Psychotherapy Brown Bag

Prolonged exposure therapy for PTSD: Separating fact from fiction

by Craig Bryan, Psy.D., ABPP

My previous blog regarding common misconceptions about the chronicity and treatment of PTSD trauma has garnered a great deal of discussion. In that blog, I specifically mentioned two psychotherapies that have garnered an especially large amount of scientific support for recovery from PTSD: prolonged exposure therapy and cognitive processing therapy. The explicit mention of prolonged exposure therapy, in particular, garnered a wide range of reactions, many of which conveyed there are widespread misunderstandings about this particular therapy. In the continued spirit of clarifying and countering common myths about trauma and PTSD, I want to take some time to clarify some of the most commonly-reported concerns about prolonged exposure therapy (which I will refer to as “PE” from this point forward).

 

CONCERN 1:

PE’s effectiveness is based on randomized controlled trials (RCTs) that only include “simple” cases. “Complex” cases characterized by multiple or repeated traumas, multiple diagnoses, and suicide risk are routinely excluded from these studies. The results of these RCTs therefore do not reflect “real life” trauma.

 

This criticism of RCTs is certainly true for some, but not the majority of, RCTs evaluating PE. It is a very common misconception that RCTs only take “simple” cases, but that’s not an accurate statement. It’s not possible for me to review every single PE study that has ever been published, but here are just a few that counter this concern:

 

1. Resick et al. (2002), Resick et al. (2012), Gradus et al. (2013): These three studies report various results from the same trial comparing PE to other treatments. Among those receiving PE, 84% of participants had a history of major depressive disorder, 52% had a history of alcohol dependence, and 18% had a history of panic disorder. Over 30% reported suicide ideation during the past week. 86% of the sample had experienced multiple traumatic events in their lives. These additional traumas included being sexually assaulted more than once (48%), physically assaulted (54%), or the target of attempted rape (36%), and 41% experienced child sexual abuse. On average, participants reported 6.4 lifetime traumas.

 

1. Foa et al. (2005): 67% of the sample had more than one diagnosis, the most common comorbidities being depression (41%), social anxiety (20%), phobias (20%), generalized anxiety (14%), and panic (12%). In addition, the majority of the sample had experienced multiple instances of trauma: 63% experienced adult rape, 76% experienced nonsexual assault, and 81% experienced childhood sexual abuse.

 

1. There are several RCTs currently underway in the DOD, VA, and civilian settings that are similarly marked by chronic, repeated trauma exposure and diagnostic comorbidities. Approximately 15-25% of the participants in these trials report recent suicide ideation.  

 

CONCLUSION: The majority of individuals enrolled in PE treatments studies are not “simplistic” cases, unless “simplistic” means being diagnosed with multiple disorders, being repeatedly abused and traumatized in life, and being suicidal.

 

CONCERN 2:

Very few trauma survivors treated with PE recover from PTSD, and when they do, the majority relapse within one year.

 

“Recovery” from PTSD can be defined in many different ways, but the most common and straightforward definitions of recovery often include (a) magnitude of clinical improvement and/or (b) loss of diagnosis. Loss of diagnosis is often considered to be the more stringent definition because it’s harder to demonstrate. Using that as our guide, we know the following:

 

1. Across 70 studies, Bisson et al. (2013) reported that the overall effect size of trauma-focused cognitive-behavioral therapy (TFCBT), which includes PE, as compared to no treatment or “usual” mental health treatment was -1.62. To put this number into perspective, an effect size value of -.20 is often considered to demonstrate “small” improvements, -.50 suggests “moderate” improvements, and -.80 suggests “large” improvements. An effect size of -1.62 across 70 studies therefore means that patients are doing enormously better as compared to those who receive no therapy at all and/or typical mental health treatment.

 

1. When comparing TFCBT such as PE to non-trauma-focused CBT, Bisson et al. (2013) reported that the overall effect size was -.27 and was -.48 when compared to non-cognitive behavioral therapy. TFCBTs such as PE are therefore slightly to moderately better than other forms of psychotherapy.

 

1. In an exhaustive review of PTSD treatments, the Institutes of Medicine (2008) reported the rates of loss of PTSD diagnosis following PE from 29 studies: 60%, 40%, 95%, 71%, 53%, 40%, 28%, 60%, 50%, 94%, 93%, 91%, 80%, 93%, 50%, 76%, 92%, 39%, 25%, 37%, 50%, 65%, 90%, 41%, 48%, 59%, 75%, 58%, and 25%. This yields an average recovery rate of 61%.

 

1. With respect to relapse, research suggests that very few patients in PE who recover from PTSD experience a relapse afterwards:

 

• Foa et al. (2005): only 8% showed reliable worsening during the next 12 months;
• Resick et al. (2012): only 5% experienced a relapse at any point during the 10 years following treatment (yes, that’s right: a 10-year follow-up study);
• Powers et al. (2010): patients receiving PE did better than 86% of patients receiving other therapies, and during follow-up 76% of patients receiving PE were still doing better than others. Although not a perfect indicator of relapse, we could use these numbers to infer that 11.6% experienced a “relapse” of some kind.

 

CONCLUSION: Although there is variability among different studies conducted with different populations, overall, the majority of patients with PTSD who receive PE will no longer have PTSD at the end of treatment. The overall size of this improvement tends to be somewhat larger than that seen in other treatments, and is very large compared to no treatment at all. Only a small proportion (5-10%) will experience a relapse or other clinical worsening during the next 1 to 10 years. Think about that for a moment: with less than 15 hours of therapy on average (the average duration of PE), the chances of recovering from PTSD are nearly 2 in 3, and if you do recover, the odds of staying recovered for at least 10 years is around 90%. There are few treatments in the mental health disciplines that have performance stats that good.

 

CONCERN 3:

Patients receiving PE are much more likely to drop out of treatment than patients receiving other treatments.

 

This is a common concern about PE, but let’s look at what the science tells us about dropout in PE as well as psychotherapy in general.

 

1. When looking at individual PE trials, the typical dropout rate is usually between 25-35%. For example:

 

• Resick et al. (2002): 27% dropout rate;
• Foa et al. (2005): 32% dropout rate;
• Tuerk et al. (2011): 34% dropout rate.

 

1. When looking across 70 published studies, Bisson et al. (2013) reported that the dropout rate among trauma-focused cognitive behavioral therapies (which includes PE) was 19% on average as compared to 13-15% on average in non-trauma therapies.

 

1. In a different analysis of 669 studies, Swift & Greenberg (2012) reported that the mean dropout rate in psychotherapy as a whole is around 20%. Regardless of the type of treatment received, patients with PTSD had the relative highest rates of dropout (25%). This means that patients with PTSD are among those most likely to drop out of therapy of any kind, even therapies other than PE.

 

CONCLUSION: Out of 100 patients, 20-33 drop out of PE early as compared to 13-25 who drop out of other therapies early. Overall, this suggests that there is some truth to the statement that PE has higher dropout rates, but this rate is not much higher than dropout rates among trauma survivors in general, regardless of the treatment received.

 

CONCERN 4:

Conclusions about PE are based only on treatment completers, which inflates the apparent effectiveness of the treatment.

 

This concern is based on a well-known truth about most medical treatments and psychotherapies: those who finish the treatment as prescribed generally do better than those who drop out of therapy early. As we discussed above, in real life approximately 1 in 5 patients who will drop out of treatment early. High quality clinical trials therefore often analyze the results in two ways: treatment completer and intent to treat analyses.

 

• A treatment completer analysis ignores dropouts and pays attention only to those who finish a treatment. This provides us with an understanding of how effective the treatment is if someone gets the full dose.

 

• An intent to treat analysis includes all participants who start the treatment, even if they don’t finish it. In other words, both the treatment completers and the dropouts are included. This type of analysis provides us with an understanding of how effective the treatment is with respect to the overall population of individuals accessing care. This is a more conservative estimate, but it is more generalizable to “real life.”

 

Intent to treat analyses are considered “best practice” for clinical trials, as it enables us to estimate the number of patients who will improve in “real life” if they were to simply walk in the door and start receiving a particular treatment. As applied to PE, Resick et al.’s (2002) intent to treat analysis suggested a 53% recovery rate from PTSD following PE, but the treatment completer analysis suggested an 83% recovery rate. By comparison, the recovery rate was only 2% for those who did not receive treatment. Similar findings were reported by Foa et al. (2005) and Tuerk et al. (2011), suggesting those who complete PE fare approximately 50% better than the overall group of patients who start (but don’t necessarily finish) PE. 

 

CONCLUSION: For every 100 patients who seek out treatment for PTSD, 50 will recover if they start PE (even if they don’t complete the treatment), but over 80 will recover if they finish the treatment. By comparison, only 2 will recover if they receive no treatment at all. This suggests that someone who starts PE, even if they don’t finish the treatment, is approximately 25 times more likely to recover from PTSD than someone who receives no treatment at all! Thus, although treatment completer analyses do yield higher recovery rates, it’s notable that receiving any amount of PE will likely be highly beneficial to trauma survivors.

 

CONCERN #5:

PE shouldn’t be offered to anyone because the treatment didn’t work for me/my family member/my friend.

 

This is the most difficult concern to respond to, because it highlights the limits of our knowledge: we do not yet have any treatment that is 100% effective for PTSD. Unfortunately, although PE often leads to positive outcomes, it does not work for all trauma survivors all the time. We’re starting to understand, for instance, that there are different types of trauma that may influence how a person responds to treatments like PE.

