U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Gartlehner G, Forneris CA, Brownley KA, et al. Interventions for the Prevention of Posttraumatic Stress Disorder (PTSD) in Adults After Exposure to Psychological Trauma [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2013 Apr. (Comparative Effectiveness Reviews, No. 109.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Interventions for the Prevention of Posttraumatic Stress Disorder (PTSD) in Adults After Exposure to Psychological Trauma

Interventions for the Prevention of Posttraumatic Stress Disorder (PTSD) in Adults After Exposure to Psychological Trauma [Internet].

Show details

Discussion

We conducted a systematic review of the efficacy and comparative effectiveness and harms of psychological, pharmacological, and emerging interventions for the prevention of PTSD in adults exposed to psychological trauma. Because of the ongoing controversy about whether different treatments are efficacious at all, we first assessed evidence for the efficacy of the treatments of interest and then proceeded to assess comparative effectiveness. We used this approach because our preliminary searches and input from experts during the topic refinement process suggested that we would find little head-to-head comparative evidence; we realized that we would likely need to rely on indirect evidence to attempt to draw conclusions about comparative effectiveness.

Below, we summarize the main findings and strength of evidence (SOE) by Key Question (KQ). We then discuss the findings in relationship to what is already known, applicability of the findings, implications for decisionmaking, limitations, research gaps, and conclusions.

Key Findings and Strength of Evidence

KQ 1. Efficacy and Comparative Effectiveness

The evidence for the effectiveness of psychological, pharmacological, and emerging interventions to prevent PTSD is limited. Overall, 52 studies met eligibility criteria for KQ 1.27,47,51,8397,99114,116133 Of these, 35 were rated as high risk of bias and omitted from the main data synthesis.27,47,51,8388,90,9297,99114,116118 Tables 38 and 39 summarize the main findings and the SOE for KQ 1.

Table 38. Summary of findings and strength of evidence for the efficacy of psychological, pharmacological, and emerging interventions to prevent PTSD and reduce PTSD symptom severity.

Table 38

Summary of findings and strength of evidence for the efficacy of psychological, pharmacological, and emerging interventions to prevent PTSD and reduce PTSD symptom severity.

Table 39. Summary of findings and strength of evidence for the comparative effectiveness of psychological, pharmacological, and emerging interventions to prevent PTSD and reduce PTSD symptom severity.

Table 39

Summary of findings and strength of evidence for the comparative effectiveness of psychological, pharmacological, and emerging interventions to prevent PTSD and reduce PTSD symptom severity.

We identified trials that reported on one or more of eight different psychological interventions: debriefing, cognitive behavioral therapy (CBT), CBT combined with hypnosis, cognitive therapy (CT), prolonged exposure therapy (PE), psychoeducation, self-help materials, and supportive counseling (SC). We included two trials that reported on two different medications: hydrocortisone and escitalopram. In addition, we identified one trial reporting on an emerging intervention: collaborative care. All these studies evaluated the efficacy of the intervention against an inactive control such as a waitlist, usual care, or placebo.

From these studies, we concluded that debriefing is not effective in preventing PTSD or reducing the severity of PTSD symptoms in civilian victims of crime, assault, or accident trauma at 6-month followup (low SOE) (Table 38). We had insufficient data (single study) to determine the efficacy of debriefing at 2- or 6-week followup, as well as at 11-month followup. From a single study involving civilian trauma patients requiring surgical hospitalization, we concluded that collaborative care produces a greater decrease in PTSD symptom severity at 6, 9, and 12 months after injury compared with usual care (low SOE). However, data addressing whether there was a difference in PTSD diagnosis 12 months after injury were not conclusive (insufficient SOE). For most interventions—namely, CBT, CBT combined with hypnosis, CT, PE, psychoeducation, self-help material, SC, and the two pharmaceuticals—we had single studies with small treatment arms (generally fewer than 80 subjects). This paucity of information led us to conclude that the evidence was insufficient to support their efficacy in preventing PTSD or reducing the severity of PTSD symptoms. When assessed, we arrived at the same conclusion for the effectiveness of these interventions in reducing comorbid symptoms of anxiety and depression.

