Brief Cognitive Therapy for Panic Disorder A Randomized Controlled Trial

Transcription

1 Page 1 of 10 Journal of Consulting and Clinical Psychology August 1999 Vol. 67, No. 4, by the American Psychological Association For personal use only--not for distribution. Brief Cognitive Therapy for Panic Disorder A Randomized Controlled Trial David M. Clark Paul M. Salkovskis Ann Hackmann Adrian Wells John Ludgate Michael Gelder ABSTRACT Cognitive therapy (CT) is a specific and highly effective treatment for panic disorder (PD). Treatment normally involves hr sessions. In an attempt to produce a more costeffective version, a briefer treatment that made extensive use of between-sessions patient selfstudy modules was created. Forty-three PD patients were randomly allocated to full CT (FCT), brief CT (BCT), or a 3-month wait list. FCT and BCT were superior to wait list on all measures, and the gains obtained in treatment were maintained at 12-month follow-up. There were no significant differences between FCT and BCT. Both treatments had large (approximately 3.0) and essentially identical effect sizes. BCT required 6.5 hr of therapist time, including booster sessions. Patients' initial expectation of therapy success was negatively correlated with posttreatment panic anxiety. Cognitive measures at the end of treatment predicted panic anxiety at 12-month follow-up. Adrian Wells is now at the Department of Clinical Psychology, Manchester University, Manchester, United Kingdom. This research was funded by grants from the Medical Research Council of the United Kingdom and the Wellcome Trust. We are grateful to Hester Barrington-Ward, Sarah Durbin, Anke Ehlers, Melanie Fennell, Carolyn Fordham-Walker, Freda McManus, Anthony Morrison, and Christina Suraway for their assistance. Correspondence may be addressed to David M. Clark, Department of Psychiatry, University of Oxford, Warneford Hospital, Oxford, United Kingdom, OX3 7JX. Electronic mail may be sent to David.Clark@psych.ox.ac.uk Received: May 13, 1998 Revised: November 26, 1998 Accepted: November 30, 1998

2 Page 2 of 10 During the 1980s, several effective cognitive behavioral treatments for panic disorder were developed. The two that have been most extensively evaluated are the panic control treatment (PCT) developed by Barlow, Craske, and colleagues and the cognitive therapy program developed by Clark, Salkovskis, Beck, and colleagues. Although the two treatments differ in emphasis, they have many common ingredients and appear to be similarly effective. Originally they involved 12 to 15 one-hour sessions. In this format, one or both have been shown to be superior to equally credible relaxation-based treatment, supportive psychotherapy, alprazolam, imipramine, and placebo medication (see Barlow & Lehman, 1996, and Clark, 1996, for reviews ). Recently, researchers have started to investigate whether it might be possible to obtain similarly good results with briefer forms of the treatments. If so, scarce health care resources could be used to provide effective treatment for a larger number of patients. Two studies have investigated briefer forms of panic control therapy. Craske, Maidenberg, and Bystritsky (1995) found that a four-session version of PCT was more effective than four sessions of nondirective supportive therapy. However, the authors commented that the overall clinical outcome of the four-session version was considerably less than the outcome typically achieved with full PCT. Newman, Kenardy, Herman, and Taylor (1997) reported a small-scale pilot study in which full PCT was compared with 6 hr of palmtop-computer-assisted PCT. Both treatments produced significant improvement. However, at the end of treatment, but not at follow-up, a significantly greater proportion of patients were panic free after full PCT than after computer-assisted PCT. One study has attempted to abbreviate cognitive therapy. Black, Wesner, Bowers, and Gabel (1993) devised a shortened (eight-session) version of cognitive therapy, which included additional psychological procedures that they developed. This abbreviated version of cognitive therapy was not significantly different from placebo medication and achieved a panic-free rate (32% of the intention-to-treat sample), which is less than half the rate obtained in any other study of cognitive therapy (see Clark, 1996, for a review ). The present study represents a further attempt to develop a brief version of cognitive therapy. To maximize the amount of change achieved in each therapy session, we developed a set of self-study modules covering the main aspects of therapy and asked patients to complete the modules prior to therapy sessions. Design Method Patients were initially assigned to full cognitive therapy (FCT), brief cognitive therapy (BCT), or a wait-list control condition. Patients in FCT had up to 12 one-hour sessions in the first 3 months, whereas BCT patients had 5 sessions. Both groups had up to 2 booster sessions in the next 3 months. Patients on the wait list received no treatment for 3 months, after which they were assigned randomly to one of the two forms of cognitive therapy. Assessments, which included ratings completed by an independent assessor who was unaware of treatment allocation, were at pretreatment/wait list, posttreatment/wait list, 3-month posttreatment follow-up, and 12-month posttreatment follow-up. Patients All Oxfordshire general practitioners, psychiatrists, and psychologists were sent a letter requesting referrals

