in response to any reported intrusions.
Psycho-Oncology 18: 11471155 (2009) DOI: 10.1002/pon

Appraisal in anxious cancer patients

1149

Outpatient screening
Of 870 cancer patients approached in the outpatient clinics at the Royal Marsden NHS Foundation Trust, 506 (58%) completed and returned the Hospital Anxiety and Depression Scale (HADS; [19]). Sixteen (2%) patients declined to participate at the screening stage; 1 was recently widowed, 1 had poor English, 2 patients were too distressed and 12 did not specify a reason for refusal. A further 348 (40%) of patients did not return the HADS. At the screening phase, there were no signicant dierences between responders and non-responders on time since diagnosis [t(856) 5 0.37, P 5 0.71] or sex [w2(1) 5 1.22, P 5 0.30]. However, non-responders were signicantly younger than responders [t(865) 5 3.19, Po0.01] and were signicantly more likely to be of Non-Caucasian ethnicity [w2(1) 5 16.36, Po01]. Patients who scored 8 or above on the anxiety subscale (28%) were categorised as anxious [19] and selected for the interview stage of the study. Of these 141 anxious patients, 66 (47%) responded and were interviewed. For the interview stage, responders and non-responders did not dier on time since diagnosis [t(139) 5 1.21, P 5 0.23], age [t(139) 5 0.62, P 5 0.53], ethnic origin [w2(1) 5 0.32, P 5 0.57] or disease stage [w2(2) 5 0.49, P 5 0.78]. However, responders were signicantly more likely to be female than non-responders [w2(1) 5 4.53, Po0.05]. Responders and non-responders did not dier on anxiety [t(139) 5 0.89, P 5 0.32], or depression [t(139) 5 1.83, P 5 0.07].

139 patients combined were interviewed and 67 reported intrusive cognitions. Of the total sample, 55 patients had breast cancer (39.5%), 6 patients had gynaecological cancer (4.3%), 20 patients had urological cancer (14.4%), 30 patients had lymphomas (21.6%), 9 patients had lung cancer (6.5%), 11 patients had gastro-intestinal cancer (7.9%) and 8 patients had head and neck cancers (5.8%). Seventy-one patients were interviewed on the telephone (51.1%) and 68 patients were interviewed in person (48.9%).

Measures
Screening

Hospital Anxiety and Depression Scale (HADS): The HADS [19] is a 14-item self-report scale developed for the measurement of depression and anxiety in physically ill populations. The scale has been validated in cancer patients [21].
Interview session

Psychological medicine referrals

Of 278 patients referred to the Psychological Medicine Service, 97 (35%) completed and returned the HADS. Twenty-four (8%) scored below cut-o on the HADS and thus a total of 73 patients were interviewed. Twenty-eight (10%) declined to participate; two felt too ill, two did not want to talk about their illness, two did not feel anxious, 1 person did not have enough time and 21 did not give a reason for their refusal. A further 153 patients (55%) did not respond to invitations to interview. There were no signicant dierences between responders and non-responders on age [t(273) 5 0.55, P 5 0.58], time since diagnosis [t(272) 5 0.50, P 5 0.62], ethnic origin [w2(1) 5 3.39, P 5 0.07] or sex [w2(1) 5 0.36, P 5 0.22].

Final interview sample

A priori power calculations indicated that to detect a medium eect size (r 5 0.30, a 5 0.05) with 0.80 power requires a sample size of 64 participants [20] and therefore at least 64 patients were required to report intrusive cognitions. From outpatient screening and psychological medicine referrals,
Copyright r 2009 John Wiley & Sons, Ltd.