 

This limitation to our knowledge doesn’t mean that PE (or any other scientifically-supported trauma therapy) should be abandoned or dismissed, however. Consider this logic as applied to something like cancer. Chemotherapy is a useful and scientifically-supported treatment for cancer, but sometimes individuals with cancer nonetheless die of cancer despite receiving chemo. Because chemo is such an uncomfortable treatment with many very unpleasant side effects, we might conclude that it’s reasonable to withhold this “harmful” treatment from other cancer patients because it doesn’t work as well as we’d like for everyone. This would be very irresponsible, as it would effectively restrict a potentially life-saving treatment from a large number of cancer patients. We must be cautious not to make the same mistake for trauma survivors.

 

CONCLUDING THOUGHTS

 

Because the dissemination of false information about treatments like prolonged exposure therapy could potentially get in the way of trauma survivors taking their lives back, we must commit ourselves to understanding the facts about PTSD and its treatment. Too many people are tormented by the shadow of past traumas, and they deserve nothing less than accurate information about those treatments that can serve as a light in their darkness. Although PE may not work for all trauma survivors all the time, it has consistently proven to be a very good option for many individuals with PTSD.

**********

Dr. Craig J. Bryan, PsyD, ABPP, is a board-certified clinical psychologist and the Executive Director of the National Center for Veterans Studies at The University of Utah. He is a veteran of the Iraq War, and is a nationally-recognized researcher and clinician in the areas of trauma and suicide.

No, PTSD is NOT a chronic condition (at least it doesn't have to be)

by Craig Bryan, PsyD, ABPP

The recent NYTimes piece about a unit of Marines who have struggled with combat-related PTSD and suicide has reignited public conversation about military and veterans mental health, suicide prevention, and mental health treatment. During the course of these discussions, many myths and misconceptions about PTSD and its treatment have arisen. Because some of these misconceptions could potentially serve as barriers to good, effective treatment, I wanted to take some time to desconstruct and address two myths about treating PTSD among military personnel and veterans.

Myth #1: Suicide among military personnel and veterans is the result of combat trauma

First, I want to make sure to remind us all that less than half of veterans who die by suicide deployed in any capacity, and even fewer experienced combat directly. I realize that the NYTimes piece focused on a combat unit and their struggles with suicide, but if there’s one thing I’ve learned over the past decade, it’s that when we hear poignant and tragic stories such as this one, we easily forget that trauma and suicide impact even those service members and veterans who have not deployed.

Myth #2: PTSD is a chronic condition that cannot be effectively treated

Second, I would like to address what appears to be some misconceptions about trauma and PTSD. Specifically, I want to address the implication that (a) PTSD is a chronic, untreatable condition, and (b) that treatment for PTSD must be long-term by default. We now have approximately 30-40 years of research clearly demonstrating that PTSD can be effectively treated to remission, but these effects seem to be limited to only a handful of therapies. Unfortunately, most mental health clinicians don’t know how to do these therapies, and (even more stunning) very few even know that they exist.

There is now overwhelming consensus by trauma experts that certain forms of trauma-focused therapy (e.g., prolonged exposure therapy, cognitive processing therapy) are incredibly effective and lead to remission from PTSD in 2 out of 3 cases on average, and in some cases in as many as 9 out of 10 of cases (Institute of Medicine, 2008). These treatments typically last for only 10-12 outpatient sessions on average. If you were to deliver these therapies in the typical one-hour-per-week format, this means recovery can occur in less than 3 months on average. We now have anecdotal evidence, published case reports, as well as controlled scientific studies that these therapies can be administered every day instead of once per week (like usual) and you can get the same benefits (Ehlers et al., 2014). Indeed, in my own clinical work, I periodically accept military veterans for two weeks of intensive cognitive processing therapy in which we meet every morning for therapy instead of just once per week. This intensive approach leads to the loss of PTSD just as often as it does when I meet with a patient at a slower, more traditional pace of once per week. This is incredibly hopeful news for trauma survivors: it is feasible to be effectively treated for PTSD very quickly, and the effects last for up to 10 years (Resick et al., 2012). This is a very different picture from what we have historically believed about PTSD.

The good news is that we’ve learned a LOT about treating PTSD during the past 40 years, and we’re now very good at doing so. Even more encouraging is the emergence of new data from recent and ongoing PTSD treatment studies that these really good therapies also reduce suicidal ideation (Gradus et al., 2013), contrary to a very common and pervasive myth that trauma survivors cannot be safely treated with trauma-focused therapies because they are “too fragile” or “can’t handle it.” I personally find this perspective amongst clinicians to be especially heartbreaking, as it communicates to the patient that he or she truly is broken, that they really can’t handle things, and that they are indeed incapable of living lives worth living. Is it any wonder that so many trauma survivors don’t get better when this is the message they are receiving, even if only implicitly, from the very individuals who are supposed to be trained to help them recover?

So… why do so many people (mental health clinicians included) believe that PTSD is chronic? Because the vast majority of clinicians were never trained to use state-of-the-art therapies, and the vast majority of trauma survivors have never received these therapies. The “status quo” right now is to use antiquated or untested techniques based on outdated and/or pseudoscientific ideas. We are, in essence, stacking the deck in such a way that we can only confirm the very (mis)conceptions about PTSD we hope to disconfirm.

I can’t stress this enough: it is no longer true that PTSD *has* to be chronic. It is also untrue that a trauma survivor will never be able to see and experience a reality that is not colored by their tragedy and suffering.  With the right kind of therapy delivered by the right kind of clinician, recovery from PTSD can occur in a stunningly rapid period of time for the majority trauma survivors, to include military personnel and veterans. Trauma survivors no longer have to live in darkness and despair, and they don’t have to spend years of their lives and untold fortunes for “therapy” that doesn’t work. Some treatments work better than others. We need to ensure that these better treatments become the first (not the second or the third or the tenth) treatments that trauma survivors receive.

Final Thoughts

It is easy to wave the banner of “providing the best care possible for our veterans,” but I have yet to see a concerted public effort to demand that mental health professionals receive training in anything other than scientifically supported treatments. It is also easy to beat the drum for “improved access to care” and “treatment works,” but calls for improved access to care without any discussion about the quality of that free/low cost/convenient care is not helpful. Easy access to bad therapy (even if it’s free) is not good enough.

To me, the most troubling question implicated by the NYTimes piece was not “Why doesn’t the VA do more?” but rather “Why was the VA the *only* viable option available to these men?” It’s easy to point our collective fingers of blame at the system, but we should also keep in mind that what this story implies is that these men did not have access to *any* appropriately-trained mental health professional in their communities, to include non-VA options. This is a tragedy that we must fix, not only for our veterans, but for all of our family members and friends.

 **********

Dr. Craig J. Bryan, PsyD, ABPP, is a board-certified clinical psychologist and the Executive Director of the National Center for Veterans Studies at The University of Utah. He is a veteran of the Iraq War, and is a nationally-recognized researcher and clinician in the areas of trauma and suicide.

Post #2 in the Discussion of the Evidence Base for Music Therapy in the Treatment of Mental Illnesses - Reply by Music Therapy Maven founder Kimberly Sena Moore

A few days ago, I posted an initial article from a planned discussion about the evidence base for music therapy in the treatment of mental illnesses.  I received an invitation from Kimberly Sena Moore, who runs the Music Therapy Maven blog, to have a dialogue on this issue after I had a back-and-forth with some folks on the Facebook feed for Ben Folds on this issue.  Below is her reply to my initial post (click here to read my article).  As I noted before, Kimberly and I encourage readers from all perspectives to post their thoughts in the comments section.  One of the many things I appreciate about this exchance is the cordial and respectful tone and I'm hopeful readers can keep that momentum going.

************

Musings on the State of Music Therapy Research: Strengths, Challenges, and Future Directions

by Kimberly Sena Moore

 

Earlier this month I invited Dr. Michael Anestis, a clinical psychologist and researcher, to share his perspectives on music therapy research. By his own admission, his area of focus is mental health research—which omits the medical, rehabilitation, developmental, wellness, and preventive areas of focus in music therapy research—but I felt he had fair and valid points that I am sure are not unique to him. So I asked Michael to share his thoughts on the Music Therapy Maven (link), following which I would write a response.

 

I am approaching my response from a couple of different perspectives. First is as a researcher. I have spent the last several years seeped in research, music therapy and otherwise, reading, analyzing, critiquing, and having conversations about the current status of our research literature and directions it seems to be heading. The second is as a music therapy advocate. It is a perpetual challenge in our field to disseminate and share our knowledge in a way that is accessible and understood by non-music therapists. The sentiment that “oh, you’re a music therapist, isn’t that nice?” can extend to our research. I would like to think that this perception is shifting, especially as our understanding of music’s influence on brain and behavior function continues to deepen, and I feel dialogues like this can go a long way towards advocating for our profession and its research.

 

In reading Michael’s post, his primary concern seems to be that our mental health research does not meet his (and others’) definition of having a strong evidence base. These characteristics include having multiple large trials, randomly controlled trials, reporting long term effects, and clear intervention reporting with a transparent description of interventionist training and an intervention manual. Furthermore, he is wary of music therapy’s ability to treat mental illnesses (which admittedly I am to, but I also don’t feel any one type of therapy—drug, verbal, cognitive, creative or otherwise—can alone “treat” mental illnesses, but that’s another conversation...).