3 Page 3 of 10 for a study of psychological treatments for panic disorder. Referred patients were assessed by trained clinical psychologists using the Structured Clinical Interview for DSM III R ( Spitzer & Williams, 1986 ). Acceptance criteria, which were the same as in our previous trial of FCT ( Clark et al., 1994 ), were (a) Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev.; DSM III R ; American Psychiatric Association, 1987 ) criteria for panic disorder with no, mild, or moderate agoraphobic avoidance; (b) duration of current episode of at least 6 months; (c) at least three panic attacks in the 3 weeks prior to interview; (d) panic considered as the patient's main problem; (e) age 18 to 60 years; (f) willingness to accept random allocation; (g) no depressive disorder severe enough to require immediate psychiatric treatment; (h) no previous treatment with cognitive therapy or exposure therapy for panic disorder; (i) no use of medication or, if taking psychotropic medication, on a stable dose for at least 3 months with an agreement not to change dosage; (j) no evidence of organic mental disorder, schizophrenia, alcohol or drug dependence, cardiovascular disease, asthma, epilepsy, or pregnancy; and (k) record of at least one panic attack while keeping a daily panic diary during a postinterview 2-week baseline period. Forty-nine patients met entry criteria at interview. Four patients recorded zero panic attacks in their baseline diary and were dropped from the study. Two other patients did not complete the baseline diary and withdrew on their own. The remaining 43 patients were randomized. One patient (allocated to FCT) dropped out after one session, having indicated that she was much improved and could not arrange time off work for further sessions. All other patients completed treatment. Treatments Patients in FCT were offered up to 12 weekly 1-hr sessions and received a mean of 10.4 sessions ( SD = 2.1, range = 5 12). Patients who had fewer than 12 sessions either became panic free early in the treatment or missed some sessions because of scheduling difficulties. BCT patients had 5 sessions in 3 months: Session 1 was 1.5 hr, Sessions 2 4 were 1 hr each, and Session 5 was 0.5 hr. During the first 3 months of follow-up, both groups had a mean of 1.5 one-hour booster sessions. Total therapy and boostersession time for the two treatments was 11.9 hr for FCT and 6.5 hr for BCT. FCT. FCT was the same as in the Clark et al. (1994) study and comprised a mixture of cognitive techniques and behavioral experiments, all intended to modify misinterpretations of body sensations and the processes that maintain them. BCT. BCT was a modified version of the full treatment, which was developed during a year of intensive pilot work. The same range of procedures was used, but many were first introduced in self-study modules. Patients read the self-study modules and completed the written exercises and the homework outlined in the modules before discussing a module's topic with their therapist. Patients studied a different module before each of the first four sessions. Module 1 gave case illustrations of the panic attack vicious circle and used a series of questions about thoughts and feelings to help patients identify the vicious circle in one of their own recent panic attacks. Thought challenging was introduced, and ways in which attention to body cues, avoidance, and images might maintain negative interpretations of body sensations were explained. Module 2 focused on patients' worst fears about the sensations they experienced in attacks and helped them to generate alternative, noncatastrophic explanations for the sensations. Module 3 introduced safety behaviors, explained how they prevent cognitive change, and helped patients identify their own safety behaviors. Module 4