Structured Clinical Interview for DSM-IV: All patients were asked whether their anxiety/concerns were related or unrelated to the cancer diagnosis. Following this, generalised anxiety disorder, major depression and adjustment disorder were assessed using sub-sections of the Structured Clinical Interview for DSM-IV Disorders (SCID; [22]). The diagnosis of adjustment disorder was made based on predominant symptoms; adjustment with depressed mood, adjustment with anxiety or adjustment with mixed anxiety and depression. Intrusive cognitions: A structured interview assessed the presence of repetitive intrusive phenomena, including memories, images and thoughts [23]. Denitions were based on those of Patel et al. [23]. Memories were dened as a visual image of a specic event and its surrounding context that occurred in the past. Images were dened as a specic visual picture, relating to the past, present or future that did not meet criteria for memory. For example, a family members disembodied ill face, devoid of any surrounding context would be classied as an image, whereas being in a specic hospital on a particular day with a family member who was dying would be classied as a memory. Thoughts were included if they consisted of verbal content referring to the past, present or future. If more than one intrusion was reported, patients were asked to identify and concentrate on the two most intrusive cognitions. Questions asked in relation to intrusive cognitions included a description of the intrusion, associated emotions (i.e. sadness, guilt, shame, anger, anxiety and helplessness; 0 5 not at all, 100 5 very much so), frequency (0 5 none of the time, 100 5 all of the time), duration (1 5 seconds, 2 5 minutes, 3 5 hours), interference with daily life
Psycho-Oncology 18: 11471155 (2009) DOI: 10.1002/pon

1150

K. L. Whitaker et al.

(0 5 not at all, 100 5 severely), uncontrollability (0 5 not at all, 100 5 completely) and associated distress (0 5 not at all, 100 5 severely). For memories and images, participants were asked how vivid the image was (0 5 hazy memory, 100 5 clearest and most vivid memory). For images of past events, patients were asked whether it felt as though they were reliving the memory (0 5 not at all, 100 5 very much so) and accompanying emotional and physical sensations (0 5 not at all, 100 5 very much so). For images that were not past events, patients were asked if the image was related to an event that had actually happened. Impact of Event Scale (IES): The IES [24], a 15item self-report scale, consists of intrusion and avoidance sub-scales. High internal consistency, test re-test reliability and validity have been reported [25]. Appraisal of intrusive cognitions: Response to Intrusions Questionnaire (RIQ): Six items measuring negative appraisal of intrusive cognitions from the RIQ [7] were used including Something is wrong with me, I will not achieve goals that are important to me (replaced I will not be able to do my job properly [12]);I am inadequate, I cannot cope, Some day I will go out of my mind, I have a psychological problem. Patients were required to rate from 1 (totally disagree) to 7 (totally agree) for each item. Internal consistency for the negative appraisal scale was high (a 5 0.86). Coping with intrusive cognitions: Patients were asked to rate on a 100-point scale (0 5 not used at all, 100 5 used very much) the extent to which they engaged in three coping strategiesdistraction (I try to distract myself), suppression (I try to push the intrusions out of my mind) and rumination (I dwell on it). The items were adopted from previous research (e.g. [9]).

Recruitment, demographic and clinical influences on reporting intrusive cognitions

There was no signicant dierence in whether patients reported intrusive cognitions according to whether they were interviewed on the telephone or in person [w2(1) 5 0.36, P 5 0.55]. However, patients recruited from Psychological Medicine were signicantly more likely to report intrusive cognitions than patients recruited from outpatient screening [w2(1) 5 5.36, Po0.05, Cramers V 5 0.20]. Age, sex, ethnic origin and marital status had no eect on whether patients reported intrusive cognitions (P40.05). Independent t-tests revealed that time since diagnosis signicantly aected whether patients reported intrusive cognitions [t(137) 5 2.46, Po0.05, r 5 0.21] and this relationship remained [R 5 0.18, Po0.05] after controlling for sample type (Psychological Medicine or Outpatient). The group of patients reporting intrusive cognitions had signicantly longer post diagnosis (mean 5 49.60 months, SD 5 58.67) than patients not reporting intrusive cognitions (mean 5 28.90 months, SD 5 39.22). The presence of intrusive cognitions was not related to treatment type [w2(1) 5 0.00 P 5 0.98] or disease stage [w2(2) 5 3.21 P 5 0.20]. Finally, the group of patients reporting intrusive cognitions were signicantly more likely to meet criteria for DSM-IV diagnoses (30/67) than those not reporting intrusive cognitions (17/72), w2(1) 5 6.95, Po0.01.

