Treatment of Posttraumatic Stress Disorder: An Assessment of the Evidence (2008)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Summary T he Institute of Medicine (IOM) Committee on Treatment of Post- traumatic Stress Disorder (PTSD) was charged by the Department of Veterans Affairs (VA) to review and assess the evidence on the efficacy of pharmacologic and psychologic treatment modalities for PTSD (see Box S-1 for the complete Statement of Task). The committee was given five major tasks: review the scientific evi- dence and make conclusions regarding efficacy; note restrictions of the conclusions to certain settings, populations, and so on; comment on gaps and future research; answer several questions related to the goals, timing, and length of treatment; and finally, note areas where the evidence base is limited by inadequate attention or poor quality. This report contains the committee’s conclusions about the strength of the evidence regarding the efficacy of various treatment interventions. There are two important qualifiers of the committee’s underlying objective in responding to its charge. First, the committee was not asked to develop clinical practice recommendations, but to reach evidence-based conclusions that would inform policy decisions. Second, concluding that the evidence is inadequate to determine efficacy is not the same as concluding that a treat- ment modality is inefficacious. In responding to its charge, the committee found the evidence inadequate to determine the efficacy of most treatment modalities (see Statement of Task II.C.3). The committee did not examine the many factors that contribute to recommendations for clinical practice, including clinician and patient preferences, access, safety, availability, cost, alternatives, local practice patterns, medicolegal issues, and ethical con- cerns. The committee did not conclude that the evidence for any treatment  

 TREATMENT OF POSTTRAUMATIC STRESS DISORDER BOX S-1 Statement of Task I. The Department of Veterans Affairs has asked the IOM to convene a new committee to review the literature on various treatment modalities (including pharmacotherapy and psychotherapy) and treatment goals for individuals with PTSD. II. Specifically, the committee will conduct an evidence-based review on best treatment practices and types and timing of specific interventions, and com- ment on the prognosis of individuals diagnosed with PTSD (and existing comorbidities). As part of its assessment, the IOM committee shall: a. Develop descriptive evidence tables including type of study and identify potential bias and generalizations of the study. The committee shall also search for and classify systematic and narrative reviews on the topic of treatment and recovery of individuals with PTSD. b. The committee shall examine and classify the existing studies on various treatment modalities for PTSD. The committee will report on the highest levels of evidence available. For each study the committee will consider the quality of design and execution, and will be guided by the following classification: I Randomized controlled trial II-1 Controlled trial without randomization II-2 Cohort or case-control study II-3 Time series or uncontrolled experiment III Opinion of respected authority, case report, and expert committee c. The committee shall consider the following framework to make conclusions about the strength of the available evidence for treatment modalities: 1. Evidence is sufficient to conclude the efficacy of X in the treatment of PTSD. (A qualifier of magnitude may be added if appropriate.) modality­ was suggestive that it was ineffective or harmful (see Statement of Task II.C.4 and C.5). The committee conducted a systematic and comprehensive search of the relevant published literature and identified a total of 2,771 studies, and from that list included only randomized controlled trials (RCTs; placebo- controlled pharmacotherapy trials and wait-list or similar controls in the psychotherapy trials) in its review. The committee identified 37 RCTs on pharmacotherapies and 52 studies on psychotherapies (see Chapter 2 for more details about the committee’s methods). The committee excluded 