4 Page 4 of 10 reviewed the outcome of experiments in which patients were encouraged to drop safety behaviors during their attacks and in feared situations, helped patients identify triggers for the attacks, consolidated the alternative explanations that had already been developed, and outlined a relapse-prevention program. Supplementary handouts, each of which dealt with a common catastrophic thought (e.g., "I'll faint" or "I'm having a heart attack"), were also prepared and used at the therapist's discretion. The timing of the sessions was as follows: Session 1 occurred in Week 1, Session 2 in Week 3, Session 3 in Week 5, Session 4 in Week 8, and Session 5 in Week 11. As in FCT, patients were asked not to enter feared situations more than usual until Week 5. Therapists Four clinical psychologists with experience in the use of cognitive and behavioral treatments for anxiety served as therapists and administered both treatments. Before starting the trial, each had specific training in cognitive therapy for panic disorder and had at least one supervised practice case in each version of cognitive therapy. Regular individual supervision was provided throughout the trial. Measures Treatment credibility, panic attacks, general anxiety, agoraphobic avoidance, panic-related cognition, and depression were assessed with the same measures as used in the Clark et al. (1994) study. Table 1 provides a full listing. To be classified as panic free, patients had to record zero panic attacks in their diary during the last 2 weeks and be rated as panic free by the assessor. Statistical Analysis A two-step approach was adopted. First, as in the Clark et al. (1994) study, a single panic anxiety composite measure was created and analyzed. If the composite revealed significant between-groups differences, individual panic anxiety measures were also analyzed. Following Rosenthal and Rosnow's (1991) recommendation, we generated the composite by standardizing ( M = 0, SD = 1) patients' scores on each of the 11 panic anxiety measures and then averaging across the measures. Characteristics of Patients Results Patients' mean age was 34 years ( SD = 11.1). Mean duration of the current episode of panic disorder was 3.7 years (range = ). Sixty-two percent were female. Fifteen percent had no agoraphobic avoidance, 63% had mild agoraphobic avoidance, and 22% had moderate agoraphobic avoidance. Sixtytwo percent had previously received some form of treatment for emotional problems. Thirty-two percent were on a stable dose of a psychotropic medication (mainly low-potency benzodiazepines or betablockers). There were no significant differences between the groups in any of these characteristics. Suitability of Treatment and Expectation of Improvement Patients rated the two versions of cognitive therapy as equally logical (for FCT, M = 8.4, SD = 1.9; for BCT, M = 8.7, SD = 1.5), indicated that they would be equally likely to recommend them to a friend (for FCT, M = 7.4, SD = 1.7; for BCT, M = 7.8, SD = 1.3), and had equivalent expectations of improvement

5 Page 5 of 10 (for FCT, M = 8.8, SD = 1.6; for BCT, M = 9.1, SD = 1.4). Use of Self-Study Modules in BCT Inspection of the case notes indicated that BCT patients made extensive use of the self-study modules. All the patients completed written assignments in all four modules, except for 1 patient, who completed three modules. Effects of Treatment Table 1 shows patients' scores before and after treatment and during wait list. 1 We performed two-way (Group Time) repeated measures analyses of variance (ANOVAs) on these data. For the panic anxiety composite, and for every individual measure, there was a significant Group Time interaction. We used Tukey tests to compare the groups at each time point. At pretreatment/wait list, no tests were significant (all p s >.30), indicating that the groups did not differ before the start of treatment or wait list. At posttreatment/wait list, both FCT and BCT were superior to the wait-list control condition on every measure (all p s <.005). There were no significant differences between FCT and BCT (all p s >.35). Table 2 shows the percentages of patients who were panic free and the percentages who achieved high endstate functioning at each assessment. Following Craske, Brown, and Barlow's (1991) recommendation, we defined high end-state functioning as panic free and as an assessor-scored panic-related distress disability rating of 2 or less ("slight"). Chi-square analyses indicated that both FCT and BCT were superior to the wait-list control condition in terms of the proportion of patients who became panic free (for FCT, χ 2 [1, N = 14] = 11.8, p <.001; for BCT, χ 2 [1, N = 14] = 9.6, p <.01) and the proportion of patients who achieved high end-state functioning (for FCT, χ 2 [1, N = 14] = 15.0, p <.001; for BCT, χ 2 [1, N = 14] = 12.6, p <.001). FCT did not differ from BCT on either measure. Maintenance of Treatment Gains All treated patients provided 3- and 12-month follow-up data. Table 1 shows patients' scores at the followup assessments. To investigate whether the gains achieved in therapy were maintained, we compared panic anxiety composite scores from the follow-up assessments with patients' posttreatment scores using a Group (FCT vs. BCT) Time (posttreatment vs. 3-month follow-up vs. 12-month follow-up) ANOVA. There were no significant main effects or interactions, indicating that, for both FCT and BCT, the gains achieved in treatment were maintained at follow-up and that the two treatments did not differ during the follow-up period. 2 Table 2 shows the percentages of patients who were panic free and the percentages who achieved high endstate functioning at the 3- and 12-month follow-ups. Both groups of patients also maintained their gains on these measures. Additional Analyses To gain a clearer impression of the magnitude of the response to FCT and BCT, we calculated effect sizes for the panic anxiety composite (see Table 3 ). The two treatments had large (approximately 3.0) and essentially identical effect sizes. To further explore the apparent equivalence of the two treatments, we adopted Rogers, Howard, and Vessey's (1993) confidence interval approach to equivalence testing. Using