Intrusive cognitions and anxiety

In order to investigate the relationship between number of intrusive cognitions and anxiety level, the total sample was divided into three groups according to pre-dened criteria [26]; mildly anxious (n 5 53, HADS 5 810), moderately anxious (n 5 55, HADS 5 1114) and severely anxious (n 5 31, HADS 5 1521). A one-way ANOVA indicated a signicant overall eect of anxiety category on the number of intrusive cognitions reported [F(2,136) 5 3.40, Po0.05]. A polynomial contrast analysis showed that the mean number of intrusions increases (contrast estimate 5 0.27, Po0.05) as anxiety level increases. The linear relationship between anxiety and intrusive cognitions may explain why a higher number of patients from Psychological Medicine reported intrusive cognitions than patients screened in outpatient clinics as the former were signicantly more anxious.

Results Participant characteristics

The demographic, clinical and psychological characteristics of the total sample, patients screened through Psychological Medicine, and patients screened through outpatients, are presented in Table 1. Comparisons showed that patients recruited from Psychological Medicine referrals were signicantly more anxious and depressed than patients meeting the cut-o for anxiety via outpatient screening. They were also more likely to be diagnosed with anxiety or depression using the SCID compared with outpatients. Patients from Psychological Medicine were more likely to have advanced disease, whereas patients from outpatient screening were more likely to have locally advanced disease. Psychological Medicine patients had signicantly longer post diagnosis and were quicker to return the HADS.
Copyright r 2009 John Wiley & Sons, Ltd.

Number and type of intrusive cognitions

Sixty-seven patients (48%) reported an intrusive cognition, of whom 14 patients reported at least one additional intrusion. Of these, three patients reported two intrusive memories, seven patients
Psycho-Oncology 18: 11471155 (2009) DOI: 10.1002/pon

Appraisal in anxious cancer patients

1151

Table 1. Demographic, clinical and psychological characteristics of the total sample (N 5 139), patients recruited from Psychological Medicine (n 5 73) and patients recruited from Outpatient Clinics (n 5 66)
Characteristic Total sample Psychological Medicine patients Outpatients Difference between Psychological Medicine and outpatients

Continuous variables [Mean (SD)] Age Months since diagnosis Days since HADS completion HADS anxiety score HADS depression score Categorical variables [n (%)] Sex Male Female Ethnic origin White British Other Marital status Single Married/living with a partner Separated/divorced Widowed Cancer stage Early Locally advanced Advanced Unknown Treatment On treatment Post-treatment Undecided DSM-IV diagnoses Does not meet criteria Generalized Anxiety Disorder (GAD) Major Depressive Disorder (MDD) Both GAD and MDD Adjustment disorderdepression Adjustment disorderanxiety Adjustment disordermixed
Po0.05. Po0.01 (two-tailed).

52.22 (12.18) 38.88 (50.45) 7.14 (6.70) 11.97 (3.05) 7.14 (4.03)

51.08 (12.45) 47.41 (59.04) 6.03 (7.50) 12.84 (3.09) 8.27 (4.26)

53.48 (11.85) 29.44 (37.03) 8.38 (5.48) 11.02 (2.73) 5.88 (3.37)

t(137) 5 1.16, t(137) 5 2.12, t(137) 5 2.09, t(137) 5 3.67, t(137) 5 3.65,

P 5 0.25 Po0.05, r 5 0.18 Po0.05, r 5 0.18 Po0.01, r 5 0.30 Po0.01, r 5 0.30

43 (31) 96 (69) 101 (72.7) 38 (27.3) 20 95 20 4 (14.4) (68.3) (14.4) (2.9)

26 (35.6) 47 (64.4) 50 (68.5) 23 (31.5) 11 50 11 1 37 6 25 5 (15.1) (68.5) (15.1) (1.4) (50.7) (8.2) (34.2) (6.8)

