- Email: [email protected]
Pathogenesis of three lesions causing lameness of the foreleg in horses
Attention Training Reduces Intrusive Thoughts Cued by a Narrative of Stressful Life Events: A Controlled Study Yasmine Nassif1 and Adrian Wells1,2 1 2
University of Manchester, UK NTNU Trondheim, Norway
Objectives:
Intrusive thoughts are common distressing symptoms that occur after exposure to stressful life events. This study tested the idea, based on metacognitive theory, that such intrusions may be ameliorated by the Attention Training Technique (ATT; Wells, 1990). Method: Participants who reported distressing intrusions were randomly allocated to two sessions of ATT (plus homework) or a filler task and were exposed to a narrative recording of their traumatic experience before and after the intervention. Frequencies of intrusions occurring during the narrative were measured. A measure of self-reported attention flexibility was also examined. Results: Participants in the ATT condition showed a significant reduction in intrusion frequency and an increase in attention flexibility compared with the control group. ATT was associated with a 3.4 times greater reduction in the incidence of intrusions. Conclusions: ATT may be an effective technique for reducing symptoms of stress exposure, reducing the incidence of intrusions when exposed to stimuli associated with stressful C 2013 Wiley Periodicals, Inc. J. Clin. Psychol. 70:510–517, 2014. events. Keywords: attention training; metacognition; attention flexibility; stressful life events; intrusive thoughts
Intrusive thoughts are a common symptom caused by normal life stresses such as bereavement and by more extreme negative events (Durham, McCammon, & Allison, 1985; Gold, Marx, Soler-Baillo, & Sloan, 2005). They are often unwanted, can be cued or uncued, and are associated with significant distress. Intrusions appear to predict long-term responses to stress (Davidson & Baum, 1993) and may be related to biological outcomes (Baum, Cohen, & Hall, 1993). The metacognitive model of stress-related intrusions (Wells, 2009) proposes that intrusive thoughts and other symptoms such as increased arousal prime mental simulations of coping with stressful events such that basic control programs can be established for regulating cognition and action in the future. Such symptoms normally subside as cognition is retuned to threat-free processing. However, intrusions and stress symptoms persist when the individual engages in a style of thinking that strengthens threat-related processing configurations. The style of most importance is the Cognitive-Attentional Syndrome (CAS), which comprises worry/rumination, a preoccupation with memory, maintaining attention on potential threat, and coping behaviours that ironically impair mental control. This thinking style maintains the perception of threat and subcortical fear networks fail to be downregulated. Consistent with these proposed effects, studies demonstrate that the tendency to engage extended processes of worry/rumination is associated with higher levels of intrusive thoughts in the days after analogue stress exposure (Wells & Papageorgiou, 1995). Furthermore, the tendency to use worry to control unwanted thoughts differentiates individuals with acute stress from those without after traffic accidents (Warda & Bryant, 1998) and is a prospective predictor of the development of posttraumatic stress disorder (PTSD; Holeva, Tarrier, & Wells, 2001) or stress symptoms (Roussis & Wells, 2008). In metacognitive therapy (Wells, 2009) techniques have been designed to modify metacognitive control and enhance flexibility so that individuals can refrain from responding to intrusions
Please address correspondence to: Adrian Wells, Academic Division of Clinical Psychology, University of Manchester, Rawnsley Building, MRI, Oxford Road, Manchester, M13 9WL. E-mail: [email protected] JOURNAL OF CLINICAL PSYCHOLOGY, Vol. 70(6), 510–517 (2014) Published online in Wiley Online Library (wileyonlinelibrary.com/journal/jclp).
C 2013 Wiley Periodicals, Inc. DOI: 10.1002/jclp.22047
Attention Training and Intrusive Thoughts
511
with extended processing in the form of the CAS. One such technique is Attention Training (Wells, 1990, 2007), the effects of which have been explored in cases of panic disorder (Wells, 1990), social phobia (Wells, White, & Carter, 1997), hypochondriasis (Papageorgiou & Wells, 1998), recurrent major depression (Papageorgiou & Wells, 2000), and psychosis (Valmaggia, Bouman, & Schuurman, 2007). In each case it was associated with significant improvement in symptoms. Moreover, there is preliminary evidence that an augmented version is associated with neuropsychological changes detectable with functional magnetic resonance imaging in depressed patients (Siegle, Ghinassi, & Thase, 2007). However, each of these studies used small samples or case-series. Furthermore, no studies have yet examined the possible effects of the technique on stress symptoms. Attention Training Technique (ATT) comprises a set of auditory stimuli (normally six to eight) presented simultaneously at different spatial locations over a period of about 12 minutes. The person is instructed to focus attention on selected sounds (selective attention) or spatial locations, and then to switch attention between different sounds (attention switching) and locations, before allocating attention to all sounds simultaneously (divided attention). The rationale for ATT emphasizes that the technique is not to be a distraction from or a means of controlling inner experiences. It is to increase metacognitive awareness and enable inner events such as thoughts to occur as “background noise” without resorting to worry, rumination, suppression, or any form of coping. Homework practise of ATT is usually prescribed to strengthen metacognitive change. In the present study, we sought to examine the effects of ATT on intrusions in a sample that had experienced stressful life events. As a paradigm, we were specifically interested in intrusions cued by a narrative of the events as this afforded some control over the current environment. We hypothesized that participants receiving ATT would have fewer intrusions than a control group when each group was exposed to the narrative of their stressful event. We also aimed to explore the effects of the technique on metacognition and specifically self-reported attentional flexibility. We hypothesized that the ATT group would show greater subjective ratings of flexibility after intervention than the control group. As a control group, we used an attention filler task in place of the ATT training sessions; this was included to control for effects of distraction before and after exposure to the narrative that might otherwise account for the effects of ATT.
Method Design This study used an experimental design with one experimental group and one control group. Each participant was tested on two separate occasions with an average between-session interval of 1.93 days (standard deviation [SD] = 1.23). The experimental group received two sessions of ATT and practised homework between the sessions. The control group received a filler task that comprised digit and letter detection. The study was approved by the University of Manchester School of Psychological Sciences research ethics committee.
