The National Academies of Sciences, Engineering and Medicine
Office of Congressional and government Affairs
At A Glance
: Treatment of Posttraumatic Stress Disorder
: 04/01/2008
Session: 110th Congress (Second Session)
: David Matchar

Director and Professor of Medicine, Center for Clinical Health Policy Research, Duke University Medical Center, and Member, Committee on Treatment of Posttraumatic Stress Disorder, Board on Population Health and Public Health Practice, Institute of Medicine, The National Academies

: House
: Veterans’ Affairs Committee, Subcommittee on Health


Statement of

David Matchar, MD
Director and Professor of Medicine
Center for Clinical Health Policy Research
Duke University Medical Center
Member, Committee on Treatment of Posttraumatic Stress Disorder
Board on Population Health and Public Health Practice
Institute of Medicine
The National Academies

before the

Subcommittee on Health
Committee on Veterans’ Affairs
U.S. House of Representatives

April 1, 2008

Good morning, Mr. Chairman and members of the Committee. My name is David Matchar. I am Director and Professor of Medicine at the Center for Clinical Health Policy Research at Duke University Medical Center and served as a member of the Institute of Medicine committee which produced the report Treatment of Posttraumatic Stress Disorder: An Assessment of the Evidence.1 The Institute of Medicine was chartered in 1970 as a component of the National Academy of Sciences. This study was sponsored by the Department of Veterans’ Affairs as part of an ongoing series of reports on the health of veterans.

The Department of Veterans’ Affairs charged the Institute of Medicine committee with several specific tasks. We were asked to: (1) review the evidence and make conclusions regarding the efficacy of available treatment modalities; (2) note restrictions of the conclusions to certain populations; (3) answer questions related to treatment goals, timing and length; (4) note areas where evidence is limited by insufficient research attention or poorly conducted studies; and (5) comment on gaps and future research.

To respond to its first task, making conclusions regarding efficacy, the committee developed methods using generally accepted international standards for conducting a systematic qualitative review. This included developing key questions, specifying the literature search strategy, inclusion and exclusion criteria, key quality criteria (such as assessor blinding or independence, and treatment of missing data), and judging the weight of the body of evidence. The committee’s conclusions were ultimately based on its judgments of the sufficiency of the body of evidence for each category or class of treatment. Here, I should make an important distinction between what the committee did, which was to evaluate the evidence, and clinical practice guidelines. The committee was not asked to recommend what therapies clinicians should use or not use. Making such recommendations is the work of professional associations (such as the American Psychiatric Association) and guidelines are also developed by government agencies such as the VA. Clinical practice guidelines have different purposes and frequently include a very broad range of considerations.

The committee focused its review on randomized controlled trials (RCTs) because their design is most bias resistant to answer questions of efficacy, and because the statement of task asked that we review the highest level of evidence available, which was RCTs in most cases. Application of the committee’s inclusion criteria (such as, studies that were published in English, were based on Diagnostic and Statistical Manual criteria, and included a PTSD outcome measure) narrowed the list of nearly 2,800 articles down to 89 RCTs, 37 studies of treatment with medications, and 52 studies of treatment with psychotherapy. Among the medication studies, the committee found studies of drugs such as selective serotonin reuptake inhibitors (SSRIs) and anticonvulsants.

The evidence on pharmacotherapy in general was limited, with relatively few studies meeting inclusion criteria and free of significant methodological limitations. Even among the SSRIs, with the most substantial evidence base, the committee was struck by inconsistencies in the results of studies, and serious methodologic limitations. The committee found the evidence for SSRIs (and all other drug classes for which RCTs were identified) inadequate to conclude efficacy. The report provides comments on several of the drug classes indicating areas where evidence might be suggestive in important subgroups.

The committee grouped the psychotherapy studies empirically into categories as actually examined in the literature, and did not attempt to enter the debates in the field about how the various therapies may be related at the level of theory. Among the psychotherapies, the committee identified studies where the therapy being investigated was exposure therapies alone or in combination with another component, cognitive restructuring, one or more types of coping skills training, Eye Movement Desensitization and Reprocessing (EMDR), other psychotherapy, and group format therapy. (The term exposure therapies refers to a family of therapies that include confronting the trauma-related memories or stimuli.)

The committee judged the evidence for exposure therapy sufficient to conclude efficacy. The evidence for all but one of the remaining psychotherapy categories (including the broad “group therapy” category) was inadequate to conclude efficacy. The evidence on other psychotherapies, such as hypnosis and brief eclectic psychotherapy was so limited that the committee did not form conclusions at all.

The committee’s conclusions of inadequacy regarding evidence for most treatment modalities should not be misinterpreted as if they are clinical practice guidelines. Finding that the evidence is inadequate is not a determination that the treatment does not work. It is an honorable conclusion of scientific neutrality. The committee recognizes that clinical treatment decisions must be made every day based on many other factors and considerations, such as patient preference, availability, ethical issues, and clinical experience, that we were not asked to address, and we did not.