SUMMARY  2. Evidence is suggestive but not sufficient to conclude the efficacy of X in the treatment of PTSD. (The committee may note inconsistencies in the data.) 3. Evidence is inadequate to determine the efficacy of X in the treatment of PTSD. 4. Evidence is suggestive that X treatment is ineffective in treating PTSD. 5. Evidence is suggestive that X treatment is harmful in the treatment of PTSD. d. For each of the conclusions above, the restriction of the conclusion regardi­ng the population, provider, setting [of] intervention, or comparator intervention will be noted. III. As part of its assessment, the IOM committee shall note limitations in the evi- dence base and make suggestions for further research that could strengthen the evidence or address research gaps in the treatment of PTSD. IV. In conducting its work, the committee shall consider the following questions in relation to treatment modalities (including pharmacotherapy and psycho- therapy) and treatment goals for individuals diagnosed with PTSD. a. What are the goals of PTSD treatment? • What is the definition of recovery? • For what proportion of patients is recovery possible? • Besides recovery, what other outcomes would benefit patients? b. Does evidence support the value of early intervention? c. How long should treatment continue? • What is the impact of a hiatus in treatment? • What is the impact of periodic reexamination for asymptomatic patients? V. The committee shall note when the evidence base does not allow for respond­ ing to these questions due to insufficient research attention or poorly con- ducted studies. nonrandomized and uncontrolled studies for several reasons. It is extremely difficult to answer questions of efficacy in an uncontrolled way because of the variability of treatments, outcome measures, disease course, and patient choice. RCTs are the most reliable form of evidence for efficacy, and the committee found that the characteristics of the disorder, its measurement, and its treatment are sufficiently heterogeneous that observational studies were unlikely to provide useful evidence beyond the data available from RCTs. Therefore, per part II.B of the Statement of Task, all studies included in this review are classified as level I evidence. 

 TREATMENT OF POSTTRAUMATIC STRESS DISORDER ISSUES IN PTSD TREATMENT RESEARCH The committee encountered several noteworthy issues in its review and evaluation of the evidence base. First, there is some suggestion that there may be differences between civilian populations and veteran populations with PTSD in their response to treatment and to various types of treatment (Stein et al., 2006; van der Kolk, 2007). However, the committee cannot comment conclusively on this matter because the evidence neither demon- strates that there are differences between the two populations, nor does it show that the two groups are indistinguishable in their response to treat- ment. The committee also notes that the populations of veterans with PTSD now returning from Iraq and Afghanistan might be different enough from U.S. veterans from previous wars such that studies of the latter populations (mostly dating back to the Vietnam conflict) may be minimally informative about treatment efficacy in veterans of the recent conflicts. Second, the committee examined the question of treatment efficacy in PTSD in general populations, not just PTSD in veterans, but found it strik- ing that so few of the studies were conducted in populations of veterans. Third, the committee found problems in the design and performance of studies, many apparently due to the difficulties of conducting research in this clinical domain (Harvey et al., 2003). Design problems included lack of assessor blinding or assessor independence in the psychotherapy ­studies, small sample size, and lack of follow-up for individuals who dropped out before the trials ended. The problems of high dropout rates and weak h ­ andling of missing data, which have the potential to introduce significant bias, were frequent in both pharmacotherapy and psychotherapy studies and are discussed in Chapter 5 and Appendix D. High dropout rates are a particular problem in this domain, and regardless of how they are handled, they reduce the certainty of study results. Often studies reported data only on those completing therapy, a strategy biased in favor of showing a treat- ment effect. Those studies incorporating a strategy to deal statistically with the dropouts usually used “last observation carried forward,” a method that may bias results in either direction depending on context. The committee sought to address these issues by taking the following steps: 1. basing conclusions on evidence satisfying basic quality criteria (see Box S-2 and Chapter 2);   The Cochrane systematic review of pharmacotherapy for PTSD notes the following: “ . . . combat veterans (this subgroup has been regarded as more resistant to treatment, and is arguably more likely to have more chronic and severe symptoms, to have comorbid depression, and to be male)” (Stein et al., 2006: 7). 