6 Page 6 of 10 panic anxiety composite scores and setting alpha at.05, we were able to reject the hypothesis that BCT is more than 31% less effective than FCT at posttreatment. At 3- and 12-month follow-ups, the values are 25% less and 16% less effective, respectively. Logically, it was possible that the similar outcome for FCT and BCT could have been achieved because FCT was less effective in this trial than in previous trials. To investigate this possibility, we compared panic anxiety composite scores from our previous trial ( Clark et al., 1994 ) with those in the present study (see Figure 1 ). We used Group Time repeated measures ANOVAs to compare effects of treatment and maintenance of gains in the two studies. There were no differences at any assessment between FCT or BCT in the present study and FCT in the Clark et al. study. We also compared BCT and FCT in the present study with the two other treatments (applied relaxation and imipramine) in the Clark et al. study. We found that BCT and FCT were superior to applied relaxation at posttreatment and at both follow-ups ( p s <.01) and that BCT and FCT were superior to imipramine at posttreatment and at the 12-month follow-up ( p s <.05) but not at 3-month follow-up. These results exactly parallel those observed with the FCT condition in the Clark et al. study. Approximately a third of patients in the present study were on a stable dose of psychotropic medication at the start of treatment or wait list. No patients increased their medication during the trial, but 50% (5/10) of treated patients (2 FCT, 3 BCT) and 25% (1/4) of wait-list patients discontinued their medication between the pretreatment/wait-list and posttreatment/wait-list assessments. Given this pattern of results, it seems highly unlikely that medication could account for the effectiveness of FCT and BCT. As a further check on this point, we repeated the Group (FCT vs. BCT vs. wait list) Time (pretreatment/wait list vs. posttreatment/wait list) ANOVAs that assessed the effectiveness of treatment, excluding patients who were taking medication at the start of the trial. The results were the same as in the total sample. On all the measures in Table 1, FCT and BCT were superior to wait list at the posttreatment/wait-list assessment and did not differ from each other. Predictors of Treatment Response We investigated several possible predictors of treatment response using partial correlations. For these analyses, FCT and BCT groups were combined. We computed partial correlations between posttreatment panic anxiety composite scores and a range of pretreatment variables (depression, general anxiety, episode duration, treatment suitability, and expectation of improvement), controlling for pretreatment scores. Patients' ratings at the end of Session 1 of the extent to which they thought the treatment would be successful for them were negatively correlated with posttreatment panic anxiety composite scores, r (25) =.50, p <.01. No other partial correlations were significant. Predictors of Long-Term Outcome Consistent with the cognitive theory of panic, Clark et al. (1994) found that patients who had a residual tendency to misinterpret bodily sensations at the end of treatment had a worse outcome during posttreatment follow-up. To investigate whether this was also true in the present study, we computed partial correlations between end-of-treatment measures of misinterpretation of body sensations (the Brief Body Sensations Interpretation Questionnaire Panic scale [BBSIQ Panic; Clark et al., 1997 ] and the Agoraphobic Cognitions Questionnaire [ACQ; Chambless, Caputo, Bright, & Gallagher, 1984] ) and 1-year follow-up panic anxiety composite scores, holding constant posttreatment scores. For these analyses, cognitive measures were excluded from the panic anxiety composite. As in the Clark et al. (1994) study, posttreatment BBSIQ Panic scores, but not posttreatment ACQ scores, were significantly correlated with