 

I feel we should first acknowledge and feel proud of what the music therapy profession has accomplished research-wise. Our early leaders made an intentional effort to focus on developing a research base and we have accumulated a large body of literature in our almost 65-year history. Furthermore, we have accomplished this in spite of some real challenges. One of those challenges is numbers. Although our profession shows steady growth, we are still just shy of having 6,000 board certified music therapists in the country. And how many of those are conducting research? Furthermore, we are a broad field that works across multiple ages, treatment domains, and diagnoses. Although an individual music therapist will not practice that broadly, our collective body of literature is spread thin between the mental health, medical, behavioral, developmental, educational, wellness, preventive, and neurorehabilitative fields.

 

All this is not to provide excuses, but to acknowledge that, yes, we have had challenges and there are some very real areas in which we can and need to improve our research output. We need to continue our efforts to build a stronger base of evidence. Part of this relates to the criteria Michael laid out, such as incorporating larger Ns, conducting multisite studies, and reporting long-term effects. But that is also not the only criteria through which to judge a body of evidence. Evidence is also strengthened by conducting quality research that is intentionally laid out, theory-based, and transparent.

 

Below are 5 ways that I feel that music therapy research could be strengthened and pushed to the next level. As you may glean, I don’t feel the answer lies in jumping to large, randomly controlled clinical studies. Though these are an important form of evidence, I feel we also need to work on improving the quality and depth of our research, leveraging what has been done in the past, and building on the larger literature base. These ideas are not unique to me as other music therapy researchers are writing, producing, and discussing these. They also may not be the ONLY ways to improve our output. But I do feel they warrant mention and serious consideration.

 

1) More Music Therapy-Focused Systematic Reviews

A systematic review is like a super-strength literature review. There are many different types of systematic reviews (e.g. meta-analysis, meta-synthesis, Cochrane review, narrative synthesis, and more), but they share common features, such as creating a clearly defined research question and conducting the review in a methodologically rigorous way. The intention of a systematic review is not only to review the literature on a given research question, but to synthesize the results and use that synthesize to make informed, evidence-based decisions regarding clinical treatment and future research directions.

The appeal of conducting systematic reviews in music therapy is that it allows the researcher to leverage research already completed, then combine and synthesize the results, allowing for the possibility of, among other things, the creation of new knowledge or a more clearly understood grasp of music therapy intervention efficacy. In other words, it builds new evidence based on what has been done before.

For more information about systematic reviews, the upcoming issue of the Journal of Music Therapy will include an article I co-authored defining and describing the process to conduct a systematic review (Hanson-Abromeit & Sena Moore, in press).

 

2) Conducting Translational Research

Translational research is that which takes information from basic science research and translates it for practical, clinical study or use. With music therapy, this may involve, for example, taking research from the music perception and cognition literature and using it to inform the creation and/or study of a music therapy intervention. As with systematic reviews, part of the appeal of conducting translational research is that it allows one to leverage previously existing information. Furthermore it builds knowledge from the bottom up as the foundational framework is based in scientific evidence.

Click here for more information on translational research.

 

3) Creating Support with Theoretical and Conceptual Frameworks

As I understand it, theoretical and conceptual frameworks are very closely related and there is not a single agreed-upon definition for either. Ultimately, though, each relates to using the existing literature to support a line of inquiry by identifying the research variables and relationships between them prior to conducting the study. Thus the researcher has identified what is needs to be studied or measured and, more importantly, why. In music therapy-speak, this is similar to the difference between choosing an intervention or song because you think the client will like it or it will be “fun” and choosing a particular intervention because it is the most appropriate for a given client need at that time. Developing a conceptual framework is time-intensive—from personal experience it takes a lot of reading, writing, synthesis, processing, dialogues, and a willingness to be patient with the process—but ultimately it will lead to higher quality research due to the depth involved in developing the framework.

Click here to read more about the importance of conceptual and theoretical frameworks.

 

4) Engaging in Collaborative Research

As indicated earlier, the music therapy profession has a challenge with numbers. There are almost 6,000 board certified music therapists in the state—much less than many other healthcare and education professionals, I imagine—and a small percentage are trained researchers. Plus, to state it simply, the academics in our field have the knowledge of research and research process, but the clinicians in the field are working with the clients. It makes sense, then, to combine forces and develop more collaborative clinician-academic research projects. I have started to see more collaborative music therapy research occurring and expect more will happen in future years. What’s the benefit of encouraging or promoting this type of research? In addition to the benefits of working with an interdisciplinary research team where each person brings his or her own sets of skills, experiences, and resources, conducting collaborative research allows the possibility for more research to be conducted as it opens the doors for non-academics to participate in research.

 

5) More Transparent Intervention Reporting

Music therapy is a complex profession, as are other therapies that work with human behavior. In these fields and others there is a call for more transparent intervention reporting. What does this mean? It means that researchers are being called to describe any treatment intervention in an explicit and detailed fashion. In music therapy research this means detailing not just how the music and/or music experience is facilitated (e.g. tempo, melody, lyrics, instruments, intervention steps), but also any elements of therapist effectiveness. The benefits of incorporating explicit intervention reporting is that it will accomplish not only a real opportunity to replicate studies, but also the ability to break down the music therapy treatment and analyze exactly what was contributing to client (or study participant) change.

Click here for more information about the importance of transparent music therapy intervention reporting.

Conclusion

The song “We’ve Come a Long Way Baby” comes to mind when I think about music therapy research. We HAVE come a long way in our 65-year history, but we’re also not finished. There are some real and significant ways in which we can improve the quality and depth of our research output and continue to develop a strong base of evidence. I am one of many who are having these conversations and making these changes. If you have ideas or suggestions to contribute to this conversation, please share them in the comment section below.

Resources and References

Center for Clinical and Translational Sciences (n.d.) What is Translational Research? Retrieved from http://ccts.uth.tmc.edu/what-is-translational-research" target="_self">http://ccts.uth.tmc.edu/what-is-translational-research”>http://ccts.uth.tmc.edu/what-is-translational-research

Hanson-Abromeit, D. & Sena Moore, K. (in press). The systematic review as a research process in music therapy. Journal of Music Therapy.

McGaghie, W. C., Bordage, G., & Shea, J. A. (2001). Problem statement, conceptual framework, and research question. Academic Medicine, 76(9), 923-924.

Robb, S. L. (2012). Gratitude for a complex profession: The importance of theory-based research in music therapy. Journal of Music Therapy, 49(1), 2-6.

Robb, S. L., Burns, D. S., & Carpenter, J. S. (2011). Reporting guidelines for music-based interventions. Journal of Health Psychology, 16, 342-393.

About the Author: Kimberly Sena Moore, MM, MT-BC is a board certified music therapist and a Ph.D. Candidate at the University of Missouri-Kansas City, where her clinical and research interests center on music therapy and emotion regulation development in neurodevelopmentally at-risk children. Through her role as Regulatory Affairs Associate for the Certification Board for Music Therapists, she is involved in state-level advocacy, legislative, and policy issues affecting music therapy practice. Online, Kimberly co-hosts the Music Therapy Round Table podcast, and blogs at Your Musical Self for Psychology Today and Music Therapy Maven 

The importance of testing treatments for mental illness before we sell them - Cross posting of a blog post discussion with the Music Therapy Maven blog

Greetings folks.  Below is a post I wrote that is currently posted on a blog called Music Therapy Maven (http://www.musictherapymaven.com/guest-post-the-importance-of-testing-treatments-for-mental-illness-before-we-sell-them/).  Kimberly Sena Moore, who runs the blog, will be posting a reply on her blog, which I will cross post here, and from there the discussion may continue. We both encourage folks to comment on pieces from both perspectives.  Our hope is to have a friendly discussion that encourages folks to make informed decisions about the role of music therapy in mental health care.

"The importance of testing treatments for mental illness before we sell them"

by Michael D. Anestis, Ph.D.

Last week, I had a really interesting back-and-forth on the Facebook feed for Ben Folds.  Ben had posted a link to a white paper recently released regarding music therapy and noted its “indisputable evidence on effectiveness.”  This posting caught my attention and I weighed in, which prompted some interesting responses and some understandable concerns.  Before I get into the nature and rationale of my response, let me explain a little bit about where I’m coming from on all of this.

I am primarily a suicide researcher.  I’m an Assistant Professor in the Department of Psychology at the University of Southern Mississippi and the director of the Suicide and Emotion Dysregulation lab.  When I was still a graduate student, however, I started a blog called Psychotherapy Brown Bag.  The driving force behind PBB was to take esoteric psychology studies, translate them into laymen terms, and disseminate those findings to a broad audience so that the gap between science and the general public could shrink a bit.  Much of the work on that site is devoted to describing research on treatments for mental illness and, over the years, I have developed a passion for discussing the evidence base for specific treatments.  Indeed, I now teach our annual seminar course on empirically supported treatments for adults for the clinical doctoral students at USM.  All this being said, I’m coming at this from a the perspective of somebody who cares immensely about the effective treatment of mentally ill individuals, who believes firmly in the importance of science guiding heath care, and who spends a bizarrely large amount of time reading and evaluating treatment research.

Back to the issue at hand.  When I read Ben’s post, I was concerned because, although I appreciate his passion for people to receive help, I know for a fact that the evidence base for music therapy as a treatment (stand alone or adjunctive) for any mental illnesses is limited.  Unfortunately, because most treatment research is never discussed in the media, most consumers of health care are unaware of the data supporting (or not supporting) certain treatments for certain conditions.  On top of that, most clinicians do not provide empirically supported treatments.  As such, consumers are often not equipped to know which treatments have the best support, clinicians are not selling those approaches, and the potential impact of messages from influential folks (like Ben) that deviate from the actual evidence to cause harm increases.