SUMMARY  BOX S-2 Criteria to Assess a Study’s Quality • Assembly of comparable groups (randomized,** similar distributions of known confounders). • Maintenance of comparable groups (i.e., minimal attrition, crossovers, or con- tamination, good adherence). Use of intention to treat (ITT) analysis. • Measurements equal, valid, and reliable (validated PTSD outcome measure, double masking in pharmacotherapy studies** and assessor blinding or at least assessor independence** in psychotherapy studies). • Loss to follow-up causing missing outcome data: — Differential loss to follow-up no greater than 15% absolute difference be- tween groups.** — If approximately equal loss to follow-up in each arm, study quality is af- fected by the analytic methods used to handle missing data: o Up to 10% missing outcome data acceptable without formal missing data methods employed (i.e., may use completer analysis or last ob- servation carried forward [LOCF]). o Between 10% and 40% missing outcome data acceptable depending on validity of missing data analytic method employed (e.g., for lower proportions, single imputation, for higher proportions, likelihood-based methods, multiple imputation, sensitivity analysis). o Use of LOCF decreases study quality as the percentage dropout in- creases, severely if dropout exceeds 30%. Completer analysis is not acceptable.** o No more than 40% loss to follow-up in any arm.** **Indicates a criterion that if absent (or if the authors do not disclose) is a major limitation that limited the study’s usefulness to the committee in reaching its conclusion regarding efficacy.  2. providing commentary to put the conclusion statements in the broader clinical and research context (see Chapters 3 and 4); and 3. describing opportunities and making recommendations for improv- ing the validity and applicability of future PTSD treatment studies (see Chapter 5). Third, the committee found that the evidence fails to address the effects of high rates of comorbidity among veterans with PTSD, especially major 

 TREATMENT OF POSTTRAUMATIC STRESS DISORDER depression, traumatic brain injury, and substance abuse. Thus the commit- tee’s conclusions regarding efficacy overall may not apply to the substantial proportion of veterans with one or more important comorbidities. Further, the committee notes that the evidence is mostly silent on the acceptability, efficacy, or generalizability of treatment in ethnic and cultural minorities, as few studies stratified results by ethnic background. The committee ex- pects that the psychotherapies in particular might pose special challenges in different cultural groups but was unable to comment because none of the studies addressed it. A recommendation on important subpopulations is provided in Chapter 5 and below. CONCLUSIONS Below, the committee’s conclusions about each class of treatment are provided, first for the pharmacotherapy modalities and then for psycho- therapy modalities. Evidence tables summarizing key data and references are provided in Chapter 3 for pharmacotherapy and in Chapter 4 for psychotherapy. Pharmacotherapies The committee reviewed 37 pharmacotherapy studies and divided them by class where the number of studies made that useful, and into more general categories for small numbers of studies for a given class. Head-to- head studies in classes not proven efficacious on the basis of randomized placebo-controlled trials were excluded. • The committee reviewed two RCTs of alpha-adrenergic blockers. The studies that were excluded were open-label trials, a retrospec- tive chart review, and a study that did not use an overall PTSD outcome measure. • The committee reviewed eight studies of anticonvulsants and ex- cluded five (all open label, one a maintenance study). • The committee reviewed 10 RCTs of novel antipsychotics (namely, olanzapine and risperidone) and excluded three studies that were open label or head-to-head. • The committee included one study of benzodiazepines and ex- cluded all that were open label or did not include an overall PTSD outcome (e.g., focus on sleep only). • The committee found the literature on selective serotonin ­reuptake inhibitors (SSRIs) most extensive of all classes of medication. The committee included 14 studies in its review and excluded 15 studies­. Of the seven studies judged most informative with 

SUMMARY  respect to ­efficacy, four showed a positive effect on primary PTSD outcomes, and three did not. The largest trial conducted in male combat veterans used LOCF with 30% dropout and did not dem- onstrate an improvement in primary PTSD outcomes. • The committee’s review included four RCTs of MAOIs (monoamine oxidase inhibitors) (two each phenelzine and brofaromine) and ex- cluded four additional studies that were open label, uncontrolled, or for one study, a head-to-head comparison with moclobemide. • In its review of other antidepressants, the committee identified one RCT each for the following drugs: tricyclic antidepressants imipramine and amitriptyline, mirtazapine, and nefazodone. The committee also reviewed two large RCTs of venlafaxine. • In the category of “other drugs,” the committee reviewed one study of inositol and one study of cycloserine. The committee also made note of one RCT of opioid antagonist naltrexone in patients with alcohol dependence, which did not meet inclusion criteria, that it suggested a benefit to using naltrexone in an important subpopulation. For the all drug classes and specific drugs reviewed in each of the fol- lowing classes, the committee concludes that the evidence is inadequate to determine efficacy in the treatment of PTSD: • alpha-adrenergic blocker prazosin, • anticonvulsants, • novel antipsychotics olanzapine and risperidone, • benzodiazepines, • MAOIs phenelzine and brofaromine, • SSRIs, • other antidepressants, and • other drugs (naltrexone, cycloserine, or inositol). Important comments are appended to the conclusions for alpha- a ­ drenergic blockers, novel antipsychotics, benzodiazepines, and SSRIs. One committee member does not concur with the committee’s consensus on two conclusions—on SSRIs and novel antipsychotic medications—and offers alternate conclusions (see Appendix H). Please refer to Dr. Thomas Mellman’s minority opinion on this conclusion in Appendix H. Please refer to Dr. Thomas Mellman’s minority opinion on this conclusion in Appendix H. 