A small number of studies evaluated the comparative effectiveness of two or more psychological interventions with one another or with a pharmacological intervention. Our meta-analyses of trials that compared CBT with SC in a sample of participants with acute stress disorder found that, at both the end of treatment and at 6-month followup, CBT was no more effective than SC for preventing PTSD (low SOE), reducing symptoms of anxiety (moderate SOE), or reducing symptoms of depression (low SOE) (Table 39). By contrast, related meta-analyses of the CBT compared with SC trials found that, at both the end of treatment and at 6-month followup, CBT was more effective than SC in reducing the severity of PTSD symptoms as measured by the IES (moderate SOE).

For many interventions, we had insufficient evidence to draw conclusions about efficacy. This, coupled with the fact that this knowledge base largely comprises single studies with small sample sizes, we concluded that the evidence was insufficient to determine the comparative effectiveness of most of the psychological interventions in preventing PTSD or reducing PTSD symptom severity. This includes whether Battlemind training is more effective than standard postdeployment debriefing for military personnel in the UK, CBT plus hypnosis is more effective than CBT or SC for civilian victims of nonsexual assault or motor vehicle accidents (MVA), various forms of debriefing, alone or in combination with psychoeducation for crime victims, and CT compared with PE for hospitalized survivors of MVAs or terrorist attacks. Only one study compared psychological interventions (CT and PE) with a medication (escitalopram), so we could not draw any conclusion about the comparative effectiveness of an selective serotonin reuptake inhibitors (SSRI) with a psychological intervention. No study evaluated the comparative effectiveness of two or more medications with each other. Finally, when assessed, we drew the same conclusion (insufficient evidence) for the comparative effectiveness of these interventions in reducing comorbid symptoms of depression and anxiety.

KQ2. Timing, Intensity, or Dosing

The evidence on the impact of timing, intensity, or dosing on the effectiveness or risk of harms of interventions used for the prevention of PTSD is scarce. Overall, four studies addressed timing and dosing questions;98,115,134,135 two of these were rated as high risk of bias.98,115 We found no studies on the impact of intensity of intervention for any psychological or emerging interventions. Table 40 summarizes the main findings and the SOE for KQ 2.

Table 40. Summary of evidence of the impact of timing, intensity, and dosing on the effectiveness of interventions and strength of evidence.

Table 40

Summary of evidence of the impact of timing, intensity, and dosing on the effectiveness of interventions and strength of evidence.

One randomized controlled trial (RCT) addressed the impact of timing of a psychological intervention. Immediate debriefing (within 10 hours) compared with late debriefing (after 48 hours) led to victims experiencing significantly fewer posttraumatic symptoms (insufficient SOE) (Table 40). No evidence was available on the impact of timing for any other psychological, pharmacological, or emerging interventions or any other outcomes.

In one RCT, dosing of sedation (light vs. deep) in critically ill patients did not affect either posttraumatic symptoms or symptoms of depression or anxiety (insufficient evidence). We did not find any eligible evidence on the effect of dosing for any other pharmacological or emerging interventions to prevent PTSD.

KQ 3. Subgroups

Evidence is also sparse on whether the effect of early interventions differs across demographic groups, psychiatric diagnoses, or personal risk factors for developing PTSD. Six studies that met our inclusion criteria27,107,120,129,131,134 included subgroup analyses; two were rated high risk of bias.27,107 Table 41 summarizes the main findings and the SOE for KQ 3.

Table 41. Summary of evidence and strength of evidence for the effect of early interventions in various subgroups.

Table 41

Summary of evidence and strength of evidence for the effect of early interventions in various subgroups.

Two trials assessed whether sex modified the effect of early psychological interventions on PTSD symptoms.129,134 Both reported consistent results that the effects of early psychological interventions on PTSD symptoms were similar for men and women (Table 41). However, because neither trial reported the magnitude of the estimated effect or its precision, we graded the SOE as low.

