by Michael D. Anestis, M.S.
Over the past couple weeks, as the debate over the utility of empirically supported treatments (ESTs) has played out across a number of forums, one of the critiques of the EST approach that I have encountered the most often is the idea that identifying particular forms of treatment are more effective for particular diagnoses denies clients the ability to choose their own path. The general basis for this argument is that not everyone will respond to an EST and, as such, clinicians need the flexibility to provide alternative approaches (other arguments, such as the Dodo Bird hypothesis have been addressed at length on PBB as well). That point, however, is not lost on EST advocates, who merely believe that, unless we have a systematic way of identifying ahead of time who will and will not respond to the most well supported treatment, it makes no sense to guess. Instead, the EST (or EST's, if more than one is available) should be used first and, if assessments indicate a lack of progress and a referral to another therapist is not possible, alternative, unsupported approaches are a reasonable path. Still, the idea of a hierarchical list of effective treatments bothers some as they believe it forces clients to receive a particular form of care that may run counter to their desires and, as such, the client might be more likely to drop out of treatment and less likely to respond positively to the protocol.
What's interesting about that idea, however, is that it is based purely upon guesswork. It sounds logical and, in fact, in my own training as a therapist, I was always told to provide choice when multiple forms of therapy have been shown to result in equivalent outcomes for a particular presentation, but the idea that not having a choice in treatment will have detrimental effects on outcome is not based upon any form of evidence. Today, I would like to discuss a study by Yan Leykin and colleagues, published in 2007 in Behavior Therapy. In this study, the authors attempted to fill this knowledge gap with an investigation of whether receiving one's preferred treatment in a randomized controlled trial of antidepressants versus cognitive therapy for depression would result in better outcomes and fewer drop outs.
In treatment research, whether we are examining psychological treatments or medical treatments, the gold standard approach is considered to be the randomized controlled trial (RCT). RCT's randomly assign participants to a condition of the study (e.g., antidepressant medication versus cognitive therapy) in order to ensure that the results of each form of therapy are accounted for by the therapy itself rather than by the fact that particular people more or less likely to respond chose to receive that treatment. In this sense, RCT's maximize internal validity. This all makes very good sense from a scientific perspective, but many people understandably argue that this is problematic because, in "real world" clinical practice, clients want to choose their therapy and are not randomized at all. This, of course, brings us back to the original question discussed in today's article: does receiving the treatment we prefer impact our response to treatment and the likelihood that we will drop out?
Leykin and colleagues (2007) examined this question within the context of a larger RCT that ultimately found equivalent responses to antidepressant medications and cognitive therapy for moderate to severely depressed individuals, with both treatments outperforming placebo. All of the participants in the study met DSM-IV criteria for depression as determined through a structured diagnostic interview and had scores of 20 or higher on the Hamilton Rating Scale for Depression (Hamilton, 1960) at two assessments at least one week apart (ensuring that they were not simply having a bad day when tested). More than 70% of the participants met criteria for at least one other Axis I diagnosis (e.g., anxiety disorders) and nearly 50% of the sample met criteria for at least one personality disorder. As such, the participants in this study were not unlike clients seen in standard "real world" practice environments either in severity of impairment or number of diagnoses.
Before being randomized to their treatment condition, participants were asked to indicate their preference (if any) for antidepressant medication versus cognitive therapy. Only participants who indicated a preference were included in these analyses, resulting in 68 participants who preferred antidepressants and 41 who preferred cognitive therapy (total n = 109). Importantly, however, participant preference for a particular treatment had no bearing on whether or not that individual received a particular treatment.
The results of this study are likely to surprise a lot of readers. As it turns out, individuals who received the treatment they preferred did not respond better than did individuals who did not receive the treatment they preferred. Additionally, individuals who received the treatment they preferred were no less likely to drop out of treatment than were individuals who did not receive their preferred treatment. In other words, even with two treatments as different from one another as antidepressant medication and cognitive therapy, a client's preference for a particular treatment may not have a substantial impact on whether he or she responds to and remains in treatment.
As the authors noted, there were a couple important limitations to consider in this study before jumping to any firm conclusions about its meaning. First of all, there was a partial failure of randomization in that clients were more likely to receive their preferred treatment even though they were assigned at random. Secondly, the results only reflect individuals who agreed to be randomized in the first place. In other words, these results can not tell us whether receiving the preferred treatment would be more meaningful in terms of results for individuals who feel strongly enough that they are not willing to participate in an RCT and run the risk of receiving a treatment other than their preferred choice. Future research that includes both randomized and non-randomized participants could potentially address that issue; however, such a design would have inherent methodological and theoretical difficulties that would render results difficult to interpret.
In this study, although the treatments were quite different from one another, both options had strong empirical support as effective treatments for depression. As such, even when clients received the treatment that was not their preferred choice, they were certain to be provided treatment with a strong scientific foundation. But what about other scenarios? Unfortunately, in "real world" practice, clients are often given the option of receiving treatments that lack any empirical support indicating that they are effective in resolving the problems that brought them in to treatment. An effective sales pitch, exposure to misinformation, or any other number of variables can lead the client to prefer the unsupported treatment and, if we align ourselves with the belief that receiving the preferred treatment will impact a client's outcome, suddenly individuals with mental illness become highly vulnerable to receiving untested treatments for illogical reasons. Although the Leykin et al (2007) study is by no means conclusive regarding the impact of treatment preference, the data are strong and indicate that, although intuitively appealing, the idea that treatment preference highly impacts outcome is simply incorrect. When two or more empirically supported treatments are available for a given disorder (e.g., antidepressants, cognitive-behavioral therapy, and interpersonal psychotherapy for depression), it makes sense to offer choice and, in doing so, perhaps strengthen the therapeutic alliance, but when only one choice exists with scientific support, there is no reason to believe that offering the option of unsupported treatments for those who prefer them will result in a better response. What matters is receiving the proper treatment, not receiving the treatment that sounds the most appealing.
If you would like to learn more about empirically supported treatments for mental illness, we recommend the following items, all of which are available through our online store of scientifically-based psychological resources:
- A Guide to Treatments that Work
by Peter Nathan and Jack Gorman
- Clinical Handbook of Psychological Disorders, Fourth Edition: A Step-by-Step Treatment Manual
by David Barlow
- Science and Pseudoscience in Clinical Psychology
by Scott Lilienfeld, Steven Lynn, and Jeffrey Lohr
- Cognitive Therapy Techniques: A Practitioner's Guide
by Robert Leahy
Mike Anestis is a doctoral candidate in the clinical psychology department at Florida State University
Subscribe in a reader