17 (25.8) 49 (74.2) 51 (77.3) 15 (22.7) 9 (13.6) 45 (68.2) 9 (13.6) 3 (4.5) 34 (51.5) 14 (21.2) 10 (15.2) 8 (12.1) 35 (53.0) 30 (45.5) 1 (1.5) 55 (83.3) 0 (0) 0 (0) 1 (1.5) 8 (12.1) 1 (1.5) 1 (1.5)

w2(1) 5 1.58, P 5 0.21

w2(1) 5 1.35, P 5 0.25

w2(1) 5 0.00, P 5 0.97

71 (51.1) 20 (14.4) 35 (25.2) 13 (9.3) 65 (46.8) 69 (49.6) 5 (3.6) 92 (66.2) 3 (2.2) 2 (1.4) 2 (1.4) 12 (8.6) 17 (12.2) 11 (7.9)

w2(2) 5 9.02, Po0.01 Cramers V 5 0.27

30 (41.1) 39 (53.4) 4 (5.5.) 37 (50.7) 3 (4.1) 2 (2.7) 1 4 16 10 (1.4) (5.5) (21.9) (13.7)

w2(2) 5 1.44, P 5 0.23

w2(1) 5 16.51, P 5 0.00 Cramers V 5 0.35 (between patients meeting criteria or not)

reported an intrusive memory and an intrusive thought, one patient reported an intrusive image and an intrusive thought and three patients reported an intrusive memory and an intrusive image. Of the 81 intrusions reported in total, 31 were intrusive memories, 8 were intrusive images and 42 were intrusive thoughts. Fifty-eight (72%) of the intrusive cognitions related to the persons own experience of having cancer and 13 (16%) of the intrusions related to a relatives illness, injury or death (three specically from cancer). In total, 75% of reported intrusions were specically related to cancer and 10 (12%) were unrelated to illness or death. Four of the eight reported images were related to a past event and the four remaining intrusive images were future
Copyright r 2009 John Wiley & Sons, Ltd.

oriented. Thirty-four (81%) of thoughts were future oriented, 1 was related to past event, 1 was both future and present oriented and 1 was both future and past oriented. See Table 2 for examples of dierent types of intrusive cognitions reported.

Characteristics of intrusive cognitions

Table 3 shows the mean characteristics of intrusive cognitions. Of the total sample of cognitions (n 5 81), intrusive cognitions were reported to occur just over half the time in the past week, to interfere moderately with daily life, to be moderately to severely distressing, and to be severely uncontrollable. Intrusions most often lasted for minutes rather than being eeting or lasting for hours. To investigate the dierence between specic characteristics of visual
Psycho-Oncology 18: 11471155 (2009) DOI: 10.1002/pon

1152

K. L. Whitaker et al.

Table 2. Typical content of cancer-related and cancer-unrelated intrusive cognitions

Timescale Intrusions related to cancer Memory Past Image Future Thought Future Intrusions unrelated to cancer Memory Past Image Future Thought Future Content Having high-dose chemotherapy, smell of sweet corn and messing the bed because I am incontinent Looking down on myself at my own funeral and seeing my friend and family who are crying and hearing music playing I am going to die, the cancer is not going to go away Boss at work, going to attack me and feeling humiliated Horrific monster type faces with sharp horrible teeth, snarling and mutating coming towards me Fears about flat being destroyed

Table 3. Characteristics of all intrusive cognitions (n 5 81), visual intrusions (n 5 33) and verbal intrusions (n 5 34)
Characteristic All intrusions Mean (SD) 56.30 (24.05) 49.93 (32.43) 81.43 (23.83) 69.94 (26.49) 39.59 (14.31) 18.69 (9.15) 20.90 (7.81) 17 (21) 52 (64.2) 12 (14.8) Visual intrusions Mean (SD) 56.82 (24.17) 45.45 (34.45) 85.00 (24.01) 70.00 (27.64) 42.33 (12.93) 19.64 (8.47) 22.70 (7.37) 9 (27.3) 20 (60.6) 4 (12.1) Verbal intrusions Mean (SD) 59.85 (19.95) 54.09 (31.40) 75.15 (23.69) 69.85 (23.60) 34.32 (13.28) 15.68 (8.09) 18.65 (7.39) 1 (3) 27(79.4) 6 (17.6) Difference between visual and verbal intrusions