One trial tested the effect of a debriefing intervention on subgroups defined by sex, history of depression, and history of child abuse, but it did not report magnitude or precision of effects (SOE insufficient in all cases).129 No differences were found for any of these subgroup comparisons for the effect of debriefing.

Two trials provided inconsistent findings on whether baseline severity of PTSD symptoms modified the effect of early psychological interventions (SOE insufficient).120,131

KQ 4. Harms

Little evidence exists addressing the absolute risks and/or comparative risks of harms from early interventions to prevent PTSD. Four studies assessed harms;84,93,131,135 two were rated as high risk of bias.84,93 Table 42 summarizes the main findings and the SOE for KQ 4.

Table 42. Summary of findings and strength of evidence about harms.

Table 42

Summary of findings and strength of evidence about harms.

A three-armed RCT (low risk of bias) considered absolute risk in patients presenting to an outpatient psychiatric clinic after assault or an accident.131 In a subgroup of patients with early hyperarousal, those receiving emotional debriefing experienced higher PTSD severity at 6 weeks than those not receiving such debriefing (Table 42). The investigators did not find this difference in this subgroup at either 2 weeks or 6 months or in any other subgroups (insufficient evidence). We found no other trials of psychological or pharmacological interventions that provided information on risks of early interventions.

One randomized open-label study considered comparative risk of harms from light versus deep sedation for patients requiring mechanical ventilation.135 The two groups did not differ with regard to rates of mortality (whether during their stays in the intensive care unit or their overall hospitalization) or in the incidence of adverse events (organ dysfunction, hypertension, and tachycardia). The evidence was insufficient (single study) for us to draw any conclusions about harms for this intervention.

Findings in Relationship to What Is Already Known

As stated in the introductory chapter of this review, variability of types of trauma, contexts in which they occur, and individual differences of those exposed to traumatic events are likely to prohibit a “one size fits all” model for preventive intervention. Preventing PTSD is not easy in part because each individual responds to stress differently; thus, predicting who will develop PTSD is similarly difficult. Other challenges include identifying people at risk for PTSD and, from a logistics perspective, conducting “early” interventions in the aftermath of a traumatic event.

Our results found insufficient evidence to support the effectiveness of most psychological, pharmacological, and emerging interventions that have been studied to prevent PTSD. The two primary reasons are that (1) we generally had only one or two studies that addressed each intervention and (2) most of these studies had relatively small sample sizes.

Departments of Veterans Affairs and Defense Guidelines

Clinical practice guidelines published in October 2010 by the U.S. Department of Veterans Affairs and Department of Defense (VA/DoD)136 listed several possible early interventions: crisis intervention and stress management, the PIE (proximity, immediacy, and expectations) model for combat stress reactions, brief CBT, debriefing, and pharmacological interventions (benzodiazepines and sedatives). However, these guideline developers concluded that, among these, only time-limited CBT (four or five sessions) is the most effective in preventing PTSD in samples of survivors of sexual assault, nonsexual assault, and accidents.

We concluded that short-term CBT (5 to 6 weeks) is no more effective than SC in preventing PTSD for samples of civilian trauma survivors (motor vehicle accidents [MVAs], industrial accidents, and nonsexual assault) with acute stress disorder. We did find, however, that CBT was more effective than SC in reducing PTSD symptom severity (moderate SOE) with these same trauma samples.

Consistent with the VA/DoD guidelines, we concluded that debriefing was not effective in preventing PTSD or reducing PTSD symptom severity. The guidelines note that short-term use of benzodiazepines improved sleep and PTSD symptoms but that long-term use was associated with a higher incidence of PTSD at 6 months. One of the prospective cohort studies that met our eligibility criteria (but rated high risk of bias),100 which examined the efficacy of early intervention with benzodiazepines, reported that the treatment and control groups were not significantly in PTSD or anxiety symptoms at 1-month and 6-month followup. These guidelines also stated that no studies (in their evidence base) had evaluated the effectiveness of antidepressants for preventing PTSD. We identified one study that compared an SSRI with placebo and other psychological interventions,122 but it provided insufficient evidence to determine the absolute and comparative effectiveness of an SSRI compared with either CT or PE for preventing PTSD or for reducing PTSD symptom severity.