Here’s what I mean when I discuss whether or not a treatment has a strong evidence base:

 

• Have there been multiple large trials in which individuals diagnosed with a specific condition (using reliable diagnostic methods) who received Treatment A were compared to individuals who did not receive Treatment A?
• In those trials, did the individuals receive no treatment, “treatment as usual,” or Treatment B (a different treatment with evidence supporting its use in that context)
  • If the comparison group was a waitlist, we only know that Treatment A is “better than nothing”
  • If the comparison group was “treatment as usual,” we only know that Treatment A is better than standard care in some setting, which is almost always devoid of scientific evidence
  • If the comparison group was Treatment B (with a strong evidence base), this is more useful, as it provides evidence that Treatment A offers something incrementally useful
• Were participants randomized to treatment?  Random assignment ensures that individuals are not deliberately placed in treatments in such a way that folks getting one treatment a systematically different from one another in a way that would influence results (e.g., more severe patients in one condition versus the other).  Now, randomization doesn’t always succeed in actually distributing all relevant variables evenly, but that is something that can be tested, reported, and controlled for statistically.
• In those trials, did the people making the assessments know what treatment the participants received?  If so, that could influence results (giving better scores to people in the treatment of choice), so I want raters to be blind to treatment condition
• Did the clinicians have adequate training in the treatment?
• Did the clinicians receive supervision and were they monitored for fidelity to the treatment protocol?
• Did the clinicians use a treatment manual?
• What measures did they use to assess outcomes and did they selectively report statistically significant results while overlooking important non-significant results?
• What were the effect sizes (e.g., how BIG was the effect)?
• Did the researchers use intent-to-treat analyses to account for people who drop out of treatment or did they only consider people who completed the full trial (remember, folks doing poorly are often more likely to drop out, so not considering them can skew the findings)
• Did the trial provide long-term follow-up information?  If not, we have no idea if the effects are enduring (and indeed, some treatments begin to show differential impact relative to other treatments AFTER treatment is over)
• Did the authors report a specific mechanism or mechanisms they believe explain the impact of treatment and did they test whether changes in that variable(s) explained their results?

These are a lot of questions that might seem far removed from the emotional idea of helping somebody in need.  Believe it or not, I’ve left a lot of important ones out here.  The thing is, as mechanical and restrictive as some of these ideas might seem, they are absolutely pivotal in determining whether what we think is working is actually doing the job and how its performance compares to other treatments for whatever illness is being examined.  For newer treatments, building up an evidence base like this is difficult.  It takes time, money, training, coordination of large groups of people, and immense amounts of energy.  I can completely understand the impulse to skip those steps, follow intuition, and run with a treatment that you feel is likely to help.  The problem with that is, time and again, wonderful sounding and intuitively appealing treatments have been shown to not work or – worse yet – cause harm.

I don’t think music therapy causes direct harm.  I’m suspicious about its ability to treat mental illnesses, but my opinions (and anyone else’s) on that are meaningless.  This is an empirical question and proponents of the treatment have the same ability to test their chosen treatment as anyone else.  It is simply beholden upon them to do so.  Selling a treatment before we know if it works creates opportunity costs by directing people away from evidence-based care and towards a treatment of unknown utility.  New treatments often lag in the kind of evidence I mentioned above and, for that reason, they should also lack in promotion and implementation.  Music therapy may one day have this type of evidence base for at least some mental illnesses.  If it does, I’ll champion it as loudly (or melodically, given the nature of the treatment) as anyone else.  Until then, I simply remain cautious.

This is already getting long, but let me leave you with a few links that explain things far better than I ever could.

For a description of the evidence for particular treatments for specific mental illnesses visit: http://www.psychologicaltreatments.org/

For a description of the amazing work of Paul Meehl, who for so long clarified the relative value of data versus intuition, visit: http://www.psych.umn.edu/faculty/grove/112clinicalversusstatisticalprediction.pdf

To read a description of why science is so incredibly important in clinical psychology, read a great speech by McFall by visiting:
http://horan.asu.edu/ced522readings/mcfall/manifesto/manifest.htm

Psychodynamic versus congitive behavioral therapy in the treatment of generalized anxiety disorder and social anxiety disorder

by Michael D. Anestis, Ph.D.

 

Well now...it's been a while.  I started my job as an assitant professor in the Department of Psychology at the University of Southern Mississippi around a year ago and just a week or so ago got around to asking the legal department for the go ahead to actively post on this site while self-identifying as a faculty member.  That being said, approval was granted and here we are live with new content.  My vision for this site going forward is sporadic posts by Joye, myself, and friends/colleagues of ours with a large chunk of content coming from graduate students.  When Joye and I were grad students we gained a TON from forcing ourselves to stay current on literature and to disseminate our knowledge.  My hope is to provide a similarly valuable experience for current grad students and, in doing so, help them learn to translate their scientific accomplishments into language that consumers can identify with and use to make informed choices about their mental health care.

My plan with this first post is to pick up where we left off by discussing a topic that frequently made its way to the forefront in PBB discussions: the evidence base for psychodynamic psychotherapy in the treatment of various mental illnesses.  I chose this topic because I've come across a couple of news articles centering on one or both of the studies I'll cover today and, in these articles, the notion that the results support equivalence between psychodynamic therapy (PT) and cognitive behavioral therapy (CBT) was advanced.  Such claims always make me curious to what extent the coverage of the data (and discussion of the data by the authors themselves) actually matches the results of the study.

Both of the studies I'll be reviewing came from the research team of Falk Leichsenring, whose work we have covered previously a number of times.  At one point, we discussed controversial work he published claiming that long-term psychotherapy is superior to shorter term approaches (link; link).  Another time, Joye wrote a piece summarizing a randomized controlled trial (RCT) that he and his colleagues conducted in which PT was compared to CBT in the treatment of generalized anxiety disorder (GAD; link).  The first study we'll cover today features 12-month follow-up from that GAD study.

Study 1 - PT vs CBT in the treatment of GAD - 12-month follow-up

If you click on the link above, you'll see from Joye's original article that, in their prior publication on this topic, Leichsenring and colleagues reported that, at the end of treatment, CBT had outperformed PT in terms of worry, trait anxiety, and depression (the core features of GAD).  At 6-month follow-up, the between group difference on depression became non-significant, but CBT continued to outpeform PT in the terms of worry and trait anxiety.  In numerous forums, these results were touted as equivalence between the two treatments.  To me, however, this couldn't be further from reality.  No matter how small the effect, a significant and persistent between group difference on the primary symptoms of the disorder that brought somebody into treatment is meaningful.

Now to the newer results.  Simone Salzer, Christel Winkelbach, Frank Leweke, Eric Leibing, and Falk Leichsenring (2013) published 12-month follow-up data from this same sample (PubMed description).  The results mirrored those from the 6-month follow-up, with CBT outperforming PT on worry and trait anxiety, the core symptoms of GAD.  What this tells us is that PT is far from useless and certainly not harmful with respect to the treatment of GAD.  It also tells us that it is a suboptimal choice for an individual suffering from GAD barring circumstances in which they do not respond to CBT (which will happen for many individuals, as is the case with any treatment for any condition).

Before moving onto study 2, I want to make a quick point.  In prior conversations on topics similar to this, folks have pointed out that the focus on symptoms is problematic and that PT is better suited to treat the individual as a whole.  I have two responses to this line of thinking and it seems important to mention them now rather than waiting to see if the topics re-emerge as folks read this.  First, symptoms are remarkably important (and in the cases of symptoms like suicidal ideation and behavior, they're actually life-threatening).  Individuals whose lives are crippled by panic attacks, the inability to be in front of others without fear of harsh evaluation, depressed mood, hopelessness, affective lability, non-suicidal self-injury, and other symptoms of mental illness will no doubt express great relief if a treatment is able to help them escape these problems.  Second - and this is at least equally important - there is no evidence that PT actually does outperform CBT or any other evidence-based treatment for specific mental illnesses in the treatment of other more wholistic variables (e.g., quality of life).  If there was, that would make for an interesting conversation; however, thus far, that evidence simply is not there.

Study 2 - PT vs. CBT in the treatment of social anxiety disorder

In the second study I want to review today, Leichsenring and colleagues reported results from a large RCT comparing PT, CBT, and a waitlist control condition in the treatment of social anxiety disorder (SAD; PubMed Description).  Before discussing the results, I want to highlight a number of things that the authors did extremely well.

• Participants were randomized to treatment condition, which ensures that individuals in each condition do not systematically differ from one another and skew results spuriously to favor one treatment versus another
• The study was well-powered, with 209 individuals randomized to receive CBT, 207 randomized to receive PT, and 79 randomized to the waitlist control condition
• Treatment time was equal across conditions, ensuring that one treatment didn't outperform the other simply because individuals in that condition received more treatment
• Both treatments were manualized and treatment integrity was checked regularly and thoroughly, ensuring that therapists in each condition were acutally providing the treatment they were supposed to and doing so well
• Therapists in each condition were trained extensively in the use of the specific manual utlized in the study, ensuring that the therapists were qualified to provide the treatment at a high level.
• Assessments were conducted by individuals blind to the treatment condition of the participants, meaning that ratings of severity and improvement were not influenced by the expectations of the individual conducting the assessment.
• The authors used intent-to-treat (ITT) analyses, which include the data from individuals who drop out of treatment early.  This ensures that one treatment does not gain an unfair advantage over another because individuals who drop out due to treatment failure aren't included in the data.  On a side note, attrition was fairly equal across the three conditions, with 24% of individuals in CBT, 28% of individuals in PT, and 27% of individuals on the waitlist dropping out prior to the end of treatment.