Participants Fifty-two subjects responded to an advertisement for the study asking for volunteers who had experienced a stressful life event that still caused distress. Of those, 42 met the study inclusion criteria and entered the study. Inclusion criteria were as follows: (a) having experienced a stressful life event in the past that still caused distress when remembered; (b) the event occurred at least a month or more prior to the study (to rule out the presence of acute stress and a higher probability of spontaneous recovery); and (c) a minimum level of distress of 30% on narrating the traumatic event was specified to reduce floor effects. Of the 42 subjects, 32 were female (76.2%). Age range was 18 to 48 years (mean [M] = 22.05; SD = 5.55). Participants were randomly allocated to either two sessions of ATT or two sessions of the distraction task. There were 21 participants in each group. In the experimental group,
512
Journal of Clinical Psychology, June 2014
there were 15 females (71.4%). Participants were given the choice of receiving a small payment or course credit for their participation.
Procedure and Materials All subjects were tested individually. On arrival at the laboratory they first completed the Detached Mindfulness Questionnaire (DMQ; Nassif & Wells, 2007). This measure has five subscales that assess level of meta-awareness and nature of relationship with thoughts. These dimensions of metacognition are conceptualized as adaptive or maladaptive in the metacognitive model of psychological disorder. The subscales and internal consistencies for the current sample are as follows: detachment/observing self (e.g., “I can step back from my thoughts and see them as separate from me”; alpha = .76), meta-awareness (e.g., “I usually know what I’m thinking about if someone asks me”; alpha = .62), thought control (e.g., “ I usually try to control or stop my thoughts even when they are upsetting”; alpha = .62), attention flexibility/low conceptual processing (e.g., “I can concentrate on my work even if I’m worried about something”; alpha = .77), and cognitive decentring (e.g., “I usually know when my thoughts don’t mean anything”; alpha = .50). In this study, we were specifically interested in the attention flexibility/low conceptual processing subscale. Subjects were then asked to narrate their stressful event, which was audio-recorded. They were instructed that they would not be interrupted or asked about the event and to use the present tense when narrating. They were asked the following questions, based on an instruction used by Foa, Molnar, and Cashman (1995, p. 679): “Tell me about the event in as much detail as you can remember. This includes details about the surroundings, your activities, the other peoples activities, how you felt and what your thoughts were during the [stressful event].” The level of distress experienced while narrating was measured to determine that participants met the study inclusion criterion of a score of at least 30%. Distress was assessed with the following item: “How distressing was it for you to narrate the stressful episode?” Responses were provided on a 11-point scale ranging from 0% (not at all) to 100% (very much). Ten participants scored less than 30% and were excluded from continuation in the study. After a 5-minute “cooling down” period, participants listened to their narrative. They were asked to rest for 5 minutes before being asked to give a rating of the number of intrusions during exposure to the narrative. This measure consisted of the following item: “Place a circle around the number on the scale below to indicate the number of involuntary thoughts you had about your personal stressor while listening to the narrative.” Responses were given on a point-interval scale ranging from 0 (none) to 12 (or more). Next, a self-attention rating was completed in which responses to the item “At this moment in time how much is your attention focused on yourself or on your external environment” were given on a bipolar scale ranging from −3 (entirely externally focused) to +3 (entirely self-focused). This rating forms part of the ATT protocol (Wells, 2009) and is used as a marker that attention is being diverted away from CAS-related processes. The ATT or filler task was then administered and participants engaged in their respective attention procedure for a period of 10 minutes. In the ATT condition subjects listened to and followed the directions on a recording. In the control condition, participants were asked to circle each instance of a specified target letter or digit in a set of random letter matrices. The control task was included here to control for time and the possible effects of distraction caused by ATT, which would reduce carry-over processing linked to the previous narrative exposure. After these manipulations the self-attention rating was readministered. Participants in the ATT group were also asked to complete a compliance measure, which consisted of the item: “How much were you able to engage in the techniques when instructed? Please choose a percentage.” The response scale ranged from 0% (not at all) to 100% (completely). The ATT group were given the recording of the ATT and were asked to practise this at least once before the second session. The ATT group received homework because we were uncertain about the lowest effective dose of ATT. The control group did not receive any homework instructions. At the second session, the ATT group indicated the number of times they had practised ATT. All participants completed the self-attention rating and then engaged in 10 minutes of ATT
Attention Training and Intrusive Thoughts
513
or the letter-finding task. The control task was used here to control for effects of distraction that might affect levels of anticipatory processing prior to exposure to the narrative. The selfattention rating was readministered along with the compliance rating. Participants then listened to their narrative and the intrusion measure was administered as in session 1. Finally, the DMQ was given and participants were debriefed.
Overview of Statistical Analyses Independent sample t tests were run to examine any pretraining group differences in distress, narrative duration, intrusions, and DMQ-attention flexibility scores. Analysis of covariance (ANCOVA) and negative binomial regression (NBR) were used to compare group differences in intrusions, while controlling for this variable at the start of the experiment and to control for the duration of the narrative (exposure time) when testing group differences. Differences in attention flexibility were tested with ANCOVA.
Results Types of Stressors Reported Stressful events reported by participants were broad and included the death of a close family member by natural causes or disease (N = 9), death of a loved one by accident (N = 6), receiving the initial shock of discovering a serious illness to self or to significant others (N = 9), having been involved in an accident (N = 10), and having survived a difficult period in the past that brought extreme changes to self (such as emotionally abusive relationship, unwanted pregnancy, and moving to a new country; N = 8).