 TREATMENT OF POSTTRAUMATIC STRESS DISORDER Psychotherapies The committee’s search of the psychotherapy literature resulted in 52 studies. The committee organized the psychotherapy treatments into several categories based on how they appeared in the literature; this categorization also enabled the committee to draw meaningful conclusions. The majority of the studies reviewed included one or more cognitive-behavioral therapy (CBT) approaches. The largest proportion of CBT studies included an e ­ xposure-based therapy. The committee recognized that exposure is fre- quently administered in combination with another CBT technique, and that led the committee to group together studies with exposure and exposure plus something else (such as cognitive restructuring or a coping skills train- ing modality [e.g., relaxation]). The next largest category was eye movement desensitization and reprocessing therapy, or EMDR. Although EMDR has a CBT component, the committee evaluated this research separately from exposure and other CBT in recognition of the ongoing debate about the theoretical underpinnings of EMDR and the contribution of various EMDR components in PTSD treatment (Foa et al., 2000; Power et al., 2002). The committee also examined cognitive restructuring studies separately, in cases where the approach was not explicitly combined with exposure. The committee then reviewed coping skills therapies such as relaxation and bio- feedback. The committee identified a few other psychotherapies with fairly limited evidence and assessed their results as a group. The “other” category included hypnotherapy and psychodynamic therapy. Finally, the committee reviewed studies employing a group format psychotherapy. As with the pharmacotherapy studies, the committee first considered studies that compared the intervention of interest to a control. In the case of the psychotherapy studies, the control generally was assignment to a wait list, and less frequently to minimal care or usual care. In some studies, the control was active, and the committee considered those studies next. Finally, head-to-head studies in classes of psychotherapy not proven effica- cious on the basis of randomized, wait list, or equivalent-controlled trials were excluded. The committee reviewed 23 RCTs of exposure-based treatments, some of which included in the same treatment condition (or arm) exposure plus cognitive restructuring or exposure plus coping skills training. The committee finds that the evidence is sufficient to conclude the e ­ fficacy of exposure therapies in the treatment of PTSD. The committee reviewed a small number of studies comparing exposure to another psychotherapy approach. However, this body of literature was 

SUMMARY  characterized by many limitations, making it impossible to reach a conclu- sion regarding the equivalency of exposure and any other psychotherapy. The committee also reviewed studies of EMDR, cognitive restructuring, coping skills training, and other psychotherapies: • The committee reviewed 10 RCTs of EMDR compared to wait list and other psychotherapy approaches or wait list alone. Many studies were excluded because they were comparison trials, did not have a comparison group, or only a portion of the sample had diagnosed PTSD. • The committee reviewed three RCTs of cognitive restructuring. • The committee reviewed four RCTs of coping skills and excluded one study because it did not have a control or comparison group. The committee concludes that the evidence is inadequate to determine the efficacy of the following psychotherapy modalities in the treatment of PTSD: • EMDR • cognitive restructuring • coping skills training In the category of “other psychotherapies”, the committee reviewed a total of four RCTs of eclectic psychotherapy (two studies), hypnotherapy, psychodynamic therapy, and brainwave neurofeedback. Based on these single trials, the committee felt that it would be inappropriate to reach a conclusion regarding the efficacy of these treatments. Finally, the committee reviewed four studies utilizing a group therapy format. The committee concludes that the evidence is inadequate to determine the efficacy of therapy delivered in group formats in the treatment of PTSD. FINDINGS AND RECOMMENDATIONS In response to VA’s important questions related to recovery and the length and timing of PTSD treatment, and considering the gaps in the research, the committee makes eight recommendations. More detail is provided in Chapter 5. 