The authors had six main outcome measures.  The first two were measures of diagnostic status.  Remission, defined as a score on the The Liebowitz Social Anxiety Scale (LSAS) of less than or equal to 30, measures what proportion of the individuals in each condition was below clinical levels on SAD symptoms by the end of treatment.  Response, defined as at least a 31% reduction in LSAS score, served as a measure of what proportion of the individuals in each condition exhibited a strong response to treatment.  The last four were self-report measures of key SAD and SAD-related symptoms: the LSAS, the Social Phobia and Anxiety Inventory (SPAI), the Beck Depression Inventory (BDI), and the Inventory of Interpersonal Problems (IIP).

So...what did they find?

Remission
CBT: 36%
PT: 26%
Waitlist: 9%

Response
CBT: 62%
PT: 52%
Waitlist: 15%

Significance tests revealed that both treatments were superior to waitlist on remission and response.  They further indicated that CBT outperformed PT on remission and that the two treatments were equal on response.

With respect to the self-report questionnaires, CBT outperformed PT on the LSAS, SPAI, and IIP and the two treatments were equal on the BDI.  In other words, in measures of social anxiety and interpersonal problems - the core features of SAD - CBT produced better results.  PT did not produce better results than CBT on any outcome.

The authors noted that the effect sizes for all between group differences were small and, indeed, this is an important fact to point out to readers.  Statistical significance does not necessarily mean that the two treatments produced results that were so dramatically different that one treatment can and should be dismissed as an option.  At the same time, consistent small effects favoring one treatment over the other on the specific symptoms that drove an individaul to seek treatment are not meaningless.

Speaking about these results, Dr. Leichsenring noted "As all differences between CBT and PDT are small, results do not justify (recommending) one of the treatments over the other." (link)

I find this quote difficult to comprehend.  Yes, small differences aren't huge differences, but small differences on the symptoms causing the most impairment and/or distress in an individual's life are an important consideration and precisely the type of guidance we should use in determine the front line treatment for individuals seeking mental health care.  If we were debating treatments for cancer or another physical ailment and were told that one produced consistently favorable results that were small in magnitude on the central components of the disorder, I suspect few would argue that the results were meaningless.  The situation is no different here.

Overall, I applaud any effort to study the effects of PT on specific mental illnesses and the authors did a lot of things tremendously well in these trials, but it is VITAL that when the media discusses results from studies (and the authors of those studies provide quotes in the articles) the words in the article match the data.  Otherwise, we're left with a game of telephone in which individuals who are motivated to provide one impression or another lead consumers further and further away from the actual evidence.

************

Dr. Mike Anestis is an assistant professor in the Department of Psychology at the University of Southern Mississippi and the director of the Suicide and Emotion Dysregulation Lab.

Comparison of CBT and ACT in the treatment of depression and anxiety: Differences emerge at long-term follow-up

by Michael D. Anestis, Ph.D.

I absolutely LOVE it when impressive researchers compare two evidence-based treatments to one another in the same study.  Comparative efficacy is a wonderful thing and we don't get nearly enough of it.  I just came across a great example of this, as Evan Forman, Jena Shaw, Elizabeth Goetter, James Herbert, Jennie Park, and Erica Yuen just published a follow-up to an earlier comparision of cognitive behavioral therapy (CBT) and acceptance and commitment therapy (ACT) in Behavior Therapy, this time reporting results from an 18-month post-treatment assessment point.

In the earlier study, Forman and colleagues (2007) reported that there were no differences between ACT and CBT immediately post-treatment in a randomized controlled trial (RCT) examining a range of outcomes including depressive symptoms, anxiety symptoms, and overall functioning.  This was compelling evidence and supportive of the notion that ACT is a promising emerging treatment capable of producing results comparable to gold standard evidence-based psychotherapy for certain populations.  As the authors noted in this study, however, it is always important to ask not only how folks are doing right after treatment, but also whether or not treatment gains are maintained over time.

The total sample included 132 undergraduates seeking treatment at a counseling center.  Importantly, participants were not required to meet diagnostic criteria for a DSM-IV-TR disorder.  Rather, inclusion criteria centered around moderate levels of depression and/or anxiety as measured by the BDI-II and BAI (scores greater than 9).  In most healthy undergradaute samples I've collected, the mean for the BDI-II comes in around 7, with a standard deviation of 7, so I'm not entirely comfortable with considering 9 a measure of moderate depression symptoms (in fact, "moderate" for the BDI-II is typically defined by a score of 20-28 and the maximum score is 63).  The authors noted that the goal was to keep inclusion criteria broad in order to maximize external validity (e.g., the degree to which the sample and the results reflect what would typically be seen in the "real world").  That's a valid goal, but it is VERY important to keep the context of the sample in mind when interpreting the findings.

As far as diagnoses go, the authors used structured diagnostic interviews and reported that 49.2% met criteria for an anxiety disorder and 37.1%  met criteria for a depressive disorder.

On to the results.  91 participants provided follow-up data and were included in the analyses.  That being said, the authors used both intent-to-treat analyses and a second idetentical set of analyses that only considered the completers.  I'll present the ITT results here.

The authors found that the group of individuals receiving CBT reported greater maintenance of treatment effects for depression and general functioning than did individuals who received ACT.  There was a non-significant trend (p = .08) for quality of life favoring the CBT group and no between group differences on anxiety.  Putting these results in another context, the authors ran analyses considering clinical significance by comparing the percentage of individuals in each group (CBT vs ACT) who were "reliably recovered" on each outcome measure (see the original paper for a thorough description of the meaning of that phrase).  The found the following:

• Depression (BDI-II): 81.8% of CBT patients vs. 60.7% of ACT patients
• Anxiety (BAI): 72.7% of CBT patients vs. 56.0% of ACT patients
• General functioning: 46.4% of CBT patients vs. 22.6% of ACT patients
• Quality of life: 37.8% of CBT patients vs. 22.9% of ACT patients

So, on the whole, the results pain a rather plainly superior long-term outcome for CBT relative to ACT with this particular population, but it is VERY important to keep our understanding of those results within their context.  This is a strong study, but it's only one study.  This is a valid sample, but not a particularly severe one.  The authors used protocols specific to diagnosis, but not everyone had a diagnosis and results for specific diagnoses were not compared to one another here.  Also, the results do not mean that ACT does not work...they simply point towards CBT working better in the long haul.  It will be interesting to see if other studies like this emerge and, if so, whether these results will be replicated. If they are, confidence in their meaning will increase significantly.  In the meantime, this is something to keep in mind.

************

Dr. Anestis is an incoming assistant professor in the Department of Psychology at the University of Southern Mississippi

Articles cited in this post:

Forman, E.M., Herbert, J.D., Moitra, E., Yeomans, P.D., & Geller, P.A. (2007).  A randomzied controlled effectiveness trial of acceptance and commitment therapy and cognitive therapy for anxiety and depression.  Behavior Modification, 31, 772-799.

Forman, E.M., Shaw, J.A., Goetter, E.M., Herbert, J.D., Park, J.A., & Yuen, E.K. (in press).  Long-term follow-up of a randomized controlled trial comparing acceptance and commitment therapy and standard cognitive behavior therapy for anxiety and depression.  Behavior Therapy.

New study considers therapeutic relationship and introject in dialectical behavior therapy and community treatment by experts

by Michael D. Anestis, Ph.D.

A study published in the current issue of the Journal of Consulting and Clinical Psychology by Jamie Bedics of California Lutheran University and David Atkins, Katherine Comtois, and Marsha Linehan of the University of Washington caught my eye recently.  The addressed some important issues, particularly the role of therapeutic alliance in treatment outcome for dialectical behavior therapy (DBT) relative to community treatment for borderline personality disorder (BPD).  Importantly, treatment outcome involved not only symptoms (non-suicidal self-injury; NSSI) but also factors related to general intrapsychic change (introject).

Before getting into the study itself, I want to take a moment to highlight a point raised by the authors early in their manuscript.  They noted that DBT has been shown to be superior to treatment as usual and treatment by expert clinicians with respect to suicide attempts, emergency/inpatient treatment, NSSI, anger, depression, and social and global adjustment (Kleim, Kroger, & Kosfelder, 2010; Lynch, Trost, Salsman, & Linehan, 2007).  Those are, without question, important variables and I have argued many times that people who trivialize the importance of symptom reduction relative to broader variables like quality of life are missing the mark.  I've also noted that DBT is, without question, my favorite form of treatment to administer as a clinician, so I have no desire to bash the treatment or its creators.  That being said, given the extensive literature on the efficacy and effeciveness for DBT, particularly with respect to patients with BPD, I can't help but wonder why we don't see data on the imapct the treatment has on the specific targets of the treatment modules (mindfulness, interpersonal effectiveness, emotion regulation, and distress tolerance) or on the other symptoms of BPD (e.g., affective lability, inconsistent sense of self, dissociation).  Either those variables aren't being measured or the results are not great.  Either way, that information should be readily available by now.

Moving on...

In this paper, the authors note that a frequent complaint against DBT is that it is soley focused on symptom change rather than more general intrapsychic change, with the idea being that if behaviors shift, that does not mean that the underlying pathology has been resolved.  To address this, they proposed to examine the impact of DBT on introject, defined as "an spect of an individual's personality that consists of self-directed actions including cognitive self-appraisals, and verbal and physical actions directed towards the self." (p.66)  Introject is not a construct with which I am particularly familiar, so my discussion of its meaning will be relatively minimal here.

