Is the Effectiveness of CBT Fading?
A recent meta-analysis published under the provocative title “The Effects of Cognitive Behavioral Therapy as an Anti-Depressive Treatment is Falling: A Meta-Analysis” (Johnsen & Friborg, 2015) has provoked considerable discussion (and some gloating on the Internet by those who are opposed to CBT). The authors computed the effect-sizes found in 70 studies of Cognitive Behavioral Therapy (CBT) for depression published between 1977 and 2014 and concluded “the effects of CBT have declined linearly and steadily since its introduction.” What’s going on here? Is CBT becoming less effective for some reason? Do we need to abandon CBT and try other ways of treating depression?
Johnsen and Friborg (2015) propose several possible explanations for these findings. First, they suggest that the apparent simplicity of CBT has led many to assume that CBT is easy to learn and may have resulted in practitioners attempting to provide CBT without the training, practice, and competent supervision needed to provide CBT in an efficacious manner. In short, they suggest that CBT may appear less effective because some practitioners without adequate training and supervision are delivering it poorly.
Second, they suggest that practitioners with limited experience or competence may provide more effective treatment when they are following a treatment manual. They suggest that the initial developers of CBT may have been more concerned with treatment fidelity than more recent investigators and point out that many clinical trials of CBT fail to properly describe the contents of the treatment given or to specify whether a particular treatment manual was followed or not. Again the idea is that CBT may appear less effective because poor adherence to treatment protocols results in inferior treatment.
Finally, they mention the impact of the placebo effect. When a treatment is perceived as new and innovative, both practitioners and patients may approach it with considerable enthusiasm and this can maximize the placebo effect. After the “newness” wears off and a treatment approach is perceived as being routine, neither practitioners nor patients greet it with the same level of enthusiasm and a more modest placebo effect would be expected. When the effect of any treatment is measured, one is actually measuring a combination of the actual effects of the treatment, the placebo effect, and other possible biases. As the placebo effect declines over time, or methodology improves over time, the effectiveness of the treatment appears to decline. In reality, initial studies of the effectiveness of a treatment (whether a psychotherapy, a medication, or a surgical procedure) are likely to overestimate its effectiveness because they don’t account for the placebo effect and other possible biases. This suggests that over time researchers will obtain more accurate estimates of the actual effects of the treatment. Note that the problem of the apparent effectiveness of treatments declining over time is not unique to CBT, it has been observed fairly widely (for an interesting and readable discussion of this phenomenon see Jonah Lehrer’s article at http://www.newyorker.com/magazine/2010/12/13/the-truth-wears-off#). This perspective suggests that the effectiveness of CBT isn’t actually declining, rather that we’re gradually obtaining more accurate estimates of its effectiveness.
In responding to Johnsen and Friborg’s article, Beck and Waltman (2015) point out a number of methodological issues that may have influenced the results of Johnsen and Friborg’s analysis. In particular, they note that the term “CBT” encompasses a range of treatment protocols from Beck’s original Cognitive Therapy, to a variety of modifications of Cognitive Therapy, to treatment approaches such as Behavioral Marital Therapy and Behavioral Activation that are quite different from Beck’s protocol. They also note that populations treated in the studies included in this analysis range from individuals with unipolar depression to individuals with post-partum depression, depressed alcoholics, and depressed individuals with Parkinson’s Disease. Finally they note that the meta-analysis included many studies that did not clearly describe the treatment provided and that did not assess the competence with which treatment was delivered as well as some studies that provided fairly brief courses of treatment. They point out that when a meta-analysis lumps together a variety of treatments used on with variety of different populations, this makes it difficult to interpret the results.
So, is CBT for depression losing its effectiveness? No, that’s not what Johnsen and Friborg’s (2015) meta-analysis actually shows. “Effect-size” is a statistical estimate of the effect that one variable has on another, in this case the effect that CBT has on depression. Their study shows that we’re finding lower effect-sizes in recent studies than were found in earlier studies. The odds are that this is due to a number of factors. First, it is likely that effect-sizes in the earlier studies were inflated somewhat by a robust placebo effect since both investigators and patients perceived CBT as “new.” Second, early studies of CBT for depression were careful about training therapists well and taking treatment fidelity seriously. It is possible that some more recent studies were less careful about achieving and maintaining therapist competence. Third, more recent studies have tested a variety of variations on the original CBT protocols for treating depression and it is possible that some of these variations decrease the effectiveness of treatment. Finally, a number of the more recent studies have explored CBT as a treatment for depression in specific populations such as alcoholics, women with post-partum depression, and individuals with Parkinson’s Disease. It is possible that some of these populations are more difficult to treat than individuals with simple unipolar depression. My bet is that CBT isn’t losing its effectiveness, but that we’re getting more accurate estimates of its effectiveness, exploring variations on the original CBT protocols, and testing CBT’s applicability in a wider range of populations.