Finally, the VA/DoD guidelines stated that the effectiveness of PIE has not been confirmed and that there is some evidence that it is not effective in preventing PTSD. Our literature search did not identify any studies that used the PIE (proximity, immediacy, and expectations) model, so we cannot comment on this intervention or method of intervention.

National Center for Posttraumatic Stress Disorder

The National Center for Posttraumatic Stress Disorder and the VA/DoD identify psychological first aid as an appropriate post-trauma early intervention strategy.136 In general, psychological first aid is seen as a mental health correlate of physical first aid and not as a therapeutic or preventive intervention for PTSD. Guidelines published by the VA/DoD reported that there is insufficient evidence to recommend for or against the use of psychological first aid to address post-trauma symptoms beyond 4 days. Our search of the literature did not identify any studies on the effectiveness of psychological first aid in preventing PTSD.

Other Sources

Our literature searches did not identify a large number of studies that reported on prevention interventions after the 2001 World Trade Center disasters (WTCD) in the United States. Two studies of employees at the worksite after the WTCD were identified did not meet our eligibility criteria because of lack of objective information about the interventions used.137,138 The only data available about the interventions from this team of investigators were based on participant self-report; this precluded our categorizing them in our analyses or results. Although evidence from these studies is insufficient to draw any conclusions about the effectiveness of whatever crisis intervention strategies they had used, the studies were methodologically sound and were the only studies we found that reported on data on this traumatic event. Given this, we briefly summarize characteristics and main findings of these studies.

Both studies pertain to a prospective cohort of English- or Spanish-speaking adults living in New York City on the day of the WTCD identified using random-digit dialing. The first cohort was contacted 1 year after the WTCD, and the second 2 years after the event. In telephone interviews, the investigators asked subjects whether they had attended brief sessions related to coping with the WTCD and questions about alcohol use and mental health status including PTSD symptom severity. Compared with subjects who had not attended any intervention sessions, subjects who received two to three sessions experienced a reduction in PTSD symptom severity (OR, 0.36; p<0.05) at 1 year after the WTCD137 and lower PTSD symptom severity (0.8 percentage points lower, p<0.05) during the month before the assessment.138 At the 2-year followup, 80 percent of the sample reported that the brief sessions were helpful for coping with the disaster.138

Despite the limitations of these studies (i.e., lack of randomization, inability to contact subjects without a telephone, excluding individuals who did not speak English or Spanish), their results suggest brief postdisaster crisis interventions may be effective after mass exposure to psychologically traumatic events. The reasons for the effectiveness of these interventions are unclear and warrant further investigation. Whether the findings of this intervention would generalize to other types of disasters and trauma samples also remains unclear.

Implications for Clinical and Policy Decisionmaking

Treatment guidelines from the VA/DoD outline a four-stage framework for early responses to traumatic events15: (1) provide concrete help, food, warmth, and shelter; (2) soothe and reduce states of extreme emotion and increase controllability; (3) assist survivors with distressing and repetitive reappraisals of the trauma; and (4) treat specific syndromes and disorders such as acute stress disorder, depression, and anxiety.

The first step in this framework is clear, but clinical uncertainty exists about what interventions to use to address the other three. Initially, practical considerations, such as presence or lack of availability of trained personnel or resources to implement treatment, will likely guide treatment decisions in the immediate aftermath of a traumatic event. Once safety and basic needs stabilization has occurred, which early psychological or pharmacological interventions would be most effective and the least harmful in preventing PTSD cannot be specified from our results. If individuals have access to a variety of early interventions and they do not have a preference for a particular type of intervention, our findings would support the use of brief CBT interventions over SC for possibly reducing PTSD symptom severity; this was the intervention for which we had the most information to support efficacy (although low SOE). If an individual prefers medication, our results did not identify any class of drugs that has been shown to be effective in preventing PTSD.