The authors noted that some researchers have hypothesized that the therapeutic relationship in DBT may serve as a mechanism through which intraphychic change my arrise and that other researchers have repeatedly noted that therapeutic alliance is preditive of treatment outcome (e.g., Castonguay, Constantino, & Grosse Holtforth, 2006; click here for a PBB post contradicting the proposed impact of therapeutic alliance).  Given these points, they also sought to consider the role of therapeutic alliance across outcomes, not only in DBT, but also in patients receiving community treatment from expert clinicians.

To do this, the authors utilized data from a longitudinal randomized controlled trial in which 101 adult females were randomly assigned to receive either DBT or community treatment from expert clinicians.  All participants were diagnosed with BPD through a structured diagnostic interview.  Additionally, to be included, each participant had to endorse a past history of self-injury (at least two suicide attempts or NSSI within the past 5 years and a minimum of one NSSI incident within the past 8 weeks).  Participants with a lifetime history of schizophrenia, schizoaffective disorder, bipolar disorder, psychotic disorder not otherwise specified, or mental retardation were excluded.  Community therapists were nominated by leaders in the mental health community as particularly adept in the treatment of difficult clients and they self-identified as "eclectic" or "mostly psychodynamic."  Community treatment consisted of at least one individual session per week and additional treatment prescribed as needed.

Therapeutic relationship was assessed during the active phase of treatment at 4 months, 8 months, and 12 months.  Introject was assessed at 2 weeks, 4 months, 8 months, 12 months, and post-therapy follow-up at 16 months, 20 months, and 24 months.

The authors had 5 main hypotheses and I will describe the results relevant to each below the description of the hypothesis itself.

1. DBT patients would show a more affiliative introject during the course of treatment.

The two treatment conditions did not differ on levels of affiliative introject at the onset of treatment.

As anticipated, DBT patients saw an increase in affiliative introject across treatment and follow-up.  Importantly, the patients in DBT improved more on this outcome than did individuals receiving treatment by community experts.  Equally important, DBT patients, on average, shifted from a hostile to affiliative introject prior to the end of treatment whereas individuals in community treatment did not approach this point on average until the end of 1-year follow-up.  The authors also provide data on individual clusters within these measures and I'll refer readers to the actual study for that level of detail.

2. DBT therapists would be perceived by patients as demonstrating greater levels of affirmation, protection, and control during treatment.

This hypothesis was confirmed, with DBT participants also reporting that DBT therapists exhibited higher levels of controlling behavior early in treatment while granting greater levels of autonomy as treatment progressed.

3. Higher therapist affiliation would be associated with increased introject affiliation.

When the authors examined the impact of therapist affiliation on introject measured at the same moment  or the next assessment period in DBT, the effects were non-significant; however, when they looked at specific clusters within and across treatment, there were some significant findings.  Specifically, DBT participants who reported higher levels of therapist "active love" at one time point reported higher "self-love" at the next assessment point and, similarly, DBT participants who reported higher "therapist protect" at one time point reported higher "self-protect" at the next assessment point.  Additionally, DBT participants reported a stronger, positive association between "therapist affirm" and next period ratings of "introject self-affirm" in contrast to participants receiving community treatment who, on average, reported lower levels of "self-affirm" in assessment periods following high ratings of "therapist affirm."  This last finding speaks to the possibility that the manner in which DBT therapists express affirmation may have stronger (in fact, postive relative to problematic) effects on patients' self-affirmation. 

4. Higher therapist affiliation would be associated with less frequent NSSI.

There was no main effect of therapist affiliation on NSSI.  In other words, the therapeutic relationship itself is not the determining factor on self-injurious behavior, an important point consistent with earlier PBB conversations emphasizing the importance of specific therapeutic techniques relative to "common factors" such as the therapeutic alliance (which many people use to claim that all treatments are equal when administered by quality therapists).  There was, however, a significant affiliation by treatment interaction.  In other words, the impact of relationship on NSSI depended upon the type of treatment being received.  Specifically, DBT patients who perceived their therapists as more affiliative reported less NSSI throughout treatment and follow-up and this effect was stronger than in the community treatment condition.  Importantly, in the community treatment condition, patients who reported higher "therapist affirm" also reported HIGHER levels of NSSI.  In other words, when DBT therapists use affirmation, they appear to do so in a manner that reduces NSSI behaviors whereas when community expert clinicians use affirmation, they may actually reinforce and increase the likelihood of future NSSI behavior.  In DBT, the relationship between therapist and patient is one of the primary means through which behavior is shifted in that, when a client self-injures, he or she is not able to contact the therapist for 24 hours and the next session is spent primarily conducting chain analyses to understand why the behavior occurred and how it can be avoided in the future through alternative adaptive response patterns.  Contrastingly, when the patient calls for coaching in crisis or avoids using the behavior between sessions, substantial affirmation is given, thereby modeling effective means for building intimacy in relationships (e.g., encouraging healthy behavior and not rewarding harmful behaviors through greater levels of care and attention than are experienced in response to positive behaviors).  It is possible that community clinicians were responding to their clients' self-harming behaviors through increased expressions of caring and concern, thereby granting such attention more in response to NSSI than in response to healthy behaviors and, in effect, increasing the value of NSSI behaviors to the clients.

5. The simultaneous use of emancipating, affirming, proecting, and controlling behavior (the DBT dialectic) by the therapist would predict improved outcome.

Although there was no main effect for the DBT dialectic (e.g., simultaneous emphasis on accepting the client as she is while noting the need to change), there was an interaction effect.  Specifically, DBT patients who reported that their therapists used this approach also reported less NSSI whereas community patients who reported their therapists used this approach reported more NSSI.  Now, as the authors note, it could be that the DBT pattern does not fit the approach used by community experts and that, as such, it might represent poorly conducted treatment.  Regardless, it is certainly supportive that therapists who are perceived by their patients as behaving in a manner consistent with how DBT is intended to be administered tend to have patients who end up decreasing their use of NSSI.

***

On the whole, I think this study had a TON of strengths to it, in that it expanded the type of outcomes assessed, used an impressive study design (RCT, longitudinal approach with multiple follow-ups), and strong analytical approach (hierarchical linear modeling).  That being said, I can't help but wonder why outcomes often assessed in DBT trials (e.g., suicide attempts, re-hospitalization) were not included here.  If they did not have the data, that makes perfect sense.  If there was not enough power, that makes sense.  If they simply did not report on data they had access to and enough variability to analyze, that's a problem.  Additionally - and this is a limitation that applies to almost all clinical trials - I can't help but wish the comaprison treatment was one unified approach administered by experts in that particular approach, thereby allowing for comparisons of incremental validity using the same sample rather than relying on effects aggregated in meta-analyses in which few if any of the studies actually involved the direct comparision of treatments.

What are your thoughts?

Studies cited in this post:

Bedics, J.D., Atkins, D.C., Comtois, K.A., & Linehan, M.M. (2012).  Treatment differences in the therapeutic relationship and introject during a 2-year randomized controlled trial of dialectical behavior therapy versus nonbehavioral psychotherapy experts for borderline personality disoders.  Journal of Consulting and Clinical Psychology, 80, 66-77.

Castonguay, L.G., Constantino, M., & Grosse Holtforth, M.G. (2006).  The working alliance: Where are we and where should we go? Psychotherapy: Theory, Research, Practice, Training, 43, 271-279.

Kliem, S., Kroger, C., & Kosfelder, J. (2010).  Dialectical behavior therapy for borderline personality: A meta-analysis using mixed-effects modeling.  Journal of Consulting and Clinical Psychology, 78, 936-951.

Lynch, T.R., Trost, W.T., Salsman, N., Linehan, M.M. (2007).  Dialectical behavior therapy for borderline personality disorder.  Annual Review of Clinical Psychology, 3, 181-205.

************

Dr. Anestis is a post-doctoral fellow with the Military Suicide Research Consortium

Long-term psychodynamic psychotherapy: Discussing the evidence

by Michael D. Anestis, Ph.D.

Over the past few years, I have weighed in on a number of discussions regarding the evidence-base for long-term psychodynamic psychotherapy (LTPP) as a treatment for various forms of mental illness.  My posts on these topics have generally been in response to highly publicized publications reporting remarkably strong effects for LTPP across diagnostic categories - effects the authors typically report are stronger than those of short-term treatments in general and empirically-supported treatments (ESTs) in particular.  Importantly, I am not the only one who has expressed substantial concerns with these publications.  In fact, to those who use empirical evidence as the determining factor as to whether or not a study makes a valued contribution to science, a number of studies on this topic have been largely discredited (more on this later).  That being said, it is extremely important for these debates to take place across a number of forums so that more people - whether they are psychologists, aspiring grad students, or consmers of mental health care - can make informed decisions about what is their clearest path towards a desired outcome.

Recently, PBB guest author Jim Coyne wrote a post on his Psychology Today blog critiquing a new study by Falk Leichsenring and Sven Rabung (2011) that was recently published in the British Journal of Psychiatry and which claimed to demonstrate superior effects for LTPP relative to short-term psychotherapy.  I would encourage you to read Dr.Coyne's piece for a thorough description of the shortcomings of this particular study.  That being said, the general crux of Dr.Coyne's concerns is that the data utilized by Leichsenring and Rabung are not capable of adequately answering the relevant questions and that the data they utilize do not actually depict the narrative described in their conclusions and subsequent publicity.

