Discussion of Outcome Results
Of the twenty samples included in this review, eleven involved random assignment to treatment condition—three of these were the secondary prevention studies. Of the remaining eight random assignment treatment studies, most had further limitations, such as low power to detect differences between groups, inclusion of recent victims without a control to account for natural recovery, and limited presentation of outcome data specifically for sexual assault victims. Only the Resick et al. (2002)
study had sufficient power to detect medium effect size differences between treatment conditions. Despite these limitations, much progress has been made in the last thirty years in the development and evaluation of effective treatments for sexual assault victims.
The available data suggest that several cognitive behavioral treatments are quite effective in treating PTSD, depression, and other common symptoms that sexually assaulted women are likely to experience. Notably, CPT, PE, and SIT have received the most research support. There is some evidence for benefits of CPT over PE, particularly regarding improvements in trauma-related guilt. However, both treatments appear to be effective and it would be premature to make a conclusion regarding superiority based on a single study conducted by the developers of CPT. In a study with low power for detecting differences between treatments, no significant differences were found between PE and SIT. CPT and SIT have not been directly compared. Finally, EMDR was effective in two, small-N studies. However, the benefits of EMDR beyond its exposure-related components have not been evaluated for sexual assault victims.
Other cognitive behavioral treatments not coupled into treatment packages, including cognitive restructuring, coping skills training, progressive relaxation, systematic desensitization, and assertion training have shown some treatment gains; however, the number of studies and women in each of these conditions is still limited. In addition, one psychopharmacological investigation has been conducted with sexual assault victims, but data were not presented on women's symptoms after medication usage stopped. Due to the limited data, the effectiveness of these other cognitive behavioral treatments and of pharmacological treatment need further evaluation, and if evaluated, should be compared to CPT, PE, or SIT to determine whether they are more effective than these existing treatments. Finally, supportive counseling, which probably is the most widely used treatment in rape counseling centers, offers some benefits (as seen in pre- to post-intervention improvements), but cognitive behavioral strategies appear to lead to faster and higher rates of recovery, particularly for PTSD outcomes.
Two CBT approaches for recently assaulted women have shown some promise for facilitating quicker recovery or possibly preventing symptom development. For victims within one month post-assault, Foa et al. (2006)
found that a brief CBT intervention led to faster recovery rates than supportive counseling did. A second study targeted women prior to a forensic rape exam with a focus on preventing post-assault mental health and substance abuse problems (Resnick et al., 2007b
). More studies along these lines are needed to identify the most effective ways to intervene with rape victims in the days and initial months post-rape.
The findings from this review line up with treatment recommendations for traumatized individuals or individuals with PTSD more generally. Bisson et al. (2007)
conducted a meta-analysis of treatments for chronic PTSD (symptoms for at least three months) secondary to a variety of traumas and concluded that, in general, trauma-focused treatments and EMDR led to better outcomes than stress management and that all three of these approaches were superior to other therapies, including supportive therapy, psychodynamic therapy, and hypnotherapy. These findings support the superiority of treatments that focus on the memory of a trauma event and its meaning, rather than coping skills, support, or other non-trauma-focused techniques. International Society for Traumatic Stress Studies treatment guidelines (Rothbaum, Meadows, Resick, & Foy, 2000
) designated exposure as having the most support among cognitive behavior therapies for trauma. Stress Inoculation Training was also deemed an effective treatment. The Resick et al. (2002)
study examining CPT versus PE had not yet been published; thus CPT was listed as promising, but needing more support, due to fewer published investigations.
Several studies with related populations also may provide important information for directing future treatment evaluation efforts. These studies included some sexual assault victims, but also included childhood sexual abuse survivors, physically assaulted crime victims, or victims of other types of trauma. Foa et al. (1999)
compared PE, SIT, and a combination exposure and SIT treatment in a sample of sexual and physical assault victims. All three conditions were superior to a wait list control and few between treatment differences were found. On a measure of end state functioning, PE was found to be superior, followed by SIT, then the combination treatment. Foa et al. (2005)
examined PE alone and PE with a cognitive restructuring component in a sample of women, 68.7% of whom were sexual assault victims. No added benefit was found for the cognitive restructuring component over PE alone and both treatments led to significant improvements in PTSD and depression over wait list women. Rothbaum, Astin, and Marsteller (2005)
treated a sample of adult and child rape victims with PE or EMDR and found that both conditions led to decreases in PTSD and state anxiety, with no differences between the two treatments. Finally, Taylor et al. (2003)
compared outcomes for PE, EMDR, and relaxation training (RT) in a sample of mixed trauma victims, 45% of whom had experienced a sexual assault. This study found that PE led to larger decreases in reexperiencing and avoidance symptoms than EMDR and RT, reduced avoidance symptoms more quickly than RT, and led to fewer PTSD diagnoses than RT. Taken together these results bolster the findings of this review regarding the efficacy of PE, CPT, and SIT, and suggest that future studies comparing EMDR and PE for rape victims should specifically examine reexperiencing and avoidance symptom clusters.