Evidence on the impact of timing, intensity, or dosing on the effectiveness or risk of harms of interventions used for preventing PTSD is extremely scarce. The single trial of critical incident stress debriefing (CISD) timing indicated that early CISD was more effective than late CISD in reducing the number of posttraumatic symptoms in victims of robbery.134 However, because of the lack of an inactive control group in this study, no firm conclusions about a greater efficacy of early debriefing can be drawn. It is conceivable that the difference between immediate and delayed intervention can be attributed to harmful effects of delayed debriefing relative to no beneficial effects of immediate debriefing.

With respect to subgroups, we had similarly very little evidence. Clinicians may not need to take the sex of a patient into consideration when choosing a preventive intervention for at least some types of traumas, for example, nonsexual assault crime or robbery victims (low SOE). Whether that would generalize to other types of trauma, such as sexual assault, is unclear. We found no evidence about which early interventions are more or less effective for various other subgroups of interest (defined by sociodemographic characteristics, history of psychological conditions such as depression, or history of traumas that might put victims at increased risk of PTSD).

Evidence addressing the absolute risks and/or comparative risks of harms from early interventions for the prevention of PTSD was similarly insufficient. Of note, concern has been raised as to whether emotional debriefing might worsen symptoms associated with PTSD. The one study addressing this question found higher PTSD scores with emotional debriefing than with no debriefing (test of interaction, p=0.005) at 6-week followup but no differences at 2 weeks or 6 months.131

We would like to emphasize though that these results do not apply to the treatment of PTSD, a topic that is not covered in this report.

Applicability

The scope of this review was limited to studies that enrolled adults exposed to psychological trauma who were at risk for PTSD. We did not attempt to review literature on preventive interventions for PTSD in children exposed to psychological trauma.

The included studies covered diverse populations exposed to a wide range of traumas. Nevertheless, we had little or no evidence about terrorist attacks, sexual assault, natural disaster, or combat. Many studies were conducted in civilian populations outside the United States; the applicability of their results to patients or victims within this country is uncertain. The mean age of subjects was generally in the 30s to 40s, but some studies enrolled slightly older populations. Some studies screened participants for posttraumatic symptoms before enrollment and generally included them in their samples. Although these participants did not have a diagnosis of PTSD per se (because their symptoms had lasted less than 1 month), results from studies in such selected populations with acute stress symptoms might have little applicability to average populations exposed to psychological trauma who may not have acute stress symptoms at the time of the intervention. Furthermore, most studies were conducted in clinical settings. Results from samples of patients attending a clinic might not apply to members of the general community who suffered traumatic experiences of the same type. Therefore, evidence, in general, was insufficient to determine whether findings are applicable to all those at risk for PTSD from this heterogeneous set of traumatic events or, possibly, applicable only to certain groups.

Similarly, we did not find evidence to confirm or refute whether treatments are more or less efficacious for various subgroups: patients characterized by sex, race, or ethnicity; refugee status; first responders; victims of either natural or manmade disasters; or individuals with coexisting psychiatric conditions or with a history of events that might have put them at risk of PTSD. The samples used in many studies had some subjects with the aforementioned subgroup characteristics, even if the main focus was on a different population. For instance, the trials may have included individuals with a history of multiple past traumas, service-connected disability, or coexisting psychiatric conditions such as depression. Three drawbacks posed challenges for this review and limit anything we might say about applicability: we had only a single study per intervention/population combination, many studies did not publish details about some attributes of their subjects (e.g., race or ethnicity), and generally investigators did not report whether interventions were efficacious for individuals in subgroups that they may have in fact included.

For the few interventions for which we did have low or moderate SOE for effectiveness (collaborative care over usual care and CBT over SC for people with acute stress disorder, for instance), we could not say with confidence that these results could be generalized to other patient populations or other types of trauma.

In addition, many studies were conducted outside the United States with civilian populations (not U.S. military abroad). For example, all three trials comparing CBT with SC were conducted by the same research group in Australia. Whether and how differences in ethical or cultural backgrounds and health systems affect the applicability of results to U.S. populations remains uninvestigated and unanswered.