Having read Dr.Coyne's piece, Dr.Jared DeFife felt compelled to write a response on his own Psychology Today blog expressing his disagreement with Dr. Coyne.  My goal today is to respond to Dr.DeFife's post, expressing my disagreement on a number of points.  As always, I want to be clear in pointing out that my goal is not to attack the writer as a person or to claim that psychodynamic psychotherapy is a failure, but rather to discuss whether or not the logic and empirical evidence being used to buffer an argument seem reasonable and consistent with stated conclusions.  As you might guess at this point, I do not think this is the case.  Before reading my critique, however, I would highly encourage you to read Dr.DeFife's blog for yourself and draw your own conclusions so that you can read my words with a critical and informed eye.  That is, afterall, the goal of science: competing hypotheses and intelligent conversations leading to clearer understandings (which, in time, will likely be overturned or expanded upon by further knowledge).

Early in Dr.DeFife's post, he wrote a section entitled "psychotherapy isn't a lab experiment."  Before discussing the three points he included under this title, it is worth noting an important retort: no medical procedures or treatments are lab experiments.  Like psychotherapy, they are interventions aimed at improving the lives of real people, many of whom are struggling with life-threatening conditions.  That being said, given that the stakes are so high, understanding the degree to which those interventions actually produce results that justify their cost and continued use in a marketplace of competing options seems vital.  Like medical procedures, things can happen during the course of psychotherapy that are difficult to capture in an experiment; however, like medical procedures, the well-being of patients is best safe-guarded through careful analyses of the degree to which psychotherapy matches the expectations of those who espouse a particular approach.  We don't question the need or possibility to study medical techniques and psychotherapy does not enjoy a special status that earns it different treatment in this regard.

Under this title, Dr.DeFife included three main points: (1) psychotherapy is not a standardized system implemented in a rigid manner, thereby making studying it more difficult (2) true clinical trials are "double-blind" and psychotherapy isn't, thereby diminishing the information we can gain from empirical trials and (3) therapy takes significant time and studying something complex that long is difficult and expensive.  Each of these points seems to me to involve important problems. 

With respect to the first point, there is, in fact, significant evidence that the use of treatment manuals (which, although serving as guidelines, do not require a therapist to administer treatment in a robotic manner lacking in empathy) improves outcomes and that the degree to which treatment is administered in a manner consistent with instructions is an important influence on the degree to which the treatment will prove effective for that client.  Paul Meehl would tell us that there is nothing wrong with improv in the lack of evidence and that, as such, when events take place that are difficult to predict (whether in psychotherapy, surgery, or any other medical procedure) it is the responsibility of the clinician to deviate from the anticipated course, but then follow-up any such deviation with empirical investigation.  In other words, yes, treatment is not always delivered in the exact same manner across clients, but that can be controlled for in any experiment and thus serve as a variable that is informative in terms of understanding results. 

With respect to the second point, I'm not entirely sure I understand Dr.DeFife's idea.  I agree that psychotherapy is not a double-blind procedure, as at least the clinician knows what treatment is being delivered; however, he notes at the end of that paragraph that just because we don't have double-blind data on seat belt efficacy does not mean we do not have meaningful data.  I would agree with that point - that imperfect psychotherapy data is highly valuable.  My concern is that this point might be being used to justify faith in all flawed data rather than considering each result within the context of its limitations and offering more confidence to results based upon data of greater quality.  In other words, if the evidence-base for one treatment is more flawed than the evidence-base for another treatment, saying all data is flawed and therefore equal seems to be a gross misrepresentation of reality.  If this is not Dr.DeFife's point, however, than obviously this is a moot point.

With respect to the third point - that studying psychotherapy is difficult and expensive due to time requirements - I agree, but my response is: so what?  These are important questions that require careful answers.  If somebody wants their treatment approach to be considered legitimate as a response to life-threatening conditions, they need to provide data capable of justifying that outcome.  Psychodynamic therapy has had just as much time (in fact much more) as other treatments to develop that evidence-base.  No treatment (including CBT for diagnoses for which it does not have strong evidence) should be used in the absence of evidence, even if that evidence is hard to attain. 

The next section of Dr. DeFife's blog post was entitled "a growing evidence base for psychodynamic therapy."  This portion was based mostly upon three citations: Leichsenring and Rabung (2008), Shedler (2010), and Gerber et al (2011).  I have not read the third piece yet, so I will hold off on making any comments (and get to work on reading the piece).  The first two citations, however, fall into the "widely discredited" category I mentioned earlier.  Rather than rehash the debates noted here and elsewhere countless other times, I'll simply direct you to these prior PBB articles detailing the many flaws of those studies in detail:

• Initial response to Shedler (2010)
• Response to media coverage of Shedler (2010)
• Response to Shedler (2011)
• Coverage of published critique of Leichsenring and Rabung (2008)

Interestingly, if you read the comments section in the second link, you'll see Dr.DeFife commented using many of the same arguments (and actually the exact same Woody Allen quote used later in his PT blog).  Ultimately, if you read these links, what you'll see is that the data simply are not consistent with the claims made by the publishing authors.  That does not mean that the opposite of what the authors say is true, but it means that it is inappropriate to look at those numbers and then publicize the notion that psychodynamic psychotherapy is outperforming ESTs, that longer-term treatment is needed in order to properly impact mental illness, or any other similar idea. 

The next section of Dr.DeFife's piece is entitled "critiques of psychodynamic therapy research: fast and furious."  Early in this section, Dr.DeFife writes:

"In critiquing or looking at critiques of meta-analyses, I'm always aware of the great opening monologue of Woody Allen's classic film Annie Hall: 'There's an old joke...two elderly women are at a Catskill mountain resort, and one of 'em says, 'Boy, the food at this place is really terrible.' The other one says, 'Yeah, I know; and such small portions.'" The critiques of meta-analyses and systematic reviews generally follow the exact same logic: Boy, the studies they review are really terrible.  Yeah, I know, and they didn't include enough of them!"

Quite frankly, I think this misses the point entirely.  Certainly some people critique the studies included in meta-analyses based purely upon the number included or the sample size utilized within the studies.  The larger complaint, however, is that the studies are of an atrocious overall quality and do not even come close to directly addressing the important questions.  Taking it even a step further, as detailed in the PBB links above, when the studies actually make an effort to compare psychodynamic therapy to ESTs for particular conditions, the results support the EST or equivalence.  When looking at secondary measures - measures included in a study that were not relevant to the central hypotheses and often are not measures of the severity of the condition being treated - results sometimes paint a different picture, but equating those measures with primary measures (based on a prior hypotheses) is questionable at best.

The next section of Dr.DeFife's post was entitled "what the studies really say."  I will once again refer to the links above for a detailed discussion of the data utlized in these studies.  The first and third links are particularly detailed on this point (the second is the least detailed in this regard).  Suffice to say that many view "what the studies really say" quite differently.

The final section of Dr. DeFife's post - "finding more worthwhile questiosn to investigate" - is the one with which I actually have the firmest disagreement.  Here again, the author expressed a very similar point in his earlier PBB comment: that we're better off investigaing questions other than which treatments work better for particular conditions that others.  First off, plenty of researchers are investigating such questions.  The two are not mutually exclusive.  Secondly, given the continued proponderance of non-evidence-based treatments in mental health and the willingness of certain researchers and media outlets to publicize false claims and/or broad claims based upon faulty data, I would argue that the need is as strong as ever to enforce a strict policy of contining to test the efficacy and effectiveness of particular treatments for particular conditions relative to alternative options.  In fact, I think the funding should go to proposed studies in which individuals who are experts in particular treatments administer those treatments (e.g., one study with experts in psychodynamic and cognitive behavioral treatments for depression) to samples randomly assigned to receive one treatment versus the other and in which hypotheses are made ahead of time regarding specific variables in which one treatment is expected to outperform the other.  Most consumers have no way of wading through a market of competing treatments and knowing where those treatments stand relative to one another, so the best answer may simply be impacting what comes to market in the first place by forcing treatments to perform to a particular level before being offered to clients presenting with particular needs.  Knowing that the treatment being received by a person in need is, in fact, the one that tends to produce the best outcomes for particular groups of people on particular measures seems to me to be as important a question as we can ask in this field.

Anyway, just to again reiterate one of my main points here: I disagree with Dr.DeFife's conclusions, but this has nothing to do with any sort of sense of who he is as a person.  I have never met him and would not hesitate to shake his hand and have a friendly and vigorous scientific debate.  The point here is simply that we disagree on the nature of and information provided by the data and we both believe it is important to make our case in a manner in which individuals can read multiple viewpoints and draw informed conclusions.

************

Dr. Anestis is a post-doctoral fellow with the Military Suicide Research Consortium

Predicting the likelihood of particular treatment outcomes for specific clients

by Michael D. Anestis, Ph.D.

We at PBB are strong advocates for the concept of empirically-support treatments (EST's).  As we have mentioned countless times before on the site, ESTs are treatments that have been shown through rigorous scientific research to produce the greatest results for the greatest number of individuals the highest percentage of the time for specific diagnoses.  For example, exposure plus response prevention (EXRP) has demonstrated the greatest results in the treatment of obsessive compulsive disorder (OCD).  Generally speaking, the strength of a treatment is demonstrated through effect sizes, which are statistical measures of the magnitude of an outcome and which allow for comparisons across studies.  In other words, rather than simply saying whether or not a study found a statistically significant result, which can depend entirely upon how large their sample was and which can not truly be comared from one study to the next, authors who report effect sizes also give us an idea of how their results stack up to those of other works.