7 Page 7 of 10 panic anxiety at 1-year follow-up: for BBSIQ Panic, r (25) =.42, p =.03, and for ACQ, r (25) =.24, ns. Discussion The present results indicate that by using between-sessions self-study modules, the therapist time required to deliver cognitive therapy for panic disorder can be substantially reduced without loss of effectiveness. BCT had a low dropout rate (0%), was superior to a wait-list control group, and did not differ from FCT at posttreatment or at follow-up. Effect sizes associated with BCT were essentially the same as those obtained with FCT, which performed as well in this trial as it did in our previous studies. Finally, BCT produced results that were superior to those we have previously obtained with applied relaxation and imipramine. A major reason for wishing to abbreviate FCT was to make the treatment available to a larger number of people. If this is to occur, it is necessary to show not only that BCT can be highly effective in a research setting with experienced therapists but also that it transports to more routine clinical settings, after suitable therapist training. Further research is required to assess transportability. In the absence of a component analysis, it is not possible to say what were the key ingredients in BCT. Patients made extensive use of the self-study modules, and it is our impression that the modules were very useful. However, it is important to remember that the brief treatment also involved seven sessions with a therapist. A considerable amount of ground was covered in these sessions, and it would be wrong to assume that similar results would be obtained using the self-study modules alone. For example, it seems unlikely that patients would be consistently willing to drop their safety behaviors in feared situations and during panic attacks if they had not had the opportunity to discuss, and often practice, this crucial but highly threatening maneuver with a therapist. In the present study, the first five sessions of BCT were spread over 3 months to allow an appropriate comparison with FCT and the wait-list condition. In normal clinical practice, closer spacing of sessions might be desirable. Once-weekly 1-hr sessions are common in psychological clinics, but there is no particular reason for supposing this type of scheduling is optimal. Salkovskis, Clark, Hackmann, Wells, and Gelder (1997) investigated the effectiveness of four sessions of cognitive therapy spread over 10 days in a group of patients with panic disorder and severe agoraphobic avoidance. The four-session treatment produced substantial reductions in panic, general anxiety, and avoidance and was more effective than an equivalent number of sessions of traditional exposure therapy. We have also recently reported a successfully treated case of panic disorder without agoraphobia in which most of the therapy was completed in a single 4-hr session after the patient had completed some of the self-study modules used in the present study ( Clark, 1996, pp ). Single-session treatment is now common for simple phobias ( Öst, 1989 ). Our case report suggests that single-session treatment may also be viable for at least some panic disorder patients. Individuals who indicated at the end of the first session that they thought the treatment was particularly likely to be successful for them showed relatively greater improvement during treatment. Safren, Heimberg, and Juster (1997) recently reported a similar finding with group cognitive behavioural treatment for social phobia. Both results suggest that for some patients treatment outcome could be improved by therapists placing greater emphasis on the assessment and enhancement of treatment credibility. Finally, the present study replicated Clark et al.'s (1994) finding that cognitive measures taken at the end of treatment were significant predictors of panic anxiety 1 year later. This suggests that cognitive behavior