Preintervention Differences Distress from narrative. All participants reported being distressed by listening to their narrative. Ratings ranged from 30% to 100% and there were no group differences in distress level when listening to the narrative at the outset of the study, t(40) = 1.38, p = .177. For the experimental group, M = 65.48, SD = 19.23, and for the control group, M = 57.14, SD = 20.04. The magnitude of the difference was small to moderate (η2 = .05). Duration of narrative. There was no significant group difference in the duration of the narrative (in seconds) across the two groups (ATT group: M = 392.48, SD = 306.03, 95% confidence interval [CI] between 253.17 and 531.78; control group: M = 282.48, SD = 124.72, 95% CI between 225.70 and 339.25); t(26.47) = 1.53, p = .139, η2 = .055. However, it was important to control for the duration of narrative when comparing groups on distress and frequency of intrusions as different exposure times may affect the rate of intrusions and the degree of attention flexibility measured. Intrusions and attention flexibility. Intrusions occurred across the two groups during narrative exposure. In the first session, in the ATT group (before any intervention), the mean score for intrusions was 3.24 (SD = 2.05, 95% CI between 2.31 and 4.17). In the control group, the mean score for intrusions was 3.00 (SD = 1.82, 95% CI between 2.17 and 3.83). This difference was not statistically significant, t(40) = .40, p = .69. Independent samples t tests showed no statistical difference between the groups at preintervention on the five subscales of the DMQ, including the one of main importance-attention flexibility (experimental group: M = 14.33, SD = 4.48, 95% CI between 12.30 and 16.37; control group: M = 14.81 SD = 4.60, 95% CI between 12.72 and 16.90); t(40) = −.34, p = .74. The magnitude of the difference in means was very small. Attention rating, compliance, and homework. In the first session, all participants in the ATT group reported an external shift in attention. All but two participants reported a two-point
514
Journal of Clinical Psychology, June 2014
Table 1 Frequency of Intrusion Occurrences Reported During Narrative Exposure in the Experimental and Control Group: Premanipulation (Session 1) and Postmanipulation (Session 2) Experimental (ATT) group N = 21 Number of intrusions 0 1 2 3 4 5 6 7 8 9
Control (filler task) group N = 21
Session 1 (N = 21)
Session 2 (N = 21)
Session 1 (N = 21)
Session 2 (N = 21)
0 4 3 7 5 0 0 0 1 1
6 9 4 0 1 0 1 0 0 0
2 3 3 5 3 3 2 0 0 0
2 5 5 2 3 2 1 0 0 1
shift on the Self-Attention Rating Scale, while the other two reported a shift of only one point. In the second session, all except two participants reported an external shift in attention. The other two persons reported no change in attention rating. No participant reported an internal shift of attention after ATT. The pattern of attention rating was random in the control group in the two sessions. In the first session, 11 people reported an external shift in attention, six people reported no change, and one person reported an internal shift of attention after completing the task. In the second session, three people reported an external shift in attention, seven people reported no change, and eight people reported a shift toward self-focused attention. There was no difference between the groups in direction of attention before the manipulation at time 1 (ATT group: M = 1.29, SD = 1.01, 95% CI between .83 and 1.74; control group; M = .76, SD = 1.38, 95% CI between .14 and 1.39); t(40) = 1.41, p = .17. There was no difference before the manipulation at time 2 (ATT group: M = .19, SD = 1.08, 95% CI between −.68 and .30; control: M = .43, SD = 1.5, 95% CI between −1.1 and .26); t(40) = .59, p = .56. However, after the manipulation, the differences were significant at time 1 (ATT group: M = −1.14, SD = 1.28, 95% CI between −1.72 and −.56; control group: M = .05, SD = 1.66, 95% CI between −.80 and .71); t(40) = −2.40, p = .02. There were also differences at time 2 (ATT group: M = −1.71, SD = 1.06, 95% CI between −2.19 and −1.23: control group: M = −.10, SD = 1.20, 95% CI between −.69 and .50); t(40) = −4.43, p < .0005. These data suggest that the ATT was operating as intended. Compliance with ATT appeared to be good and ranged between 50% and 90% in the first session, and between 50% and 100% in the second session. Homework practice ranged from one to three sessions (mean = 1.81), which brings the total number of ATT sessions to three to five (mean = 3.81).
Primary Variables: Postintervention effects Intrusions. Intrusion scores decreased at the end of the second session to 1.29 (SD = 1.46, 95% CI between .62 and 1.95) for the ATT group and to 2.76 (SD = 2.21, 95% CI between 1.76 and 3.77) for the control group. Descriptive statistics for frequency of intrusions in the two groups across the two sessions are presented in Table 1. All except one participant in the ATT group reported fewer intrusions associated with their narrative in the second session. This participant reported an additional three intrusions. Six participants who received ATT reported no intrusions in session 2. The occurrence of intrusions in the control group seemed to be random, as some participants had more intrusions (N = 9), some had the same number (N = 3), and some had fewer intrusions (N = 9).
Attention Training and Intrusive Thoughts
515
Table 2 Mean Rate of Intrusions for the ATT and Control Group in the Two Sessions Experimental group Intrusion per second
Control Group Intrusion per second
95% CI
Session 1 Session 2
95% CI
Mean
SD
Lower
Upper
Mean
SD
Lower
Upper
.010 .004
.005 .004
.008 .002
.013 .006
.013 .012
.010 .013
.008 .006
.017 .012
Note. SD = standard deviation; CI = confidence interval.
Because the narratives were of different lengths and this may have influenced the number of intrusions reported, it was more appropriate to express the occurrence of intrusions in terms of a ratio, which would take duration of narrative into consideration. The mean rates of intrusions (number of intrusions divided by duration of narrative in seconds) are presented in Table 2. These results seem to suggest a threefold drop in intrusions in the ATT group at the end of the second session, and no significant change in the rate of intrusions in the control group. Because the occurrence of intrusions (seen as a count variable) reported by participants can never assume a negative value, and intrusions are viewed in terms of occurrence or nonoccurrence of an event, i.e., incidence rate of intrusions, this is a special type of Poisson distribution (Walters, 2007). There are limitations of the Poisson such as the inability of this distribution to adjust for the presence of a zero count in the analyses and the assumption that the rate of intrusions is equal among participants in the same group, which is not necessarily true in our case. A more suitable analysis of incidence rate which adjusts for these limitations is provided by NBR (Walters, 2007). We therefore tested the effect of ATT on rate of intrusions (incidence rate) after manipulation using NBR. Incidence rate at the first session was also controlled to adjust for initial rate of intrusion scores. The independent variable was type of group. Results from the NBR indicated an incidence rate ratio of 3.40 (p = .001). This means that the reduction in the incidence of intrusions is more than three times greater in the group that received ATT, compared to the control group. This difference was significant (p = .001), with a 95% confidence interval ranging between 1.88 and 6.16. These results indicate that an average of 3.81 sessions of ATT lowered the occurrence of intrusions when compared with a control group when exposure time and initial incidence rate were controlled.