Is CBT still “the gold standard” in treating depression? I really dislike using that phrase in reference to CBT. Referring to CBT (or anything else) as the “gold standard” implies that it is clearly the superior option. In reality, Beck’s Cognitive Therapy (and several other forms of CBT) have been well studied and have been found to provide effective treatment for depression (and a number of other problems). This does not mean that every approach that labels itself “CBT” is effective, it does not mean that CBT has been found to be more effective than other treatment approaches, and it does not mean that other treatment approaches are necessarily ineffective. Those who have referred to CBT as “the gold standard” have been over-selling CBT and that is a mistake. Let’s be realistic about the claims we make regarding CBT. Overall, CBT has been very well-researched. The available evidence shows that adequately tested forms of CBT provide effective treatment for depression, are more effective than placebo, and are at least as effective as alternative treatment approaches with which they have been compared.
Does it make sense to lump all CBT approaches together? No, it does not. The broad category of “cognitive behavioral therapy” includes quite a variety of treatment approaches. For example, Cognitive Therapy, Rational-Emotive Therapy, Acceptance and Commitment Therapy, and Dialectical Behavior Therapy are all cognitive-behavioral therapies. While there are important commonalities among CBT approaches to treating depression, there are important differences as well. I’ve heard proponents of one CBT approach disagree quite vigorously with proponents of another CBT approach. They aren’t all the same.
Aren’t all CBT approaches to treating depression equally effective? Actually, that’s a good question. While there is research showing that a number of different CBT approaches are effective in treating depression, we don’t have an enormous amount of research comparing different CBT approaches with each other. There are some studies that compare two different CBT approaches with each other and often there is no significant difference between the two treatments. However, this doesn’t mean that all of the various CBT approaches to treating depression are equally effective. My bet is that some CBT approaches will turn out to be more effective than others. However, we’ll need a lot more research before we’ll have grounds for drawing that conclusion.
Does it really matter if you do CBT well? Of course it does, and a number of studies that support this conclusion. Unfortunately, not every therapist who thinks that they are doing CBT does it well. In particular, many practitioners focus on the cognitive interventions (such as cognitive restructuring, rational responses, and schema change) and overlook the importance of the therapeutic relationship and the importance of behavioral interventions (such as behavioral activation, exposure-based interventions, and behavioral experiments). My bet is that better training and supervision would improve the effectiveness of CBT significantly (for information about the training that we offer, go to http://www.behavioralhealthassoc.com/educationalPrograms.php).
Is CBT so wonderful that there is no need to improve the treatment of depression? No. Several CBT approaches to treating depression have been found to be effective for depressed individuals and I’m not aware of any approach to treating depression that has been found to be consistently more effective than CBT. It is really gratifying to provide CBT to depressed individuals and see how it can transform their lives. However, CBT doesn’t work equally well for everyone. There are many individuals who find that CBT is only partially effective for them and some individuals who find that CBT is not helpful to them. Depression is a major problem for depressed individuals, for those who care about them, and for society at large. We need to continue trying to improve the effectiveness of CBT, to improve the effectiveness of other approaches to treating depression, and to develop new approaches to treating depression.
Beck, A. T. and Waltman, S. (2015). Commentary on “The effects of cognitive behavioral therapy as an anti-depresive treatment is falling: A meta-analysis.” Cognitive Therapy Newsletter, 15/16, 2, p. 9.
Johnsen, T.J. and Friborg, O. (2015). The effects of cognitive behavioral therapy as an anti-depressive treatment is failing: A meta-analysis. Psychological Bulletin, advance online publication http://dx.doi.org/10.1037/bul0000015
Lehrer, J. The truth wears off. (2010). The New Yorker. Accessed at http://www.newyorker.com/magazine/2010/12/13/the-truth-wears-off#