Frequency Interference Uncontrollability Distress IES total IES avoidance IES intrusion Duration n (%) Seconds Minutes Hours

t(65) 5 0.56, P 5 0.58 t(65) 5 0.47, P 5 0.29 t(65) 5 1.70, P 5 0.09 t(65) 5 0.2, P 5 0.98 t(65) 5 2.50, Po0.01, r 5 0.30 t(65) 5 1.96, Po0.05, r 5 0.24 t(65) 5 2.25, Po0.05, r 5 0.27 Fisher exact test o0.05, Cramers V 5 0.34

Table 4. Correlates of negative appraisal (RIQ) of intrusive cognitions (n 5 67)

Negative appraisal Anxiety Depression Intrusion uncontrollability Intrusion distress IES total IES avoidance IES intrusion Distraction Suppression Rumination
Po0.05. Po0.01 (one-tailed).

Partial correlation controlling for frequency 0.29 0.31 0.27 0.34 0.15 0.23 0.01 0.08 0.10 0.33

0.37 0.42 0.31 0.42 0.22 0.28 0.08 0.17 0.19 0.41

(images and memories combined) and verbal intrusions, independent t-tests were conducted. For patients reporting two or more intrusions, one intrusion was randomly selected to be included in the analysis. Visual intrusions were associated with signicantly more subjective distress, including intrusion and avoidance, according to the IES. Visual intrusions were also more uncontrollable than verbal intrusions, although this dierence did not reach statistical signicance. A Fisher exact test showed that visual intrusions were more often shorter in duration than verbal intrusions, although both intrusions usually lasted for minutes. Intrusive memories and intrusive images were reported to be extremely vivid
Copyright r 2009 John Wiley & Sons, Ltd.

(mean 5 89.26, SD 5 15.74 and mean 5 85.63, SD 5 21.45, respectively). For intrusive memories, patients reported experiencing emotions the same as, or similar to those experienced during the actual event (mean 5 82.10, SD 5 27.57) and reported moderately reliving the memory (mean 5 54.52, SD 5 41.52). Re-experiencing physical sensations was reported a little (mean 5 37.74, SD 5 43.64).

Correlates of negative intrusion appraisal (RIQ)

Table 4 contains the correlations between negative appraisal of intrusions, anxiety, depression, intrusion-related distress and intrusion-specic coping.
Psycho-Oncology 18: 11471155 (2009) DOI: 10.1002/pon

Appraisal in anxious cancer patients

1153

Negative appraisal was positively correlated with anxiety, depression and intrusion-related distress, after controlling for intrusion frequency. Negative appraisal was also positively correlated with intrusion-specic coping, including avoidance (IES) and rumination, after controlling for intrusion frequency.

Discussion
The present study replicates and extends our previous nding that approximately half of anxious cancer patients experience frequent negative intrusive cognitions that are distressing, interfering, uncontrollable and commonly related to future concerns. It was also found that those cancer patients with intrusions tended to appraise them in a negative way, seeing them, for example, as a sign that they could not cope or had a psychological problem. Specically, mean negative appraisal of intrusive cognitions was higher in cancer patients than in ambulance workers and non-clinical populations [4,7], and was equivalent to a depressed sample [13]. This contradicts one explanation for the lower frequency and impact of intrusions in cancer patients compared with psychiatric samples [1]. Clearly cancer patients do perceive intrusions as a sign of mental fragility. The data supported the alternative explanation that lower frequency intrusions are related to lower anxiety levels in cancer than in psychiatric samples. The linear relationship between anxiety and intrusive cognitions appears to be robust, supporting the notion that experiences such as intrusive cognitions exist on a continuum from non-clinical [4,27] to clinical levels [23,28]. As predicted, negative appraisal of intrusive cognitions was related to general psychological symptoms, including anxiety and depression severity, and also to intrusion-specic characteristics, such as distress and uncontrollability. These relationships remained after controlling for intrusion frequency. This follows a pattern of results reported in PTSD samples [7,10,11], the bereaved [12], student [29] and dysphoric samples [13]. Rumination and intrusion avoidance were also associated with negative appraisal, irrespective of intrusion frequency. This is in line with the suggestion that cognitive avoidance and rumination used in response to intrusive cognitions paradoxically prevent the elaboration and integration of intrusive cognitions and help maintain negative appraisals [30]. Thus, such strategies may serve to exacerbate rather than ameliorate intrusive cognitions, which in turn lead to the maintenance of distress [3]. Overall the results support the notion that negative appraisal of intrusive cognitions plays a role in the development of emotional distress after cancer diagnosis. The present study made novel comparisons between visual and verbal intrusions in cancer
Copyright r 2009 John Wiley & Sons, Ltd.