Finally, for many interventions we lacked sufficient (or indeed any) evidence to draw conclusions about either absolute or comparative effectiveness. Consequently, we cannot draw any conclusions about applicability.

Limitations of the Comparative Effectiveness Review Process

To find relevant studies, we employed an intensive search process in multiple electronic databases; we also conducted searches for grey literature. Because of time and monetary limitations, however, we limited eligible studies to those published in English. Methods research indicates that such an approach can introduce language bias; however, in general, it may also lead to overestimates of the effectiveness of interventions.

For KQs 1 through 4, we included RCTs and prospective cohort studies. For KQ 4, focused on harms, we also reviewed retrospective cohort studies or case-control studies without any sample size limits. We chose this broad approach because preliminary searches indicated that, for many interventions, no evidence from controlled trials is available. Nevertheless, these eligibility criteria might still have led us to exclude some observational studies that might have provided useful information, particularly for generating hypotheses for future studies with more rigorous study designs.

For harms, studies conducted in other populations (i.e., those without PTSD) might have yielded useful information. Such studies could, for instance, could provide important information about adverse effects of medications that might be used to prevent PTSD or its symptoms.

If information in full-text articles was unclear or missing, we attempted to contact authors for clarification. The yield of this effort, however, was small. Despite multiple attempts to contact authors, few replied or were able to provide missing information. Lack of information regarding the timing of interventions, in particular, makes evaluating preventive interventions for PTSD difficult; the main reason is that only interventions administered within the first 3 months after the exposure to psychological trauma qualify as preventive interventions.

Finally, publication bias and selective outcome reporting are potential limitations. Although we searched for grey and unpublished literature, the extent and impact of publication and reporting bias in this body of evidence is impossible to determine.

Limitations of the Evidence Base

Overall, two major limitations characterize this body of evidence. First, no eligible evidence was available assessing the efficacy or comparative effectiveness or risk of harms of many of our eligible interventions. Despite our broad eligibility criteria, including observational studies for effectiveness and harms, we could not draw conclusions for or against benefits and harms for the majority of our interventions of interest. Even when studies assessing the effectiveness of an intervention were available, they often did not assess harms. For example, we included 17 studies for KQ 1 on the efficacy and comparative effectiveness of interventions, but only two studies of low or medium risk of bias provided data on the risk of harms. Although lack of evidence cannot be equated with lack of effectiveness or harms, incautious use of interventions without proven net benefit has the potential of causing more harms than benefits. Important questions, for instance, whether particular interventions such as single-session emotional or educational debriefing might worsen symptoms, remain unanswered.

Second, available evidence was frequently fraught with methodological shortcomings. Of the 56 studies meeting our eligibility criteria, we rated 37 as high risk of bias and only 3 as low risk of bias. Studies assessed as high risk of bias have significant flaws of various types (e.g., stemming from serious errors in design, conduct, or analysis) that may invalidate their results. Consequently, the evidence base for most critical outcomes was insufficient to draw conclusions. The evidence for only a few outcomes could be rated as low or moderate; the latter indicates reasonable confidence in effect estimates of those studies.

Research Gaps

For KQ 1, we have a long way to go before we can conclude which psychological and pharmacological interventions are effective in preventing PTSD. We had insufficient data for the majority of psychological and pharmacological interventions that have been used to prevent PTSD; moreover, many studies that we did identify were rated as high risk of bias. For the few interventions for which we did have low (collaborative care over usual care for individuals with traumatic injuries requiring surgical hospitalization) or moderate SOE for effectiveness (CBT over SC for individuals with ASD), whether these results would generalize to other trauma types or samples is simply unknown. The studies that compared CBT with SC in individuals with ASD were all conducted in Australia by the same group of researchers. Ideally, these studies would be replicated in other samples by other researchers in order to determine the generalizability of these findings. Consistent with other reviews,32,36,38 we also concluded that psychological debriefing is not useful for preventing PTSD. One of these reviews also concluded that debriefing could actually be harmful to participants and should cease;38 this is not something we can conclude from our evidence base.

