Meta-Analysis Shows the Effectiveness of Cognitive Therapy
How effective is cognitive therapy, for which disorders, and compared to what? A. Butler and J. Beck (2000) reviewed 14 meta-analyses that have investigated the efficacy of cognitive therapy
with a total of 9,138 subjects in 325 studies involving 465 specific
comparisons. Meta-analysis is a statistical approach that allows researchers to aggregate the results of multiple studies and describe
these results in a standard unit known as an effect size. In our review,
we examined how cognitive therapy outcomes compared to the
outcomes of various control groups in terms of their effect sizes. Below is a summary of
our review. For an email attachment of the full manuscript, contact
beckinst@gim.net, attention Assistant Intake Coordinator.
What is an effect size?
Effect sizes of the type we report here represent the difference between two group
outcomes in standard deviation units. Hence, an effect size of 0 would indicate no
difference between group outcomes, whereas an effect size of 0.50 would indicate that
the outcome of the cognitive therapy group was one-half a standard deviation larger
(better) than the outcome of the control group. An effect size of 1.0 (a one-standard-deviation difference) is large and would indicate that the average cognitive therapy
patient had an outcome superior to that of 84% of the control group. Effect sizes have
been categorized along a continuum of no effect (0 to 0.2), low effect (0.2 to 0.5),
medium effect (0.5 to 0.8) and high effect (greater than 0.8).
| SUMMARY OF META-ANALYTIC FINDINGS |
|
Comparisons of cognitive therapy to no-treatment, wait list, and placebo controls. |
|
|
Disorder |
Average effect size |
% of CT patients superior to controls |
|
| Adult unipolar depression |
.82 |
79% |
| Adolescent unipolar depression |
1.11 |
87% |
| Generalized anxiety disorder |
1.04 |
85% |
| Panic disorder with or without agoraphobia |
.91 |
82% |
Social phobia |
.93 |
82% |
| Childhood depression and anxiety disorders |
.90 |
82% |
| Marital distress |
.71 |
76% |
| Anger |
.70 |
76% |
| Childhood somatic disorders |
.47 |
68% |
| Chronic pain (not headache) |
.46 |
68% |
|
Comparisons of cognitive therapy with alternative treatments
Cognitive therapy versus antidepressant medication:
Cognitive therapy was somewhat superior to antidepressant medications in the treatment of adult unipolar depression (average effect size = 0.38).
Importantly, one year after treatment discontinuation, depressed patients who had been treated with cognitive therapy had half the relapse rate of depressed patients who had been treated with antidepressant medication (30% versus 60%).
Cognitive therapy versus supportive/nondirective therapies
In the small number of direct study comparisons - two for adolescent depression and two for generalized anxiety disorder - cognitive therapy was moderately superior (average effect size = 0.77).
Cognitive therapy versus behavior therapy
Cognitive therapy was equally effective as behavior therapy in the treatment of adult depression and obsessive-compulsive disorder (effect sizes = 0.05 and 0.19, respectively).
Other comparisons:
Cognitive therapy was somewhat superior to a group of miscellaneous psychosocial treatments for sexual offending (effect size = 0.35). There is no treatment (including hormonal therapy) that is superior to cognitive therapy for reducing recidivism in this population.
Large effect sizes have been found for pre-to-posttreatment improvement of bulimia nervosa symptoms using cognitive therapy.
A look toward the future
There has been a trend since the 1970's to apply cognitive therapy to an increasingly wider spectrum of disorders. Substance abuse, post-traumatic stress disorder, bipolar disorder, personality disorders, anorexia nervosa and schizophrenia are among the disorders receiving recent empirical attention.
Adapted with permission from: Butler, A. C., & Beck, J. S. (2000). Cognitive therapy outcomes: A review of meta-analyses. Journal of the Norwegian Psychological Association, 37, 1-9.
Do Stress Reactivity and Coping Style Predict Response to Cognitive Therapy of Depression?
by Andrew C. Butler, Ph.D.