patients. Contrary to prediction, similar levels of visual and verbal intrusions were reported. However, in support of previous research [1], futureoriented intrusive images were relatively uncommon, unlike in some studies of patients with anxiety disorders [2]. Instead, memories were more prominent, with the majority of patients describing contextualised imagery anchored in the past. As predicted, comparisons of visual and verbal intrusions revealed that visual intrusions were associated with signicantly more subjective distress (intrusion and avoidance) than verbal intrusions. Also, intrusive imagery was more likely to last for seconds compared with intrusive thoughts. These ndings are consistent with previous work that emphasised the eeting nature of intrusive imagery [31] and the special relationship between imagery and emotion [18]. Holmes et al. suggested that imagery has a greater impact on emotion than verbal representations of the same material, because images are more like actual percepts [32], images trigger episodes in autobiographical memory [33] and images often include personal involvement in events [18]. Although based on a specic set of denitions developed in previous studies, the conrmation that verbal and visual intrusions are distinguishable in this population supports theoretical and applied research that emphasises their independence [27,34]. So far, research has focussed almost exclusively on the presence of intrusive thoughts in cancer patients [5], leaving an entire category of intrusions unexplored. Future research should also assess whether factors such as imaging ability or hypnotisability [35] impact whether individuals experience visual intrusions. A number of unexpected ndings require explanation. For example, rumination, suppression and distraction were not associated with general distress in the present study. Previous research also found no relationship between rumination and suppression of intrusive cognitions and depression [4], although others did report a relationship [7]. Future experimental research may be required to resolve inconsistencies and elucidate the role of thought suppression on the frequency and impact of intrusive cognitions. Distraction may have been unrelated to psychological distress because it is not necessarily a negative coping strategy. Another unexpected nding was that time since diagnosis was higher in the group of patients reporting intrusive cognitions than those not reporting intrusive cognitions. This contradicts previous research, where a greater proximity to diagnosis is associated with a higher frequency and severity of stress response symptoms [36]. Reports of ongoing intrusive cognitions in the present study highlight the prolonged nature of cancer as a stressor that can precipitate intrusive symptomatology throughout the course of the disease.
Psycho-Oncology 18: 11471155 (2009) DOI: 10.1002/pon

1154

K. L. Whitaker et al.

Limitations of the present study include the inability to draw conclusions about directionality. Prospective research is required as the crosssectional design precludes causal interpretation. Also, the number of correlations conducted may increase the probability that relationships were encountered, although the consistent pattern found across correlations and concordance with previous research means it is unlikely our ndings were found by chance. Intrusion-specic coping was assessed with single-item measures that may have compromised the reliability of the assessment and underestimated the eect of intrusion coping on negative appraisal. Although we considered negative internal appraisals, other appraisals may play a role in the presence and maintenance of intrusive cognitions in cancer patients [37]. For example, internal threat appraisal encompasses individual perceptions of the self as capable, acceptable and in control, in addition to how individuals perceive their reaction to intrusive cognitions [37]. Also, external threat appraisals (e.g. disease-specic threat) have consistently played a role in psychological models of cancer such as the cognitive model of adjustment [14] and found to be related to stress response symptoms [16]. Another limitation is that PTSD was not assessed in the present study. This was because patient demands needed to be minimised while replicating previous research as closely as possible (e.g. [4]). Finally, the recruitment rates for outpatients were signicantly lower in the present study (48%) compared with our previous study [1] with prostate cancer patients (65%), which may be due to the age dierences between the samples and the nding that non-responders are signicantly younger than responders. Recruitment rates for Psychological Medicine referrals were even lower (35%), which may be because patients referred to Psychological Medicine were at the peak of their anxiety and therefore less likely to participate, although this cannot be claried as these patients did not complete the HADS. There are signicant clinical implications of the present study. For example, simply asking cancer patients about whether they experience intrusive cognitions may help normalise the experience, followed by psycho-education to explain the function of intrusions and the negative consequences of avoidance. The fact that negative appraisals associated with these intrusions have been identied suggests that the exposure treatments commonly employed with intrusive memories and images might have to be supplemented by procedures aimed at identifying and modifying the negative meanings attached to them [9]. There are thus a number of additional therapeutic approaches that might prove benecial for cancer patients reporting these experiences. Overall, the present study adds to our understanding of intrusive cognitions in cancer patients.
Copyright r 2009 John Wiley & Sons, Ltd.