10 TREATMENT OF POSTTRAUMATIC STRESS DISORDER Finding 1.  The committee found that treatment of PTSD has not re- ceived the level of research activity needed to support conclusions about the potential benefits of treatment modalities. Although prog- ress in scientific standards can be observed, and recent studies tend to provide more useful information than older studies, important limita- tions remain. There are very few large scale, multi-site initiatives of the type that has been directed toward other psychiatric disorders. The studies conducted over the nearly three decades since Diagnostic and Statistical Manual of Mental Disorders (DSM) adoption of the PTSD definition do not form a cohesive body of evidence about what works and what does not. As described elsewhere in this report, studies have used a wide variety of outcome measures and lengths of treatment (for the same treatment). Further, many studies lack basic characteristics of internal validity including high dropout rates handled with weak missing data analyses and high differential dropout among treatment arms. (Other characteristics include follow-up of all patients admitted to the trials, attention to conflict of interest, assessor independence, and length of follow-up.) Although experts in the field (Foa and Meadows, 1997; Harvey and Bryant, 2003) have called for setting research stan- dards that would strengthen methodologic quality and internal validity, more work is needed. Recommendation 1. The committee recommends that VA and other funders of PTSD research take steps to identify and require investi- gators to use methods that will improve the internal validity of the research, with particular attention to standardization of treatment and outcome measures, follow-up of individuals dropping out of clinical trials, and handling of missing data. Finding 2. The committee found that the majority of drug studies were funded by pharmaceutical manufacturers. This is an issue that has re- ceived much attention in recent years from the academic research com- munity, government agencies, patient communities, and the editors of major biomedical journals. The committee also found that many of the psychotherapy studies were conducted by individuals who developed the techniques or their close collaborators. It is important to know whether these treatments would show the same effect if implemented in other settings, requiring the confirmation and replication of these research results by other investigators. Recommendation 2. The committee recommends that VA and other funders of PTSD treatment research seek ways to give opportunities to 

SUMMARY 11 a broad and diverse group of investigators to ensure that studies are conducted by individuals and in settings without potential financial or intellectual conflicts of interest. Finding 3. The committee found that the available research leaves significant gaps in assessing the efficacy of interventions in important subpopulations of veterans with PTSD, especially those with traumatic brain injury, major depression, other anxiety disorders, or substance abuse, as well as ethnic and cultural minorities, women, and older individuals. Recommendation 3. The committee recommends that VA assist clini- cians and researchers in identifying the most important subpopulations of veterans with PTSD and designing specific research studies of inter- ventions tailored to these subpopulations. Finding 4. The committee found that research on treatment of PTSD in U.S. veterans is inadequate to answer questions about interventions, settings, and lengths of treatment that are applicable in this specific population. The committee recognizes that the successful conduct of research directly applicable to veterans will require close collaboration among funding agencies (Department of Defense, National Institute of Mental Health, National Institute of Alcohol Abuse and Alcohol- ism, National Institute of Drug Abuse), veterans’ groups, and clinical service settings. Specifically veterans groups could make considerable contributions to the design and conduct of high-quality research on the treatment of PTSD.   Recommendation 4. The committee recommends that Congress require and ensure that resources are available for VA and other r ­ elevant federal agencies to fund quality research on the treatment of PTSD in veteran populations and that all stakeholders are included in research plans. Finding 5. The committee found that studies of PTSD interventions have not systematically and comprehensively addressed the needs of veterans with respect to efficacy of treatment and the comparative ef- fectiveness of treatments in clinical use.   