Metacognition Scores To test for the effects of ATT on Attention Flexibility while controlling for this variable at the first session, as well as for exposure time, a between-groups ANCOVA was run entering Attention Flexibility at the second session as the outcome variable. Results yielded significant group differences in Attention Flexibility. Attention Flexibility increased significantly in the group that received ATT, when compared to the control group, F(1,38) = 5.19, p < .05. The mean for this group was 15.42 (standard error [SE] = .56, 95% CI between 14.28 and 16.56) and 13.58 for the control group (SE = .56, 95% CI between 12.43 and 14.72). The effect size was large (partial η2 = .12). These results are consistent with the hypothesis that ATT increases Attention Flexibility, but the mechanism of such an effect remains to be demonstrated. Further exploratory ANCOVAs were conducted on the other four metacognition subscales. No differences emerged on these dimensions, suggesting that the ATT effects on metacognition might be specific or that other dimensions of the measure may be less sensitive to change.
Discussion This study is the first to show that ATT is associated with improvement in intrusive thoughts associated with a narrative of a stressful life event. Intrusions decreased significantly more in the
516
Journal of Clinical Psychology, June 2014
group that received ATT, providing support for our hypothesis. The intervention also appears to have affected self-report attentional flexibility as measured by the DMQ. By including a control condition, we can assume that the observed effects of the ATT package are more than the result of being exposed to trauma material (narrative), as both groups narrated and listened to this material, and more than undertaking an attention (distraction) task before and after exposure to their narrative or due to passage of time and repeated testing. However, the introduction of homework in the ATT condition presents a potential confound because these participants practised more and it may have led to demand factors such as these participants expecting symptom improvement. We do not know how many doses of ATT are needed to establish optimal effects as there were few training sessions in this study, and this is an area that requires further evaluation. In addition, the longer term effects of ATT were not measured and we do not know if they are stable. Our aim at this stage was to answer a much more basic question: Does ATT have an effect on cued intrusions? Previous studies on patients with disorders such as depression and hypochondriasis (e.g., Papageorgiou & Wells, 1998, 2000) have indicated longer term effects of more intensive training but the effects in stress-exposure require future investigation. There are several other limitations in this study that may affect the generalizability of results. First, there was a skewed gender distribution and small participant numbers that do not allow for gender comparisons of the effects of ATT or for more thorough examination of change scores. Second, we do not know what the effects are on wider measures of stress symptoms. Third, the sample was drawn from a nontreatment-seeking group, and the effects of ATT on individuals with PTSD and patients cannot be inferred from the present results. From a methodological perspective, we asked participants to give a retrospective (after 5 minutes) rating of the number of intrusions experienced while listening to their narrative. Retrospective accounts may not be entirely accurate and other methods of measuring intrusion frequency, for example, with a handheld counter during the narrative, might have been used. We decided against this strategy in the present study as this introduces a concurrent task and might affect attention, thus obscuring and confounding any effects of the ATT. In conclusion, this study provides preliminary evidence that ATT might have a significant effect on intrusive thoughts in individuals experiencing distressing stress-related symptoms. Moreover, the procedure appears to modify self-report attention flexibility, but it remains to be established if this metacognition factor can be detected on objective measures and if it relates to symptom change.
References Baum, A., Cohen, L., & Hall, M. (1993). Control and intrusive memories as possible determinants of chronic stress. Psychosomatic Medicine, 55, 274–286. Davidson, L., & Baum, A. (1993). Predictors of chronic stress among Vietnam veterans: Stressor exposure and intrusive recall. Journal of Traumatic Stress, 6, 195–212. Durham, T. W., McCammon, S. L., & Allison, E. J. (1985). The psychological impact of disaster on rescue personnel. Annals of Emergency Medicine, 14, 664–668. Foa, E., Molnar, C., & Cashman, L. (1995). Change in rape narratives during exposure therapy for posttraumatic stress disorder. Journal of Traumatic Stress, 8, 675–690. Gold, S. D., Marx, B. P., Soler-Baillo, J. M., & Sloan, D. M. (2005). Is life stress more traumatic than traumatic stress? Journal of Anxiety Disorders, 19, 687–698. Holeva, V., Tarrier, N., & Wells, A. (2001). Prevalence and predictors of acute stress disorder and PTSD following road traffic accidents: Thought control strategies and social support. Behavior Therapy, 32, 65–83. Nassif, Y., & Wells, A. (2007). The Detached Mindfulness Questionnaire. Unpublished self-report scale, University of Manchester, UK. Papageorgiou, C., & Wells, A. (1998). Effects of attention training on hypochondriasis: A brief case series. Psychological Medicine, 28, 193–200. Papageorgiou, C., & Wells, A. (2000). Treatment of recurrent major depression with attention training, Cognitive and Behavioral Practice, 7, 407–413.
Attention Training and Intrusive Thoughts
517
Roussis, P., & Wells, A. (2008). Psychological factors predicting stress symptoms: Metacognition, thought control and varieties of worry. Anxiety, Stress and Coping, 21, 213–225. Siegle, G. J, Ghinassi, F., & Thase, M. E. (2007). Neurobehavioral therapies in the 21st century: Summary of an emerging field and an extended example of cognitive control training for depression. Cognitive Therapy and Research, 31, 235–262. Valmaggia, L., Bouman, T. K., & Schuurman, L. (2007). Attention training with auditory hallucinations: A case study. Cognitive and Behavioral Practice, 14, 127–133. Walters, G. D. (2007). Using Poisson class regression to analyze count data in correctional and forensic psychology. A relatively old solution to a relatively new problem. Criminal Justice and Behavior, 34, 1659–1674. Warda, G. & Bryant, R. A. (1998). Thought control strategies in acute stress disorder. Behaviour Research and Therapy, 36, 1171–1175. Wells, A. (1990). Panic disorder in association with relaxation induced anxiety: An attentional training approach to treatment. Behavior Therapy, 21, 273–280. Wells, A. (2007). The Attention Training Technique: Theory, effects and a metacognitive hypothesis on auditory hallucinations. Cognitive and Behavioral Practice, 14, 134–138. Wells, A. (2009). Metacognitive therapy for anxiety and depression. New York, NY: Guilford Press. Wells, A., & Papageorgiou, C. (1995). Worry and the incubation of intrusive images following stress. Behaviour Research and Therapy, 33, 579–583. Wells, A., White, J., & Carter, K. (1997). Attention training: effects on anxiety and beliefs in panic and social phobia. Clinical Psychology and Psychotherapy, 4, 226–232.