By focusing on sexual assault victims, this review provides specific information about post-treatment functioning and the proportion of sexually assaulted women who remain symptomatic, even if some treatment gains were made. Characteristics that may be unique to or more common in this trauma population could influence outcomes or treatment process. For example, sexual assault victims may have difficulties in intimate and sexual relationships, have concerns about being dirty or damaged related to societal ideals about female sexuality, be hesitant to disclose a trauma due to victim blaming and no independent evidence that the trauma occurred, have been assaulted by known or trusted individuals in locations to which they have ongoing exposure, be coping with forensic examination and/or ongoing legal proceedings, and experience anxiety while awaiting test results for pregnancy or sexually transmitted infections.
Treatment Non-responders and Predictors of Treatment Outcome
Despite overall symptom reductions in most studies, notable proportions of women maintained clinical levels of symptomatology at the end of treatment. Although the largest study of CPT and PE (Resick et al., 2002
) found that 15-30% of treatment completing women retained a diagnosis of PTSD and/or Major Depressive Disorder, numbers from other studies (e.g., Foa et al., 1991
) and the Resick et al. (2002)
intent-to-treat sample indicate that closer to half of women retained a diagnosis after treatment. The variability in post-treatment functioning across studies and within treatments calls for a continued focus on this aspect of treatment efficacy (as opposed to only average group change). Examining these data is integral in taking steps towards predicting for whom different treatments are most effective and determining what can be done for women who do not respond to treatment. In attempts to identify predictors of treatment outcome, Foa et al. (1991)
did not find that participant demographics, assault details, and therapy compliance ratings predicted treatment response; whereas, in a sample combining both physical and sexual assault victims, physical injury and a childhood trauma history both increased the likelihood of more severe PTSD at post-treatment (Hembree, Street, Riggs, & Foa, 2004
Frequently studies have evaluated PTSD, depression, rape-related fears, and anxiety outcomes. Additional focus on other issues that sexually assaulted women report—substance abuse, low self-esteem, suicidal ideation, relationships problems, trust, and ability to engage in new relationships—could be useful. Weaver et al. (1998)
discuss the tension between staying focused on short-term trauma treatment versus a desire to “‘fix” all of the areas that need attention” (p. 393). These authors stress the point that focusing on too many topics can derail trauma-focused work and be a type of therapeutic avoidance, yet some therapists may have concerns about not taking a more holistic approach to client recovery. Sexual assault victims are also at increased risk for revictimization (Gidycz et al., 1993
). Most revictimization research has included child sexual assault survivors. Despite numerous theories that have been put forth about what might put women with previous sexual assault experiences at greater risk (for a review, see Breitenbecher, 2001
), the extant data has not provided conclusive answers (Classen, Palesh, & Aggarwal, 2005
). Although assault responsibility lies with the perpetrator and looking for victim characteristics that lead to revictimization may be misdirected, neglecting to address risks for revictimization and safety in treatment may be a disservice to victims. Marx et al. (2001)
found that a two session intervention decreased women's risk of being raped in the following nine weeks. Replication and examination over longer follow-up is needed, however, this study offers suggestions for a revictimization prevention component, which could be delivered in ongoing treatment or in an independent, brief group framework.
Another neglected topic in the empirical treatment literature is discussion of healing, recovery, and posttraumatic growth, with the focus instead on reducing symptoms and avoiding negative outcomes. Efforts to define, quantify, and measure constructs such as meaning making and posttraumatic growth are complicated and still in an early stage (e.g., Zoellner & Maercker, 2006
), but treatment goals of reaching non-clinical levels on outcome measures may not speak to a survivor's overall level of functioning, well-being, and quality of life. One investigation of life changes following sexual assault found that women reported both positive and negative changes post-assault and that those women reporting positive changes two weeks post-assault reported lower distress one year following the assault (Frazier, Conlon, & Glazer, 2001
). More attention to growth, improved functioning in psychosocial and occupations domains, and other positive outcomes may be one avenue to improving current treatments.
Generalizability of Results: Sample Characteristics and Exclusion Criteria
Given the high rates of comorbidity, determining effective and appropriate treatments for women with comorbid trauma-related problems is an essential area for future research. Because current exposure based techniques may temporarily increase distress, they may not be appropriate for substance abusing women or if there is a risk of precipitating a relapse for prior users (Resnick & Schnicke, 1993
). Efforts to identify effective treatments for sexual assault victims with comorbid problems can build on existing joint substance abuse and PTSD interventions implemented with other populations, such as “Seeking Safety” (Najavits, 2002
). There is also support for a prolonged exposure and coping skills intervention for comorbid PTSD and cocaine dependence, which has also been used with alcohol abusers (Coffey, Schumacher, Brimo, & Brady, 2005
). Cloitre, Koenen, Cohen, and Han (2002)
have achieved promising results pairing skills training based in Dialectical Behavior Therapy and trauma-focused cognitive behavior therapy for adult survivors of child sexual abuse.