12 TREATMENT OF POSTTRAUMATIC STRESS DISORDER Recommendation 5. The committee recommends that VA take an ac- tive leadership role in identifying research priorities for addressing the most important gaps in evidence in clinical efficacy and comparative effectiveness. Potential areas for future research include: • Comparisons of psychotherapy (e.g., CBT) and medication; • Evaluation of the comparative effectiveness of individual and group formats for psychotherapy modalities; and • Evaluations of the efficacy of combined psychotherapy and medication, compared with either alone, and compared with control conditions. Combined treatment could be tested within study designs like those that have been applied in large studies for other psychiatric conditions. Finding 6. The committee found no generally accepted and used defi- nition for recovery in PTSD. Also, many studies used measures of questionable validity and reliability instead of validated, high-quality measures such as the Clinician-Administered PTSD Scale (Foa et al., 2000). The committee places the lack of agreement about recovery in context of a more general concern about identifying appropriate out- comes for PTSD research. Recommendation 6. The committee recommends that clinicians and researchers work toward common outcome measures in three general domains that relate to recovery: loss of PTSD (DSM) diagnosis, PTSD symptom improvement, and end-state functioning. The committee fur- The committee has noted with interest research on effectiveness in other areas of mental health. The STAR*D (Sequenced Treatment Alternatives to Relieve Depression) study aimed to reproduce some real-life settings in allowing participants’ choice and offering alternatives when a course of treatment did not work, and used an outcome measure of “remission” meaning becoming symptom free. Another study brought to the committee’s attention is the CATIE (Clinical Antipsychotic Trials of Intervention Effectiveness) Schizophrenia study, which compares newer atypical antipsychotics with each other and with conventional antipsychot- ics in regard to long-term effectiveness and tolerability, and also in identifying antipsychotics that work for patients who have not had success with that class of drugs. Finally, STEP-BD (Systematic Treatment Enhancement Program for Bipolar Disorder) is a long-term study of manic-depressive illness that studied treatment (both pharmacologic and psychosocial) of af- fected individuals on two “pathways”—one a naturalistic, best practices pathway that allowed patients and clinicians to choose the best course of treatment, and the other a “randomized care pathway” that involved patients in multi-site randomized controlled trials. Program participation lasted for up to 5 years to facilitate adequate follow-up.  The committee found one study that does this in the work of van der Kolk (2007) and colleagues. 

SUMMARY 13 ther recommends the following three principles be considered in the selection of outcome measures: • validity in research; • convergence on a core of common outcomes for the purpose of comparability; and • usefulness to clinicians to assess patients over time as symp- toms and function change. The committee recommends that VA assume a leadership and conven- ing role and work with other relevant federal agencies in developing these common approaches. Finding 7. The committee was unable to reach a conclusion on the value of intervention early in the course of PTSD based on the treat- ment literature it reviewed. Recommendation 7. The committee recommends that VA and other government agencies promote and support specific research on early intervention (i.e., reducing chronicity) in PTSD. The committee further recommends that future research specify both time since trauma expo- sure and duration of PTSD diagnosis, and that interventions be tested for efficacy at specific clinically meaningful intervals, as interventions might be expected to vary in effectiveness related to time since exposure and duration of diagnosis. Finding 8. The committee was unable to draw conclusions regarding opti- mal length of treatment with psychopharmacology or psychotherapy. Recommendation 8. The committee recommends that VA and other funders call for research on the optimal duration of various treatments. Trials of comparative effectiveness of different treatment lengths for those treatments found efficacious should follow. Finally, studies with adequate long-term (i.e., greater than one year) follow-up should be conducted on treatments of any length found to be efficacious. CONCLUDING OBSERVATIONS In this report the committee sought to describe the evidence regarding the efficacy of available treatment modalities for PTSD, identify some of the major issues in the field, and make recommendations to help guide further research in PTSD treatment. The committee’s findings, conclusions, and recommendations about the evidence for the treatment modalities reviewed 