The nding that the negative appraisal of intrusive cognitions plays a signicant role in anxiety and depression severity, and intrusion-specic distress, strengthens the argument that intrusive cognitions are an important area of research in psychooncology. In particular, if intrusive cognitions reported by cancer patients show similarities to those found in other clinical populations, psychological therapies developed elsewhere to reduce the impact of intrusions may provide a reduction in distress for cancer patients reporting these experiences.

Acknowledgements
This research was supported by a Cancer Research UK doctoral studentship [Grant no: C3763/A3744]. We would also like to thank the patients at the Royal Marsden NHS Foundation Trust who so kindly gave their time to participate in this research.

References
1. Whitaker KL, Brewin CR, Watson M. Intrusive cognitions and anxiety in cancer patients. J Psychosom Res 2008;64:509517. 2. Hackmann A, Clark DM, McManus F. Recurrent images and early memories in social phobia. Behav Res Ther 2000;38(6):601610. 3. Ehlers A, Steil R. Maintenance of intrusive memories in posttraumatic stress disorder: a cognitive approach. Behav Cognitive Psychother 1995;23:217249. 4. Starr S, Moulds ML. The role of negative interpretations of intrusive memories in depression. J Aect Disord 2006;93:125132. 5. Kangas M, Henry JL, Bryant RA. Posttraumatic stress disorder following cancera conceptual and empirical review. Clin Psychol Rev 2002;22(4):499524. 6. Ehlers A, Clark DM. A cognitive model of posttraumatic stress disorder. Behav Res Ther 2000;38:319345. 7. Clohessy S, Ehlers A. PTSD symptoms, response to intrusive memories and coping in ambulance workers. Br J Clin Psychol 1999;38:251265. 8. Dunmore E, Clark DM, Ehlers A. Cognitive factors involved in the onset and maintenance of PTSD. Behav Res Ther 1999;37:809829. 9. Steil R, Ehlers A. Dysfunctional meaning of posttraumatic intrusions in chronic PTSD. Behav Res Ther 2000;38:537558. 10. Dunmore E, Clark DM, Ehlers A. A prospective study of the role of cognitive factors in persistent posttraumatic stress disorder after physical or sexual assault. Behav Res Ther 2001;39:10631084. 11. Ehlers A, Mayou RA, Bryant B. Psychological predictors of chronic posttraumatic stress disorder after motor vehicle accidents. J Abnormal Psychol 1998;107(3):508519. 12. Boelen PA, van den Bout J, van den Hout MA. The role of negative interpretations of grief reactions in emotional problems after bereavement. J Behav Ther Exp Psychiatry 2003;34:225238. 13. Moulds ML, Kandris E, Williams AD, Lang TJ. The use of safety behaviours to manage intrusive memories in depression. Behav Res Ther 2008;46:573580. 14. Moorey S, Greer S. Cognitive Behaviour Therapy for People with Cancer. Oxford University Press, Oxford, 2002.
Psycho-Oncology 18: 11471155 (2009) DOI: 10.1002/pon