Research Coordinator
Cognitive therapy is well-established as an effective treatment for depression. However, not all depressed patients respond equally well and relapse is common among some patients. Surprisingly little is known about the patient characteristics that predict a less positive response to cognitive therapy for depression. Research being conducted at the Beck Institute is helping to fill this gap in the literature.
Over the past couple of years, the Beck Institute has teamed up with Larry Cohen from the University of Delaware and Kathleen Gunthert from American University to examine the potential role of stress reactivity and coping style as predictors of response to cognitive therapy for depression. Based on prior research, we hypothesized that stress reactivity – the degree to which stress activates negative cognitions and affects – will be associated with poorer outcome. We also predicted that coping style – one’s preferred strategies for coping once distress has been activated – will be related to outcome, although we’ve made no predictions about the specific coping strategies that are likely to be most helpful.
Last year, as a first step in our investigation, we conducted a pilot study with 46 depressed patients treated at the Beck Institute (Gunthert, Cohen, Butler, & J. Beck, 2003). During their first week of therapy patients completed questionnaires each day on the day’s most stressful event, depressive cognitions in response to that event, strategies used to cope with the event and/or their distress, and their level of negative affect at the end of the day. Using this daily data we calculated a stress reactivity score for each patient. This score reflected the strength of association between a patient’s stressful event appraisals and his or her negative affect ratings. A high score indicated that a patient had large increases in negative affect relative to his or her own stress appraisals about events.
Change in depression was assessed with the Beck Depression Inventory-II which patients completed before each therapy session. We used growth modeling statistics to estimate the trajectory of each patient’s BDI-II scores across treatment. The average patient showed a reduction of 15 BDI-II points across 12 sessions; the mean BDI-II score at intake was 22 and it dropped to 7 by session 12. The trajectory for the average patient showed a relatively rapid reduction in depression early in therapy followed by a more gradual reduction as therapy progressed.
The next step in our analyses was to estimate the effects of daily stress and coping variables on BDI-II trajectories. Using hierarchical linear modeling we examined these effects while controlling for initial level of depression. Thus, we tested whether patient differences in initial stress reactivity and coping style predict change in depression assuming that each patient starts therapy with the same level of depression. Our findings showed that stress reactivity did predict treatment response. Patients who were more affectively reactive to daily stressors during the first week of treatment had somewhat flatter BDI-II trajectories, meaning they had less overall reduction in depression over 12 sessions. They also tended to show a less steep reduction in BDI-II scores during the first few sessions compared to patients who were less reactive.
Which coping strategies were related to outcome? We found two strategies that predicted better outcomes and two that predicted poorer outcomes. The two strategies associated with more rapid improvement on the BDI-II were direct action (active problem solving) and relaxation. The strategies associated with less rapid improvement were self-blame and social support seeking (perhaps due to excessive reassurance seeking).
The findings from our pilot study encouraged us to design a larger study which will be conducted over the next couple of years. In this study we plan to replicate the basic design of the first study while increasing the sample size to 60 patients, adding additional measures, and having patients telephone in their responses to the daily stress and coping measures using an interactive voice response system. In addition, we will have patients complete the daily stress and coping measures a second time during the week following their sixth therapy session. This will allow us to study whether cognitive therapy reduces patients’ stress reactivity and increases their use of certain coping strategies. We also plan to follow patients after treatment has ended to study whether stress and coping variables predict how well treatment gains are maintained.
We are excited about this research both for the substantive contribution it can make regarding patient predictors of response to cognitive therapy, and because of the innovative research methodologies we are testing. If successful, this research will be unique in demonstrating the utility of daily process designs in the study of patient predictors of psychotherapy outcome.
References
Gunthert, K. C., Cohen, L. H., Butler, A. C., & Beck, J. S. (2003). Predictive Role of Daily Coping and Affective Reactivity in Cognitive Therapy Outcome. Manuscript submitted for publication.
Copyright © 2003 Beck Institute