14 TREATMENT OF POSTTRAUMATIC STRESS DISORDER in this report are not clinical practice guidelines. The committee does not intend to imply that, for example, exposure therapy is the only treatment that should be used in treating individuals with PTSD. The committee recognizes that the transparent presentation and assessment of evidence is just one part of the larger picture of PTSD treatment that includes many other factors. Further, assessing the scientific evidence may reveal areas of uncertainty. The next step in the process toward clinical decisionmaking is making recommendations for clinical practice—a step the committee was not asked to, and did not, take. The committee applied contemporary standards in evaluating ­studies, including research dating back to 1980 when PTSD was first formally d ­ efined. The principal finding of the committee is that the scientific evidence on treatment modalities for PTSD does not reach the level of certainty that would be desired for such a common and serious condition among veterans. For some modalities, for example novel antipsychotic drugs and SSRIs, the committee debated whether to characterize the body of evidence as “suggestive” or “inadequate.” It is important to emphasize that in the larger picture of PTSD treatment, had the debate ended with “suggestive” conclusions (rather than the “inadequate” conclusions the committee finally reached), the core message that better-quality research is needed would not have been rendered less urgent in consequence. The committee reached a strong consensus that additional high-quality research is essential for ­ very treatment modality. Applying the general recommendations outlined e above to exposure therapy, there is a need for better understanding of the most important and active components of exposure therapy, determining optimal administration and length of treatment, attention to principal subpopulations, and determining whether it can be effectively delivered in group format, presenting a challenging and urgent agenda for researchers and clinicians in the field. The committee views its more general findings and recommendations regarding further research to be as important as its conclusions regard- ing the evidence supporting treatment modalities. The committee became aware of the formidable challenges that researchers face in conducting high- quality studies of efficacy and comparative effectiveness. Nonetheless, the committee was able to identify studies that met the highest ­internationally accepted standards for randomized controlled trials (in assembling popu- lations, ­ administering treatment, measuring outcomes, and following up enrolled subjects), showing that such studies are possible even for such a difficult clinical condition as PTSD. As outlined in the committee’s recom- mendations in Chapter 5, setting a high standard for research on PTSD and delivering on it will require close collaboration between VA and other government agencies, researchers, clinicians, and patient groups. Thus, 

SUMMARY 15 the committee’s recommendations are its suggestions for setting a frame- work for the future that can more successfully address the critical needs of v ­ eterans who return to civilian life with the diagnosis of PTSD. REFERENCES Foa, E. B., and E. A. Meadows. 1997. Psychosocial treatments for posttraumatic stress d ­ isorder: A critical review. Annual Review of Psychology 48:449-480. Foa, E., T. Keane, and M. Friedman. 2000. Effective treatments for PTSD: Practice guidelines from the International Society for Traumatic Stress Studies. New York: The Guilford Press. Harvey, A. G., R. A. Bryant, and N. Tarrier. 2003. Cognitive behaviour therapy for post­ traumatic stress disorder. Clinical Psychology Review 23(3):501-522. Power, K., T. McGoldrick, K. Brown, R. Buchanan, D. Sharp, V. Swanson, and A. ­Karatzias. 2002. A controlled comparison of eye movement desensitization and reprocessing ­versus exposure plus cognitive restructuring versus waiting list in the treatment of post-­traumatic stress disorder. Clinical Psychology and Psychotherapy 9(5):299-318. Stein, D. J., J. C. Ipser, and S. Seedat. 2006. Pharmacotherapy for post traumatic stress d ­ isorder (PTSD). Cochrane Database Systematic Reviews (4):CD002795. van der Kolk, B. A., J. Spinazzola, M. E. Blaustein, J. W. Hopper, E. K. Hopper, D. L. Korn, and W. B. Simpson. 2007. A randomized clinical trial of eye movement desensitization and reprocessing (EMDR), fluoxetine, and pill placebo in the treatment of posttraumatic stress disorder: Treatment effects and long-term maintenance. Journal of Clinical Psy- chiatry 68(1):37-46.