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).
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:
- 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.
- 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.
- 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.
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:
- 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.
- 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.
- 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%.
- 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.
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.
- 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.
- 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.
- 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.
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.
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.
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.