One issue with reporting only effect sizes, however, is that they only tell us the type of results produced on average by a particular treatment, with no clarification regarding the degree to which a particular client is likely to experience that average result.  Many of the critics of ESTs argue that the treatments are not perfect and that some individuals truly benefit from alternative treatments.  Proponents of ESTs do not argue against that point - no treatment is perfect for everyone, whether their ailment is psychological or physical - but stress the point that we have no systematic way of identifying who those folks are ahead of time and that, as such, we're forced to guess and the amazing work of the late, great Paul Meehl repeatedly and decisively demonstrated that when clinicians guess, we tend to perform worse than data....meaning that although intuition may lead us to correctly identify some patients that would benefit more from an alternative treatment, it will no doubt also lead us to incorrectly identify a number of people who do not fit into that group and, as such, our net result will be worse.

All of this brings me to an interesting study currently in press in Behavior Therapy and conducted by Oliver Lindhiem, David Kolko, and Yu Cheng of the University of Pittsburgh.  In this article, the authors designed and tested what they refer to as the Probability of Treatment Benefit (PTB) chart, which calculates a simple to understand percentage that represents the likelihood that a particular individual will see a range of specific outcomes from a particular treatment.  The study is preliminary and only looks at percentages based on one factor in a sample from one study testing one treatment - but it's a fascinating first step.

Essentially, what the authors did was look at a prior effectiveness trial for a modular, primarily CBT-based treatment for disruptive behavior disorders in children.  With those data in hand, they looked at the likelihood of attaining particular outcomes given a client's pre-treatment levels of psychopathology.  The authors chose to base their predictions off pre-treatment symptom severity due to past research indicating that it is highly related to outcome (e.g., Kazdin & Whitley, 2006).  Importantly, the authors specified a number of outcomes they wanted to consider:

1. Treatment response - To what degree do clients see significant changes in their level symptoms over the course of treatment?  In other words, do they tend to exhibit substantially fewer or less severe symptoms than before treatment began?
2. Treatment outcome - To what extent do clients meet particular criteria post-treatment?  For example, do clients still meet diagnostic criteria for a particular diagnosis?  Are they in the noraml range for symptom levels?

The authors noted that past research has shown that individuals with higher levels of pre-treatment symptoms are more likely to have stronger treatment response but poorer treatment outcomes.  In other words, when you start with more symptoms, it makes sense that you would see greater change (e.g., regression to the mean, ceiling/floor effects) but it also makes sense that you would be less likely to end up with minimal symptoms (e.g., there is more room to change but more needs to change to reach a certain level; Reyno & McGrath, 2006).

I won't go into great detail regarding the methodology of this study as that extends beyond the scope of PBB and I'm a bit pressed for time, but please use the citations below if you would like to explore things to that level.  Ultimately, the authors did in fact find that individuals with higher levels of pre-treatment psychopathology improved more during treatment but had worse outcomes.  For example, individuals in the group with the highest initial severity levels had a 9% chance of being in the normal range post-treatment but a 64% chance of significant change in symptoms whereas individuals in the group with the lowest severity had a 68% chance of being in the normal range post-treatment but only a 21% chance of signficant change.

Ultimately, it is important to note that the authors only considered one variable (pre-treatment severity) as a predictor when many others contribute to outcomes (e.g., comorbid diagnoses), so future models need to take that into consideration.  Additionally, the authors were only able to consider probability of particular outcomes for one treatment of one set of diagnoses in one sample, so the results need to be reproduced in separate samples before we can be confident in how the chart would work for this particular treatment and they need to be tested for other treatments before we can even begin to consider the results in those contexts.  All of that being said, this marks an interesting first step.  The end result could be an easy to understand complement to effect sizes and a way to help clients understand what the results of the scientific literature mean for them as individuals relative to the population as a whole.

************

If you are interested in learning more about this or other topics discussed on PBB, we recommend that you consult our online store for scientifically-based psychological resources.

Dr. Mike Anestis is a post-doctoral fellow with the Military Suicide Research Consortium.

Articles cited in this post:

Kazdin, A.E., & Whitley, M.K. (2006).  Comorbidity, case complexity, and effects of evidence-based treatment for children referred for disruptive behavior.  Journal of Consulting and Clinical Psychology, 74, 455-467.

Linhiem, O., Kolko, D.J., & Cheng, Y. (in press).  Predicting psychotherapy benefit: A probabilistic and individualized approach.  Behavior Therapy.

Reyno, S.M., & McGrath, P.J. (2006).  Predictors of parent training efficacy for children externalizing behavior problems - a meta-analytic review. Journal of Child Psychology and Psychaitry, 47, 99-111.

The impact of obsessive-compulsive personality disorder on treatment outcomes in exposure and response prevention (EXRP)

by Michael D. Anestis, M.S.

Just about any clinician can tell you that the presence of a personality disorder can derail even treatments with monumental amounts of empirical support.  Today, I want to provide some evidence to support that notion through a quick recap of a study in press at Behaviour Research and Therapy, conducted by Anthony Pinto, Michael Liebowitz, Edna Foa, and Blair SImpson.  In this study, the authors examined the degree to which the presence of obsessive compulsive personality disorder (OCPD) predicted treatment outcomes for individuals receiving exposure and response prevention (EXRP) for obsessive compulsive disorder (OCD).

Before explaining the study, let me clarify a few points:

1. EXRP is, by far, the treatment with the greatest level of empirical support in the treatment of OCD.  Click here to read all articles we've written about this treatment.
2. OCPD and OCD, while similarly named and certainly related to one another, are distinct conditions.  Yes, it's frustrating that their names are so similar, but what can you do?

We haven't spent a ton of time discussing OCPD, so I want to give a quick recap of the DSM-IV-TR criteria for the disorder.  Like all PDs, it must be a long-standing pattern that leads to functional impairment and/or distress across multiple domains of life.  Specifically, the disorder is primarily based upon a preoccupation with orderliness, perfectionism, and cognitive inflexibilty and a diagnosis requires endorsement of four or more of the following symptoms:

• Perfectionism/preoccupation with details, lists, order to the extent that the point of the activity is lost
• Excessive devotion to work at the expense of functional relationships and liesure
• Scrupulocity and inflexibility about morality and ethics
• Inability to discard worn-out or worthless items (hoarding)
• Refusal to delegate tasks to others
• Miserly
• Rigidity and stubborness

In this particular study, the authors used data from a randomized controlled trial in which adults with OCD currently prescribed a therapeutic dose of a seletive serotonin reuptake inhibitor (SSRI) were assigned to also receive either EXRP Or stress management training.  That original study (Simpson et al., 2008) found that only EXRP was an effective augmentation for the SSRI, but that's not the focus here.

Of the 54 patients assigned to receive EXRP, 49 had been assessed for personality disorders using a structured diagnostic interview and these individuals were included in the analyses in this study.  Additionally, because research has demonstrated that the hoarding symptoms of OCPD does not adhere well to the other symptoms (Nestadt et al., 2006) and will likely be dropped in the DSM-5 description of OCPD, the authors ran their analyses both with and without this symptom included (this did not impact results either way).

The authors found that, at baseline, individuals with and without OCPD did not differ on OCD severity as measured by the Yale-Brown Obsessive Compulsive Scale (Y-BOCS) and OCPD symptoms were not significant correlated with Y-BOCS scores.  In other words, any difference in treatment outcomes associated with OCPD could not simply be accounted for by the fact that those folks had more severe OCD symptoms.

Post-treatment, the authors found that both OCPD diagnosis (yes/no does the patient meet criteria for OCPD?) and OCPD symptoms (how many symptoms of OCPD does the patient endorse?) predicted worse EXRP outcomes.  Importantly, the authors also examined whether all OCPD symptoms had a similar impact on treatment outcome. As it turns out, only one - perfectionism - predicted poorer response to treatment.

Perfectionism is not a bad thing in and of itself, but when it reaches extreme levels, it becomes highly predictive of a number of highly problematic outcomes.  This study is simply one more example of that well-established point.

So, ultimately what was the point the authors were trying to make here?  Even when applying the gold standard treatment for OCD, clinicians need to be careful to assess for OCPD, and particularly for perfectionism, as this can predict a poorer treatment response.  Although future studies would need to be conducted indicating that treatment of perfectionism results in better EXRP outcomes, this study offers an initial level of support for that notion.

************

If you would like to learn more about this or other topics discussed on PBB, we recommend consulting our online store for scientifically-based psychological resources.

Mike Anestis is a psychology resident at the University of Mississippi Medical Center and a doctoral candidate in the clinical psychology department at Florida State University.

Articles referenced in text:

Nestadt, G., Hsu, F.C., Samuels, J., Bienvenu, O.J., Reti, I., Costa, P.T. Jr., et al. (2006).  Latent structure of the diagnostic and statistical manual of mental disorders, fourth edition personality disorder criteria.  Comprehensive Psychiatry, 47, 54-62.

Pinto, A., Liebowitz, M.R., Foa, E.B., & Simpson, H.B. (in press).  Obsessive compulsive personality disorder as a predictor of exposure and ritual prevention outome for obsessive compulsive disorder.  Behaviour Research and Therapy.

Simpson, H.B., Foa, E.B., Liebowitz, M.R., Ledley, D.R., Huppert, J.D., Cahil, S., et al., (2008).  A randomized, controlled trial of cognitive-behavioral therapy for augmenting pharmacotherapy in obsessive-compulsive disorder.  American Journal of Psychiatry, 165, 621-630.