University of Manchester, UK NTNU Trondheim, Norway
Objectives:
Intrusive thoughts are common distressing symptoms that occur after exposure to stressful life events. This study tested the idea, based on metacognitive theory, that such intrusions may be ameliorated by the Attention Training Technique (ATT; Wells, 1990). Method: Participants who reported distressing intrusions were randomly allocated to two sessions of ATT (plus homework) or a filler task and were exposed to a narrative recording of their traumatic experience before and after the intervention. Frequencies of intrusions occurring during the narrative were measured. A measure of self-reported attention flexibility was also examined. Results: Participants in the ATT condition showed a significant reduction in intrusion frequency and an increase in attention flexibility compared with the control group. ATT was associated with a 3.4 times greater reduction in the incidence of intrusions. Conclusions: ATT may be an effective technique for reducing symptoms of stress exposure, reducing the incidence of intrusions when exposed to stimuli associated with stressful C 2013 Wiley Periodicals, Inc. J. Clin. Psychol. 70:510–517, 2014. events. Keywords: attention training; metacognition; attention flexibility; stressful life events; intrusive thoughts
Intrusive thoughts are a common symptom caused by normal life stresses such as bereavement and by more extreme negative events (Durham, McCammon, & Allison, 1985; Gold, Marx, Soler-Baillo, & Sloan, 2005). They are often unwanted, can be cued or uncued, and are associated with significant distress. Intrusions appear to predict long-term responses to stress (Davidson & Baum, 1993) and may be related to biological outcomes (Baum, Cohen, & Hall, 1993). The metacognitive model of stress-related intrusions (Wells, 2009) proposes that intrusive thoughts and other symptoms such as increased arousal prime mental simulations of coping with stressful events such that basic control programs can be established for regulating cognition and action in the future. Such symptoms normally subside as cognition is retuned to threat-free processing. However, intrusions and stress symptoms persist when the individual engages in a style of thinking that strengthens threat-related processing configurations. The style of most importance is the Cognitive-Attentional Syndrome (CAS), which comprises worry/rumination, a preoccupation with memory, maintaining attention on potential threat, and coping behaviours that ironically impair mental control. This thinking style maintains the perception of threat and subcortical fear networks fail to be downregulated. Consistent with these proposed effects, studies demonstrate that the tendency to engage extended processes of worry/rumination is associated with higher levels of intrusive thoughts in the days after analogue stress exposure (Wells & Papageorgiou, 1995). Furthermore, the tendency to use worry to control unwanted thoughts differentiates individuals with acute stress from those without after traffic accidents (Warda & Bryant, 1998) and is a prospective predictor of the development of posttraumatic stress disorder (PTSD; Holeva, Tarrier, & Wells, 2001) or stress symptoms (Roussis & Wells, 2008). In metacognitive therapy (Wells, 2009) techniques have been designed to modify metacognitive control and enhance flexibility so that individuals can refrain from responding to intrusions
Please address correspondence to: Adrian Wells, Academic Division of Clinical Psychology, University of Manchester, Rawnsley Building, MRI, Oxford Road, Manchester, M13 9WL. E-mail: [email protected] JOURNAL OF CLINICAL PSYCHOLOGY, Vol. 70(6), 510–517 (2014) Published online in Wiley Online Library (wileyonlinelibrary.com/journal/jclp).
C 2013 Wiley Periodicals, Inc. DOI: 10.1002/jclp.22047
Attention Training and Intrusive Thoughts
511
with extended processing in the form of the CAS. One such technique is Attention Training (Wells, 1990, 2007), the effects of which have been explored in cases of panic disorder (Wells, 1990), social phobia (Wells, White, & Carter, 1997), hypochondriasis (Papageorgiou & Wells, 1998), recurrent major depression (Papageorgiou & Wells, 2000), and psychosis (Valmaggia, Bouman, & Schuurman, 2007). In each case it was associated with significant improvement in symptoms. Moreover, there is preliminary evidence that an augmented version is associated with neuropsychological changes detectable with functional magnetic resonance imaging in depressed patients (Siegle, Ghinassi, & Thase, 2007). However, each of these studies used small samples or case-series. Furthermore, no studies have yet examined the possible effects of the technique on stress symptoms. Attention Training Technique (ATT) comprises a set of auditory stimuli (normally six to eight) presented simultaneously at different spatial locations over a period of about 12 minutes. The person is instructed to focus attention on selected sounds (selective attention) or spatial locations, and then to switch attention between different sounds (attention switching) and locations, before allocating attention to all sounds simultaneously (divided attention). The rationale for ATT emphasizes that the technique is not to be a distraction from or a means of controlling inner experiences. It is to increase metacognitive awareness and enable inner events such as thoughts to occur as “background noise” without resorting to worry, rumination, suppression, or any form of coping. Homework practise of ATT is usually prescribed to strengthen metacognitive change. In the present study, we sought to examine the effects of ATT on intrusions in a sample that had experienced stressful life events. As a paradigm, we were specifically interested in intrusions cued by a narrative of the events as this afforded some control over the current environment. We hypothesized that participants receiving ATT would have fewer intrusions than a control group when each group was exposed to the narrative of their stressful event. We also aimed to explore the effects of the technique on metacognition and specifically self-reported attentional flexibility. We hypothesized that the ATT group would show greater subjective ratings of flexibility after intervention than the control group. As a control group, we used an attention filler task in place of the ATT training sessions; this was included to control for effects of distraction before and after exposure to the narrative that might otherwise account for the effects of ATT.
Method Design This study used an experimental design with one experimental group and one control group. Each participant was tested on two separate occasions with an average between-session interval of 1.93 days (standard deviation [SD] = 1.23). The experimental group received two sessions of ATT and practised homework between the sessions. The control group received a filler task that comprised digit and letter detection. The study was approved by the University of Manchester School of Psychological Sciences research ethics committee.