Future studies also should consider sampling from underrepresented groups and examining whether culturally sensitive modifications or awareness of culture-specific attitudes about or experiences with rape could lead to better treatments. Additionally, reporting sample details, such as prior victimization history and relationship with the assailant, may help clinicians judge the generalizability of researched interventions to their clients.
Other Important Methodological Considerations for Future Research
Some of the well-designed, recent studies in the literature do use treatment manuals and monitor treatment integrity, report follow-up data (sometimes up to one year post-treatment), use blind assessors for diagnoses, and use valid and reliable measures to assess outcomes. Recent investigations also provide more discussion of women's post-treatment functioning, including reporting of effect sizes, indices of “good end state functioning” and clinically significant change, and the number of women still meeting clinical diagnostic criteria for PTSD, depression, or other relevant disorders. Studies should continue to include these methodological strengths.
In designing future studies, several key issues must be addressed. Particularly in studies including women immediately after an assault and up to three months post-assault (“recent victims”), a control group must be employed to determine whether improvement resulting from a treatment intervention is beyond the natural symptom decline that many victims evidence in the immediate aftermath of a rape (Kilpatrick & Calhoun, 1988
). Secondly, all but one study in the literature (Resick et al., 2002
) are underpowered to detect medium effect size differences between two treatments.
Another important focus of future studies will be an effort to dismantle components that may be particularly effective for specific symptoms or that could be used in a stepped approach depending on treatment response. Similarities in treatment techniques are seen among many of the existing treatments, as well as among treatments that are only described in the literature, but have not been empirically evaluated. Rather than comparing treatments with overlapping components, an attempt to identify specific empirically-supported components or principles may provide more valuable information for therapists planning interventions with their own clients. In a recent dismantling study of CPT for violence victims, of whom 31% identified adult sexual assault as their primary trauma, Resick et al. (2008)
found that women in the cognitive therapy component had greater PTSD improvement than women in the written exposure component.
Women may come to the attention of helping professionals through a variety of means. Some women are seen immediately post-assault due to injuries or for a forensic rape exam. Other women may disclose their sexual assault after several weeks or months and visit a student counseling center, a sexual assault center, or approach a private therapist. Women may also talk to their primary physician about symptoms associated with the trauma, such as sleep problems, general anxiety, depression, or pain, without disclosing the sexual assault or even without linking their own symptoms to the event. Many women may not reveal that they were raped for years following the incident. These women may seek treatment directly related to the sexual assault or for issues that are secondary to their assault or to rape-related PTSD (e.g., divorce, decreased libido, anhedonia), possibly still without divulging the trauma. These scenarios and the various helping professionals that could be approached at these different stages (e.g., medical doctor, psychologist, sexual assault advocate, lay counselor, etc.) lead to different questions for treatment. Research is needed to inform treatment decision-making and to identify the most appropriate treatments for victims at multiple phases (i.e., immediate support, prophylactic intervention or brief acute treatments, treatments for victims evidencing chronic symptoms, and treatment for women already treated with efficacious interventions, but still showing elevated symptomatology) and presenting with a variety of problems (i.e., subclinical levels of PTSD, depression, substance abuse, severe comorbid diagnoses, etc.).
Are Clinicians Using These Researched Interventions?
Most currently supported treatments for sexual assault victims include some element of exposure. However, there is evidence that exposure based interventions are not frequently used by clinicians and lay counselors who may be most likely to treat sexual assault victims. In a large survey of doctoral-level psychologists and a smaller sample of psychologists with a specialty in cognitive behavior therapy and trauma, 83% of the main sample and 35% of the specialty sample reported treating none of their PTSD clients with exposure (Becker, Zayfert, & Anderson, 2004
). The main sample endorsed a mean of 12 contraindications for exposure therapy, and listed increases in suicidality (76%), self-injury (68%), and dropout (59%) as complications of exposure. These complications and many of the contraindications have not been supported as issues particular to exposure therapy in the research literature. Other concerns about exposure therapy have been noted, including beliefs that exposure will retraumatize the victim, will take autonomy away by “forcing” the victim to recall the trauma, will not allow the victim to recover at her own pace, and will cause decompensation (Cook, Schnurr, & Foa, 2004
). Taken together, these data highlight the limited use and knowledge of exposure therapy among doctoral level psychologists. However, it is promising that attempts to train therapists with no prior experience in exposure therapy have met with success (Cahill, Foa, Hembree, Marshall, & Nacash, 2006
For therapists to initiate use of exposure techniques, considerable support often will be necessary, including ongoing supervision and consultation. Collaborations between research institutions and sexual assault advocacy organizations and trauma therapists in the community could be an avenue for providing therapists with the necessary support to institute changes in treatment approach (see Cook et al. (2004)
for additional, thoughtful suggestions for improving dissemination of empirically supported treatments). Finally, effectiveness research is needed to examine intervention outcomes for sexual assault victims treated in community settings.