Surprisingly, we had essentially no studies reporting on preventing PTSD in the aftermath of natural disasters; those that did were rated as high risk of bias. Whether this lack of studies reflects logistical difficulty of providing treatment immediately after a natural disaster or some other barrier warrants further examination. Most of the studies included in our analyses were not conducted with samples in the United States, although every year thousands of people living in the United States are exposed to tornados, hurricanes, floods, MVAs, crime, and so on. What this reflects—lack of financial support for prevention intervention, poorly designed or conducted research, fear of prematurely offering medication because of possible adverse side effects, or other issues—is not clear. Groups with strong interests in the mental health of U.S. citizens could use the results of this report to advocate for increased attention, efforts, and resources targeted at PTSD prevention.

As noted, many of the studies we identified were rated as high risk of bias. Specific and important methods flaws that we identified included the following:

  • Inadequate randomization procedures
  • High rates of loss to followup
  • Inadequate statistical approaches for data analysis (e.g., lack of intention-to-treat [ITT] analysis, or lack of statistical adjustment for significant between-group differences at baseline)

An important task of systematic reviews is to assess design and conduct of included studies and whether they provide adequate protection against bias. The methodological shortcomings of many studies conducted for the prevention of PTSD substantially limit our confidence that results accurately reflect the truth. Therefore, the focus of this report is on evidence from studies rated as having low or medium risk of bias. We summarize findings of high-risk-of-bias studies if no other evidence is available, but readers have to be aware that we have no confidence in the estimates of effect from such studies.

The inherent nature of conducting research in the immediate aftermath of a traumatic event may make addressing some of the methods problems difficult. For example, loss to followup in the wake of a natural disaster, when subjects may be living in temporary housing and then relocate during the course of the study, can be especially challenging. By contrast, other flaws are more readily addressed or under the investigators’ control, such as adjusting for between-group differences at baseline. Future research of prevention interventions for PTSD should strive to minimize these types of flaws to increase the efficacy and generalizability of their results. This gap might best be bridged through active interdisciplinary consultation with statisticians, clinical epidemiologists, and other experts in the public health sector.

For KQ 2, on timing, intensity, and dosing of interventions, eligible evidence was generally entirely missing or of high risk of bias. Even considering two studies of medium risk of bias, we could draw no conclusions about these for any of the eligible interventions. Focusing on timing, intensity, and dosing may be premature given the lack of robust support for the efficacy of any psychological or pharmacological intervention to prevent PTSD. We may not be able to answer the issues posed with KQ 2 until we have identified which interventions are effective, with which populations, trauma types, and so forth. Once we have answered this question, the field will be in a better position to address the more nuanced questions of timing, intensity, and dosing. If the ultimate goals are to prevent PTSD, reduce the severity of PTSD, and improve quality of life, we are not there yet. Issues of timing, intensity, and dosing overlap with considerations of risk, harms, and delivery of interventions in an ethical and cost-effective way. Although one cannot divorce these considerations from the preliminary question of which interventions are effective, perhaps their consideration is not so crucial at this time.

Likewise, for KQ 3, we could draw no conclusions about the impact of characteristics of traumatic exposure or subgroups defined by various personal or medical characteristics. Thus, more rigorous research is warranted with respect to the impact that demographic factors, psychiatric comorbidities, or personal risk factors might have on the effectiveness or risk of harms of interventions. This is an area of research where the VA/DoD could expand its research efforts in such a way as to benefit not only military personnel but also civilian trauma survivors. Military personnel can—and do—experience traumatic events that are not the result of direct or even indirect combat exposure. Examples even for members of the military include subgroups comprising many of the following: military police; medics and health care providers in military health care centers located outside combat zones; and victims of physical and or sexual assault, domestic violence, or even natural disasters.

More generally, other individual and personal risk factors could also be evaluated when possible. History of trauma and psychiatric comorbidity is a case in point. Such information could further advance our knowledge and understanding of how best to intervene with certain subgroups of individuals both in and outside the military culture to prevent PTSD or reduce PTSD symptom severity.