Participants Fifty-two subjects responded to an advertisement for the study asking for volunteers who had experienced a stressful life event that still caused distress. Of those, 42 met the study inclusion criteria and entered the study. Inclusion criteria were as follows: (a) having experienced a stressful life event in the past that still caused distress when remembered; (b) the event occurred at least a month or more prior to the study (to rule out the presence of acute stress and a higher probability of spontaneous recovery); and (c) a minimum level of distress of 30% on narrating the traumatic event was specified to reduce floor effects. Of the 42 subjects, 32 were female (76.2%). Age range was 18 to 48 years (mean [M] = 22.05; SD = 5.55). Participants were randomly allocated to either two sessions of ATT or two sessions of the distraction task. There were 21 participants in each group. In the experimental group,
512
Journal of Clinical Psychology, June 2014
there were 15 females (71.4%). Participants were given the choice of receiving a small payment or course credit for their participation.
Procedure and Materials All subjects were tested individually. On arrival at the laboratory they first completed the Detached Mindfulness Questionnaire (DMQ; Nassif & Wells, 2007). This measure has five subscales that assess level of meta-awareness and nature of relationship with thoughts. These dimensions of metacognition are conceptualized as adaptive or maladaptive in the metacognitive model of psychological disorder. The subscales and internal consistencies for the current sample are as follows: detachment/observing self (e.g., “I can step back from my thoughts and see them as separate from me”; alpha = .76), meta-awareness (e.g., “I usually know what I’m thinking about if someone asks me”; alpha = .62), thought control (e.g., “ I usually try to control or stop my thoughts even when they are upsetting”; alpha = .62), attention flexibility/low conceptual processing (e.g., “I can concentrate on my work even if I’m worried about something”; alpha = .77), and cognitive decentring (e.g., “I usually know when my thoughts don’t mean anything”; alpha = .50). In this study, we were specifically interested in the attention flexibility/low conceptual processing subscale. Subjects were then asked to narrate their stressful event, which was audio-recorded. They were instructed that they would not be interrupted or asked about the event and to use the present tense when narrating. They were asked the following questions, based on an instruction used by Foa, Molnar, and Cashman (1995, p. 679): “Tell me about the event in as much detail as you can remember. This includes details about the surroundings, your activities, the other peoples activities, how you felt and what your thoughts were during the [stressful event].” The level of distress experienced while narrating was measured to determine that participants met the study inclusion criterion of a score of at least 30%. Distress was assessed with the following item: “How distressing was it for you to narrate the stressful episode?” Responses were provided on a 11-point scale ranging from 0% (not at all) to 100% (very much). Ten participants scored less than 30% and were excluded from continuation in the study. After a 5-minute “cooling down” period, participants listened to their narrative. They were asked to rest for 5 minutes before being asked to give a rating of the number of intrusions during exposure to the narrative. This measure consisted of the following item: “Place a circle around the number on the scale below to indicate the number of involuntary thoughts you had about your personal stressor while listening to the narrative.” Responses were given on a point-interval scale ranging from 0 (none) to 12 (or more). Next, a self-attention rating was completed in which responses to the item “At this moment in time how much is your attention focused on yourself or on your external environment” were given on a bipolar scale ranging from −3 (entirely externally focused) to +3 (entirely self-focused). This rating forms part of the ATT protocol (Wells, 2009) and is used as a marker that attention is being diverted away from CAS-related processes. The ATT or filler task was then administered and participants engaged in their respective attention procedure for a period of 10 minutes. In the ATT condition subjects listened to and followed the directions on a recording. In the control condition, participants were asked to circle each instance of a specified target letter or digit in a set of random letter matrices. The control task was included here to control for time and the possible effects of distraction caused by ATT, which would reduce carry-over processing linked to the previous narrative exposure. After these manipulations the self-attention rating was readministered. Participants in the ATT group were also asked to complete a compliance measure, which consisted of the item: “How much were you able to engage in the techniques when instructed? Please choose a percentage.” The response scale ranged from 0% (not at all) to 100% (completely). The ATT group were given the recording of the ATT and were asked to practise this at least once before the second session. The ATT group received homework because we were uncertain about the lowest effective dose of ATT. The control group did not receive any homework instructions. At the second session, the ATT group indicated the number of times they had practised ATT. All participants completed the self-attention rating and then engaged in 10 minutes of ATT
Attention Training and Intrusive Thoughts
513
or the letter-finding task. The control task was used here to control for effects of distraction that might affect levels of anticipatory processing prior to exposure to the narrative. The selfattention rating was readministered along with the compliance rating. Participants then listened to their narrative and the intrusion measure was administered as in session 1. Finally, the DMQ was given and participants were debriefed.
Overview of Statistical Analyses Independent sample t tests were run to examine any pretraining group differences in distress, narrative duration, intrusions, and DMQ-attention flexibility scores. Analysis of covariance (ANCOVA) and negative binomial regression (NBR) were used to compare group differences in intrusions, while controlling for this variable at the start of the experiment and to control for the duration of the narrative (exposure time) when testing group differences. Differences in attention flexibility were tested with ANCOVA.
Results Types of Stressors Reported Stressful events reported by participants were broad and included the death of a close family member by natural causes or disease (N = 9), death of a loved one by accident (N = 6), receiving the initial shock of discovering a serious illness to self or to significant others (N = 9), having been involved in an accident (N = 10), and having survived a difficult period in the past that brought extreme changes to self (such as emotionally abusive relationship, unwanted pregnancy, and moving to a new country; N = 8).