Appraisal in anxious cancer patients

1155

15. Stark DPH, House A. Anxiety in cancer patients. Br J Cancer 2000;83(10):12611267. 16. Hampton MR, Frombach I. Womens experience of traumatic stress in cancer treatment. Health Care Women Int 2000;21(1):6776. 17. Widows MR, Jacobsen PB, Fields KK. Relation of psychological vulnerability factors to posttraumatic stress disorder symptomatology in bone marrow transplant recipients. Psychosom Med 2000;62(6):873882. 18. Holmes EA, Mathews A, Mackintosh B, Dalgleish T. The eect of mental imagery on emotion assessed using picture-word cues. Emotion 2008;8(3):395409. 19. Zigmond AS, Snaith RP. The Hospital Anxiety and Depression Scale. Acta Psychiatr Scand 1983;67:361370. 20. Cohen J. Statistical Power analysis for the Behavioral Sciences (2nd edn). Lawrence Erlbaum Associates: New Jersey, 1988. 21. Moorey S, Greer S, Watson M. The factor structure and factor stability of the Hospital Anxiety and Depression Scale in patients with cancer. Br J Psychiatry 1991; 158:255259. 22. First MB, Gibbon M, Spitzer RL, Williams JBW. Structured Clinical Interview for DSM-IV Axis 1 Disorders. Biometrics Research: New York, 2002. 23. Patel T, Brewin CR, Wheatley J, Wells A, Fisher P, Myers S. Intrusive images and memories in major depression. Behav Res Ther 2007;45:25732580. 24. Horowitz M, Wilner N, Alvarez W. Impact of Events Scale: a measure of subjective stress. Psychosom Med 1979;41(3):209218. 25. Sundin EC, Horowitz MJ. Impact of Event Scale: psychometric properties. Br J Psychiatry 2002;180:205209. 26. Snaith RP, Zigmond AS. HADS: Hospital Anxiety and Depression Scale. NFER Nelson: Windsor, 1994. 27. Brewin CR, Christodoulides J, Hutchinson G. Intrusive thoughts and intrusive memories in a nonclinical sample. Cognition Emotion 1996;10(1):107112.

28. Hackmann A, Ehlers A, Speckens A, Clark DM. Characteristics and content of intrusive memories in PTSD and their changes with treatment. J Traumatic Stress 2004;17(3):231240. 29. Williams AD, Moulds ML. Negative appraisals and cognitive avoidance of intrusive memories in depression: a replication and extension. Depression Anxiety 2007;0:18. 30. Michael T, Ehlers A, Halligan SL, Clark DM. Unwanted memories of assault: what intrusion characteristics are associated with PTSD? Behav Res Ther 2005;43:613628. 31. Speckens AEM, Ehlers A, Hackmann A, Ruths FA, Clark DM. Intrusive memories and rumination in patients with post-traumatic stress disorder: a phenomenological comparison. Memory 2007;15: 249257. 32. Kosslyn SM, Ganis G, Thompson WL. Neural foundations of imagery. Nature Rev, Neurosci 2001;2: 635642. 33. Conway MA, Pleydell-Pearce CW. The construction of autobiographical memories in the self-memory system. Psychol Rev 2000;107(2):261288. 34. Brewin CR, Dalgleish T, Joseph S. A dual representation theory of post traumatic stress disorder. Psychol Rev 1996;103(4):670686. 35. Kosslyn SM, Thompson WL, Constantini-Ferrando MF, Alpert NM, Spiegel MD. Hypnotic visual illusion alters color processing in the brain. Amer J Psychiatry 2000;1557:12791284. 36. Gurevich M, Devins GM, Rodin GM. Stress response syndromes and cancer: conceptual and assessment issues. Psychosomatics 2002;43(4):259281. 37. ODonnell ML, Elliott P, Wolfgang BJ, Creamer M. Posttraumatic appraisals in the development and persistence of posttraumatic stress symptoms. J Traumatic Stress 2007;20(2):173182.

Copyright r 2009 John Wiley & Sons, Ltd.

Psycho-Oncology 18: 11471155 (2009) DOI: 10.1002/pon