For KQ 4, the scarce data highlight the absence of any evidence basis for the potential harms of preventive interventions. In short, a full half of the intricate calculus of weighing therapeutic harms and benefits is missing. Considerations about using any of the interventions—psychological, psychological, pharmacological, or combinations—requires a thorough understanding of the possible risks of early interventions. It is imperative that investigators who conceptualize their interventions in terms of both benefits and harms identify and specify potential harms a priori and measure them adequately. Although this might be seemingly more important (or easier) for pharmaceutical interventions, for which adverse drug events may be better known, those studying psychological interventions are not immune from this expectation and ethical consideration.

Psychological first aid has gained rapid acceptance as a universal intervention for people in the acute aftermath of trauma, but no studies of psychological first aid met inclusion criteria for our review. Although psychological first aid was not designed as an intervention to reduce the incidence of PTSD, it may have beneficial or adverse effects on mental health among trauma survivors.139,140 Rigorous studies of psychological first aid should be conducted.

One of the key research gaps for studies of targeted prevention is the limited ability to identify people who are at high risk of developing PTSD shortly after they have been exposed to trauma. Some targeted prevention studies have focused on acute stress disorder as a marker of high risk for development of PTSD. However, only 50 percent of people who meet criteria for acute stress disorder go on to develop PTSD, and more than 50 percent of people who develop PTSD never meet criteria for acute stress disorder. Focusing on acute stress disorder to identify people at high risk for developing PTSD is, therefore, not a reasonable strategy for targeted prevention.141 Other studies have tried to identify people at high risk of developing PTSD by using screening instruments for PTSD symptoms. However, those instruments are unlikely to be helpful shortly after trauma when the vast majority of trauma survivors have some symptoms of PTSD, but few of whom go on to develop the disorder of PTSD.

The development of a clinical prediction rule to identify, shortly after exposure to trauma, those who will develop PTSD and the even smaller number who will develop chronic disabling PTSD, would be an enormous help to the field. The National Institutes of Health announced a Request for Application for development of a clinical prediction rule for PTSD, but the request was limited to research on existing longitudinal datasets. Prediction rules developed from existing datasets may have limited predictive power if key variables that predict PTSD were not measured in the original dataset. A recent study reported that a clinical prediction rule for PTSD had a sensitivity of 87 percent, a specificity of 65 percent, positive predictive value of 18 percent, and negative predictive value of 98 percent (in a population with a prevalence of PTSD at 12 months of 8 percent).142 This prediction rule would therefore seem to be promising for identifying people who will not develop PTSD, but less useful for identifying people who will develop PTSD. It is this latter group that needs to be identified for targeted prevention of PTSD to be effective. We recommend that additional work be devoted to developing a clinical prediction rule based on including all key variables predictive of PTSD (i.e., pretrauma factors, event characteristics, and peri-event responses).

Conclusions

Evidence supporting the efficacy of most interventions used to prevent PTSD is lacking. If available in a given setting, brief trauma-focused CBT might be the preferable choice for reducing PTSD symptom severity in people with acute stress disorder, collaborative care may be helpful for reducing PTSD symptom severity postinjury, and debriefing is not an effective prevention intervention.

Our findings highlight the inherent difficulties of conducting research on prevention interventions, which is often more challenging when conducting research on mental-health-related problems compared with medical or other health-related issues. Our body of evidence was highly limited because of the paucity of methodologically sound studies. Although disappointing, our findings underscore the need for ongoing research efforts in the field of PTSD prevention. Our findings lead us to conclude that the development of a clinical prediction algorithm to identify those who are at high risk of developing PTSD post-trauma exposure is perhaps a more crucial next step in the field of PTSD prevention, before determining which interventions are more effective than others. The ability to identify those most at risk for developing PTSD and then evaluating the effectiveness of prevention interventions in those individuals should be the focus of future clinical and research efforts.

Bookshelf ID: NBK133347

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (2.0M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...