Preintervention Differences Distress from narrative. All participants reported being distressed by listening to their narrative. Ratings ranged from 30% to 100% and there were no group differences in distress level when listening to the narrative at the outset of the study, t(40) = 1.38, p = .177. For the experimental group, M = 65.48, SD = 19.23, and for the control group, M = 57.14, SD = 20.04. The magnitude of the difference was small to moderate (η2 = .05). Duration of narrative. There was no significant group difference in the duration of the narrative (in seconds) across the two groups (ATT group: M = 392.48, SD = 306.03, 95% confidence interval [CI] between 253.17 and 531.78; control group: M = 282.48, SD = 124.72, 95% CI between 225.70 and 339.25); t(26.47) = 1.53, p = .139, η2 = .055. However, it was important to control for the duration of narrative when comparing groups on distress and frequency of intrusions as different exposure times may affect the rate of intrusions and the degree of attention flexibility measured. Intrusions and attention flexibility. Intrusions occurred across the two groups during narrative exposure. In the first session, in the ATT group (before any intervention), the mean score for intrusions was 3.24 (SD = 2.05, 95% CI between 2.31 and 4.17). In the control group, the mean score for intrusions was 3.00 (SD = 1.82, 95% CI between 2.17 and 3.83). This difference was not statistically significant, t(40) = .40, p = .69. Independent samples t tests showed no statistical difference between the groups at preintervention on the five subscales of the DMQ, including the one of main importance-attention flexibility (experimental group: M = 14.33, SD = 4.48, 95% CI between 12.30 and 16.37; control group: M = 14.81 SD = 4.60, 95% CI between 12.72 and 16.90); t(40) = −.34, p = .74. The magnitude of the difference in means was very small. Attention rating, compliance, and homework. In the first session, all participants in the ATT group reported an external shift in attention. All but two participants reported a two-point
514
Journal of Clinical Psychology, June 2014
Table 1 Frequency of Intrusion Occurrences Reported During Narrative Exposure in the Experimental and Control Group: Premanipulation (Session 1) and Postmanipulation (Session 2) Experimental (ATT) group N = 21 Number of intrusions 0 1 2 3 4 5 6 7 8 9
Control (filler task) group N = 21
Session 1 (N = 21)
Session 2 (N = 21)
Session 1 (N = 21)
Session 2 (N = 21)
0 4 3 7 5 0 0 0 1 1
6 9 4 0 1 0 1 0 0 0
2 3 3 5 3 3 2 0 0 0
2 5 5 2 3 2 1 0 0 1
shift on the Self-Attention Rating Scale, while the other two reported a shift of only one point. In the second session, all except two participants reported an external shift in attention. The other two persons reported no change in attention rating. No participant reported an internal shift of attention after ATT. The pattern of attention rating was random in the control group in the two sessions. In the first session, 11 people reported an external shift in attention, six people reported no change, and one person reported an internal shift of attention after completing the task. In the second session, three people reported an external shift in attention, seven people reported no change, and eight people reported a shift toward self-focused attention. There was no difference between the groups in direction of attention before the manipulation at time 1 (ATT group: M = 1.29, SD = 1.01, 95% CI between .83 and 1.74; control group; M = .76, SD = 1.38, 95% CI between .14 and 1.39); t(40) = 1.41, p = .17. There was no difference before the manipulation at time 2 (ATT group: M = .19, SD = 1.08, 95% CI between −.68 and .30; control: M = .43, SD = 1.5, 95% CI between −1.1 and .26); t(40) = .59, p = .56. However, after the manipulation, the differences were significant at time 1 (ATT group: M = −1.14, SD = 1.28, 95% CI between −1.72 and −.56; control group: M = .05, SD = 1.66, 95% CI between −.80 and .71); t(40) = −2.40, p = .02. There were also differences at time 2 (ATT group: M = −1.71, SD = 1.06, 95% CI between −2.19 and −1.23: control group: M = −.10, SD = 1.20, 95% CI between −.69 and .50); t(40) = −4.43, p < .0005. These data suggest that the ATT was operating as intended. Compliance with ATT appeared to be good and ranged between 50% and 90% in the first session, and between 50% and 100% in the second session. Homework practice ranged from one to three sessions (mean = 1.81), which brings the total number of ATT sessions to three to five (mean = 3.81).
Primary Variables: Postintervention effects Intrusions. Intrusion scores decreased at the end of the second session to 1.29 (SD = 1.46, 95% CI between .62 and 1.95) for the ATT group and to 2.76 (SD = 2.21, 95% CI between 1.76 and 3.77) for the control group. Descriptive statistics for frequency of intrusions in the two groups across the two sessions are presented in Table 1. All except one participant in the ATT group reported fewer intrusions associated with their narrative in the second session. This participant reported an additional three intrusions. Six participants who received ATT reported no intrusions in session 2. The occurrence of intrusions in the control group seemed to be random, as some participants had more intrusions (N = 9), some had the same number (N = 3), and some had fewer intrusions (N = 9).
Attention Training and Intrusive Thoughts
515
Table 2 Mean Rate of Intrusions for the ATT and Control Group in the Two Sessions Experimental group Intrusion per second
Control Group Intrusion per second
95% CI
Session 1 Session 2
95% CI
Mean
SD
Lower
Upper
Mean
SD
Lower
Upper
.010 .004
.005 .004
.008 .002
.013 .006
.013 .012
.010 .013
.008 .006
.017 .012
Note. SD = standard deviation; CI = confidence interval.
Because the narratives were of different lengths and this may have influenced the number of intrusions reported, it was more appropriate to express the occurrence of intrusions in terms of a ratio, which would take duration of narrative into consideration. The mean rates of intrusions (number of intrusions divided by duration of narrative in seconds) are presented in Table 2. These results seem to suggest a threefold drop in intrusions in the ATT group at the end of the second session, and no significant change in the rate of intrusions in the control group. Because the occurrence of intrusions (seen as a count variable) reported by participants can never assume a negative value, and intrusions are viewed in terms of occurrence or nonoccurrence of an event, i.e., incidence rate of intrusions, this is a special type of Poisson distribution (Walters, 2007). There are limitations of the Poisson such as the inability of this distribution to adjust for the presence of a zero count in the analyses and the assumption that the rate of intrusions is equal among participants in the same group, which is not necessarily true in our case. A more suitable analysis of incidence rate which adjusts for these limitations is provided by NBR (Walters, 2007). We therefore tested the effect of ATT on rate of intrusions (incidence rate) after manipulation using NBR. Incidence rate at the first session was also controlled to adjust for initial rate of intrusion scores. The independent variable was type of group. Results from the NBR indicated an incidence rate ratio of 3.40 (p = .001). This means that the reduction in the incidence of intrusions is more than three times greater in the group that received ATT, compared to the control group. This difference was significant (p = .001), with a 95% confidence interval ranging between 1.88 and 6.16. These results indicate that an average of 3.81 sessions of ATT lowered the occurrence of intrusions when compared with a control group when exposure time and initial incidence rate were controlled.