8 Page 8 of 10 therapists should aim not only for symptomatic improvement but also for marked cognitive change. References American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders ((3rded., rev.). Washington, DC: Author.) Barlow, D. H. & Lehman, C. L. (1996). Advances in the psychosocial treatment of anxiety disorders: Implications for national health care. Archives of General Psychiatry, 52, Black, D. W., Wesner, R., Bowers, W. & Gabel, J. (1993). A comparison of fluvoxamine, cognitive therapy, and placebo in the treatment of panic disorder. Archives of GeneralPsychiatry, 50, Chambless, D. L., Caputo, G. C., Bright, P. & Gallagher, R. (1984). Assessment of fear in agoraphobics: The Body Sensations Questionnaire and the Agoraphobic Cognitions Questionnaire. Journal of Consulting and Clinical Psychology, 52, Chambless, D. L. & Gillis, M. M. (1993). Cognitive therapy of anxiety disorders. Journal of Consulting andclinical Psychology, 61, Clark, D. M. (1996). Panic disorder: From theory to therapy.(in P. M. Salkovskis (Ed.), Frontiers of cognitive therapy (pp ). New York: Guilford Press.) Clark, D. M., Salkovskis, P. M., Breitholz, E., Westling, B. E., Öst, L. G., Koehler, K. A., Jeavons, A. & Gelder, M. G. (1997). Misinterpretation of body sensations in panic disorder. Journal of Consulting and ClinicalPsychology, 65, Clark, D. M., Salkovskis, P. M., Hackmann, A., Middleton, H., Anastasiades, P. & Gelder, M. G. (1994). A comparison of cognitive therapy, applied relaxation and imipramine in the treatment of panic disorder. British Journal of Psychiatry, 164, Craske, M. G., Brown, T. A. & Barlow, D. H. (1991). Behavioural treatment of panic disorder: A two year follow-up study. Behavior Therapy, 22, Craske, M. G., Maidenberg, E. & Bystritsky, A. (1995). Brief cognitive behavioural versus nondirective therapy for panic disorder. Journal of Behavior Therapy and ExperimentalPsychiatry, 26, Newman, M. G., Kenardy, J., Herman, S. & Taylor, C. B. (1997). Comparison of palmtop-computerassisted brief cognitive behavioral treatment to cognitive behavioral treatment for panic disorder. Journal of Consulting and ClinicalPsychology, 65, Öst, L. G. (1989). One session treatment for phobias. Behaviour Research and Therapy, 27, 1-9. Rogers, J. L., Howard, K. I. & Vessey, J. T. (1993). Using significance tests to evaluate equivalence between two experimental groups. Psychological Bulletin, 113, Rosenthal, R. & Rosnow, R. L. (1991). Essentials of behavioural research: Methods and data analysis ((2nd ed.). NewYork: McGraw-Hill.) Safren, S. A., Heimberg, R. G. & Juster, H. R. (1997). Clients' expectancies and their relationship to pretreatment symptomatology and outcome of cognitive behavioral group treatment for social phobia. Journal of Consulting and ClinicalPsychology, 65, Salkovskis, P. M., Clark, D. M., Hackmann, A., Wells, A. & Gelder, M. G. (1997, July). Panic disorder with severe agoraphobia: Comparing exposure with a cognitive or behavioural emphasis (. Paper presented at the 25th Annual Conference of thebritish Association of Behavioural and Cognitive Psychotherapies, Canterbury, United Kingdom.) Spitzer, R. L. & Williams, J. B. (1986). Structured Clinical Interview for DSM III R. (New York: NewYork State Psychiatric Institute.)

9 Page 9 of 10 1 Post-wait-list patients who continued to meet trial inclusion criteria were randomized to FCT or BCT. Identical results were obtained when these patients were also included in the analyses. 2 Analysis of individual measures revealed a similar pattern. Table 1. Outcome Measures at Each Assessment Table 3. Percentages of Patients Panic Free and Achieving High End-State Functioning at Each Assessment Table 4. Panic Anxiety Composite Effect Sizes at Posttreatment and at 3- and 12-Month Follow-Ups

10 Page 10 of 10 Figure 1. Panic anxiety composite scores at pretreatment/wait list (Pre), posttreatment/wait list (Post), 3- month follow-up (3mFU), and 12-month follow-up (12mFU). Solid lines represent the groups (full CT and brief CT) in the present study. Broken and dotted lines represent data from Clark et al.'s (1994) study. Wait = wait list; AR ('94) = applied relaxation; Imip ('94) = imipramine; FCT ('94) = full cognitive therapy; CT = cognitive therapy.