Next, the committee considered the issue of whether conclusions may be drawn about treatment efficacy in regard to population, provider, or setting. The committee was struck by the lack of evidence on this important issue. The Diagnostic and Statistical Manual criteria do not recognize more than one type of PTSD (such PTSD distinguished by trauma type), yet reasonable people might question whether all PTSD is the same and whether one can expect a treatment shown effective in one group, for example earthquake survivors, to also work for U.S. combat veterans. Rigorously speaking, a study only applies to the population actually studied unless there are data showing the data applies to other groups. We found no evidence either that PTSD is the same or that it’s different in veteran or VA populations compared with civilian populations. A minority opinion in the report was based on the belief that there are subgroups and the evidence should be examined separately for them, but the committee majority concluded otherwise.

VA asked the committee to comment on what the literature tells us about the meaning of recovery, the effect of early intervention, and the impact of treatment length (e.g., brief vs. prolonged therapy). The committee found no generally accepted and used definition of recovery in PTSD. We recommend that clinicians and researchers work toward common outcome measure that are valid in research, allow comparability between studies, and are useful to clinicians.

We interpreted early intervention to mean keeping cases of PTSD from becoming chronic. Intervention before the diagnosis of PTSD or before the possibility of meeting the definition of PTSD (generally, early intervention in the literature occurs immediately post-trauma, referring to a condition that’s a precursor to PTSD, such as Acute Stress Disorder) was not part of our scope, because it refers to people who do not yet have or may never develop PTSD. We could not reach a conclusion on the value of early intervention, and recommended that further research specify time since trauma and duration of PTSD diagnosis. Interventions should be tested for efficacy at clinically meaningful intervals.

On length of treatment the committee found that the research varied widely in length of treatment even for a single modality, and was not able to reach a general conclusion. We recommend that trials focus on optimal length of given treatments, and that trials of comparative effectiveness between treatments should follow. There is also a need for longer-term follow-up studies after treatment concludes.

Our last two tasks were to address areas inadequately studied, and recommendations for further research. Our overall message here is that PTSD needs more attention from high-quality research, including in veterans. The committee highlighted several research-related issues in the report, including internal validity (for example, was there blinding in the study, was there adequate follow-up of patients, were missing data handled with appropriate analyses?), investigator independence, and special populations.

As outlined in our methods and in a technical appendix, the committee found much of the research on PTSD to have major limitations when judged against contemporary standards in conducting randomized controlled trials. While recognizing that PTSD research perhaps presents special challenges, we know that high quality studies are possible because we found them in our search, and there are authorities in the field of PTSD research who have called for more attention to methodologic quality. We recommend that funders of PTSD research take steps to insure that investigators use methods to improve the internal validity of research.

The committee also noted that the majority of drug studies have been funded by the pharmaceutical manufacturers, and the majority of psychotherapy studies have been conducted by the individuals who developed the techniques or their close collaborators. The committee recommends that a broad range of investigators be supported to conduct replication and confirmation studies.

The committee recognized that PTSD is usually associated with other problems such as comorbid substance abuse, depression, and other anxiety disorders. More recently, there’s been growing concern about people with PTSD and traumatic brain injury. The research literature is not informative on this issue of patients who have PTSD and other disorders. It also does not address PTSD in special veteran populations such as ethnic and cultural minorities, women, and people with physical impairments. We recommend that the most important such subpopulations be defined to design research around interventions tailored to their special needs.

Finally, the committee made two general recommendations about research in veterans. First, the committee found that research on veterans with PTSD is inadequate to answer questions about interventions, settings, and length of treatment. We recommend that Congress require and insure that resources are available to fund quality research on the treatment of veterans with PTSD, with involvement of all relevant stakeholders. Second, the committee found that the available research is not focused on actual practice. We recommend that the VA take an active leadership role in identifying the high impact studies that will most efficiently provide clinically useful information.

In closing, I would like to highlight the three key messages of this report.

  1. Many of the studies that have looked into the effectiveness of PTSD therapies have methodological flaws and therefore do not provide a clear picture of what works and what does not work.
  2. Various pharmaceuticals and psychotherapies may or may not be effective in helping patients with PTSD; we simply do not know in the absence of good data in most cases. To strengthen study quality, we need: larger studies, longer and more complete follow-up of all participants (including those who discontinue treatment before the study is over), and better selection of which treatments to study and which to compare to each other, with priority given to the most widely used therapies. Also, greater focus on veteran populations and special subpopulations (e.g. those with traumatic brain injury, substance abuse).
  3. Given the growing number of veterans with PTSD and the seriousness of this disorder, the VA, Congress, and the research community urgently need to take steps to ensure that the right studies are undertaken to yield scientifically valid and generally applicable data that would help clinicians most effectively treat PTSD sufferers.

The committee is grateful to have had the opportunity to be of assistance to VA, and hopes that the department and Congress find the report useful in moving ahead to strengthen PTSD research.

Thank you for the opportunity to testify. I would be happy to address any questions the Committee might have.


1. The report may be viewed on the Web site of the National Academies Press.


An archived transcript of the hearing can be found on the House Veterans' Affairs Committee's Web site