Metacognition Scores To test for the effects of ATT on Attention Flexibility while controlling for this variable at the first session, as well as for exposure time, a between-groups ANCOVA was run entering Attention Flexibility at the second session as the outcome variable. Results yielded significant group differences in Attention Flexibility. Attention Flexibility increased significantly in the group that received ATT, when compared to the control group, F(1,38) = 5.19, p < .05. The mean for this group was 15.42 (standard error [SE] = .56, 95% CI between 14.28 and 16.56) and 13.58 for the control group (SE = .56, 95% CI between 12.43 and 14.72). The effect size was large (partial η2 = .12). These results are consistent with the hypothesis that ATT increases Attention Flexibility, but the mechanism of such an effect remains to be demonstrated. Further exploratory ANCOVAs were conducted on the other four metacognition subscales. No differences emerged on these dimensions, suggesting that the ATT effects on metacognition might be specific or that other dimensions of the measure may be less sensitive to change.
Discussion This study is the first to show that ATT is associated with improvement in intrusive thoughts associated with a narrative of a stressful life event. Intrusions decreased significantly more in the
516
Journal of Clinical Psychology, June 2014
group that received ATT, providing support for our hypothesis. The intervention also appears to have affected self-report attentional flexibility as measured by the DMQ. By including a control condition, we can assume that the observed effects of the ATT package are more than the result of being exposed to trauma material (narrative), as both groups narrated and listened to this material, and more than undertaking an attention (distraction) task before and after exposure to their narrative or due to passage of time and repeated testing. However, the introduction of homework in the ATT condition presents a potential confound because these participants practised more and it may have led to demand factors such as these participants expecting symptom improvement. We do not know how many doses of ATT are needed to establish optimal effects as there were few training sessions in this study, and this is an area that requires further evaluation. In addition, the longer term effects of ATT were not measured and we do not know if they are stable. Our aim at this stage was to answer a much more basic question: Does ATT have an effect on cued intrusions? Previous studies on patients with disorders such as depression and hypochondriasis (e.g., Papageorgiou & Wells, 1998, 2000) have indicated longer term effects of more intensive training but the effects in stress-exposure require future investigation. There are several other limitations in this study that may affect the generalizability of results. First, there was a skewed gender distribution and small participant numbers that do not allow for gender comparisons of the effects of ATT or for more thorough examination of change scores. Second, we do not know what the effects are on wider measures of stress symptoms. Third, the sample was drawn from a nontreatment-seeking group, and the effects of ATT on individuals with PTSD and patients cannot be inferred from the present results. From a methodological perspective, we asked participants to give a retrospective (after 5 minutes) rating of the number of intrusions experienced while listening to their narrative. Retrospective accounts may not be entirely accurate and other methods of measuring intrusion frequency, for example, with a handheld counter during the narrative, might have been used. We decided against this strategy in the present study as this introduces a concurrent task and might affect attention, thus obscuring and confounding any effects of the ATT. In conclusion, this study provides preliminary evidence that ATT might have a significant effect on intrusive thoughts in individuals experiencing distressing stress-related symptoms. Moreover, the procedure appears to modify self-report attention flexibility, but it remains to be established if this metacognition factor can be detected on objective measures and if it relates to symptom change.
References Baum, A., Cohen, L., & Hall, M. (1993). Control and intrusive memories as possible determinants of chronic stress. Psychosomatic Medicine, 55, 274–286. Davidson, L., & Baum, A. (1993). Predictors of chronic stress among Vietnam veterans: Stressor exposure and intrusive recall. Journal of Traumatic Stress, 6, 195–212. Durham, T. W., McCammon, S. L., & Allison, E. J. (1985). The psychological impact of disaster on rescue personnel. Annals of Emergency Medicine, 14, 664–668. Foa, E., Molnar, C., & Cashman, L. (1995). Change in rape narratives during exposure therapy for posttraumatic stress disorder. Journal of Traumatic Stress, 8, 675–690. Gold, S. D., Marx, B. P., Soler-Baillo, J. M., & Sloan, D. M. (2005). Is life stress more traumatic than traumatic stress? Journal of Anxiety Disorders, 19, 687–698. Holeva, V., Tarrier, N., & Wells, A. (2001). Prevalence and predictors of acute stress disorder and PTSD following road traffic accidents: Thought control strategies and social support. Behavior Therapy, 32, 65–83. Nassif, Y., & Wells, A. (2007). The Detached Mindfulness Questionnaire. Unpublished self-report scale, University of Manchester, UK. Papageorgiou, C., & Wells, A. (1998). Effects of attention training on hypochondriasis: A brief case series. Psychological Medicine, 28, 193–200. Papageorgiou, C., & Wells, A. (2000). Treatment of recurrent major depression with attention training, Cognitive and Behavioral Practice, 7, 407–413.
Attention Training and Intrusive Thoughts
517
Roussis, P., & Wells, A. (2008). Psychological factors predicting stress symptoms: Metacognition, thought control and varieties of worry. Anxiety, Stress and Coping, 21, 213–225. Siegle, G. J, Ghinassi, F., & Thase, M. E. (2007). Neurobehavioral therapies in the 21st century: Summary of an emerging field and an extended example of cognitive control training for depression. Cognitive Therapy and Research, 31, 235–262. Valmaggia, L., Bouman, T. K., & Schuurman, L. (2007). Attention training with auditory hallucinations: A case study. Cognitive and Behavioral Practice, 14, 127–133. Walters, G. D. (2007). Using Poisson class regression to analyze count data in correctional and forensic psychology. A relatively old solution to a relatively new problem. Criminal Justice and Behavior, 34, 1659–1674. Warda, G. & Bryant, R. A. (1998). Thought control strategies in acute stress disorder. Behaviour Research and Therapy, 36, 1171–1175. Wells, A. (1990). Panic disorder in association with relaxation induced anxiety: An attentional training approach to treatment. Behavior Therapy, 21, 273–280. Wells, A. (2007). The Attention Training Technique: Theory, effects and a metacognitive hypothesis on auditory hallucinations. Cognitive and Behavioral Practice, 14, 134–138. Wells, A. (2009). Metacognitive therapy for anxiety and depression. New York, NY: Guilford Press. Wells, A., & Papageorgiou, C. (1995). Worry and the incubation of intrusive images following stress. Behaviour Research and Therapy, 33, 579–583. Wells, A., White, J., & Carter, K. (1997). Attention training: effects on anxiety and beliefs in panic and social phobia. Clinical Psychology and Psychotherapy, 4, 226–232.