small groups for sexual assault/rape victims

Psychotherapy Volume 25/Spring 1988/Number 1

GROUP THERAPY FOR SEXUAL-ASSAULT VICTIMS

SUSAN ROTH ELLEN DYE LESLIE LEBOWITZ Duke University

The purpose of the study was to evaluate the effectiveness of one year of group psychotherapy in reducing the long-term psychological aftereffects of sexual assault. The treatment subjects were seven white, adult female victims who responded to standardized measures of psychological functioning at seven points throughout treatment. Evidence for treatment-related improvement of intrusions, fears, and depression were found. Variability in subject response, methodological weaknesses, and the change process are discussed.

The purpose of this article is to evaluate the effectiveness of one year of group psychotherapy in reducing the long-term psychological aftereffects of sexual assault. The members of the group were adult women who were victims on an average of eight years ago. The conceptualization of the ther- apy grew out of Horowitz's (1976) model of the psychological response to stress and trauma. It also relied heavily on knowledge gained from the treatment of Vietnam veterans. As background, we first discuss the Vietnam veteran treatment literature, then briefly summarize what is known about the psychological aftermath of sexual assault and its treatment.

The Treatment of Vietnam Veterans It has been recognized by mental health profes-

sionals since the nineteenth century that psycho- logical trauma can be the direct cause of psycho- pathology (Trimble, 1981). Nevertheless, the diagnostic category "post-traumatic stress disorder" was included for the the first time in DSM-III

Reprints may be ordered from Susan Roth, Department of Psychology, Duke University, Durham, NC 27706.

(APA, 1980). The decision to include posttraumatic stress disorder, or PTSD, in DSM-III was heavily influenced by clinical experiences with Vietnam veterans. A large percentage of these men returned from the war with psychopathology which they had not experienced in their prewar lives and which showed similarities across veterans (Figley, 1978). Mental health professionals working with veterans have argued on the basis of clinical experience that psychotherapeutic approaches which focus specifically on the traumatic experience of Vietnam have been considerably more effective than tra- ditional psychotherapy approaches, for example insight-oriented psychotherapies which focus on the individual's childhood experiences (Figley, 1978; Smith, 1985). Unfortunately, the effec- tiveness of such treatments is inadequately doc- umented in the research literature.

Trauma-focused psychotherapeutic approaches which have been used with Vietnam veterans in- clude hypnosis (Brende & Benedict, 1980; Leahy & Martin, 1967; Spiegel, 1981), behavior therapy (Fairbank & Keane, 1982; Fairbank et al., 1983; Keane & Kaloupek, 1982; Schindler, 1980), stress- oriented psychotherapy (Hendin & Haas, 1984), and rap groups or other group approaches (Berman et al., 1982; Blank, 1982; Brende, 1981; Figley, 1978; Scurfield et al., 1984; Smith, 1985; Walker & Nash, 1981; Williams, 1980). While there is at least suggestive evidence for the value of each of the approaches, there are some important dif- ferences in emphasis. For example, the goal of hypnosis is to recover in detail lost memories of the traumatic experience and to reexperience the repressed affect associated with the trauma. In behavior therapy, the goal is to restore the veteran's sense of control over his life and to establish actual control over symptoms (e.g., anxiety, flashbacks, nightmares). With stress-oriented psychotherapy the goal is to open up and explore in detail the content of traumatic experiences and the affect and meanings associated with them. Finally, in the case of group approaches, there is an emphasis on the resolution of the traumatic

82

Group Therapy for Sexual-Assault Victims

experience, facilitated by the presence of survivors of the same set of traumatic events.

While the work of Hendin & Haas (1984) and much of the group work is explicitly grounded in the work of Horowitz (1976), it is striking to us that even the behavioral approaches are not in- consistent with his model. In Horowitz's concep- tualization, the goal of psychotherapy is to help the patient work through the trauma experience through gradually "dosing" with the reexperience of the trauma event and its implications at man- ageable levels. To this end, control over disturbing symptoms is important, and as Parson (1984) and Scurfield (1985) point out, behavioral techniques might be particularly effective for this.

In summary, we find Horowitz's model of the psychological response to trauma and its impli- cations for treatment compelling and consistent with available evidence pertaining to the treatment of Vietnam veterans. We find the group approaches particularly appealing because of the extra ad- vantages afforded by the presence of other victims, and we are not alone in this view (see Scurfield, 1985, and Smith, 1985). Unfortunately there are to date no systematic outcome studies in the lit- erature of the efficacy of group treatments for Vietnam veterans.

The Psychological Aftermath of Sexual Assault and Its Treatment

Sexual assault, like war, is a traumatic life experience, and many women who have been as- saulted develop long-lasting symptoms across a wide range of problem areas, including symptoms of posttraumatic stress disorder (Cohen & Roth, 1987; Finkelhor, 1984; Herman, 1981; Kilpatrick, 1984). Furthermore, Cohen & Roth have argued that the trauma of sexual assault is extremely dif- ficult to assimilate and resolve, even for an adult. Since sexual assault has just recently begun re- ceiving attention as an important area of psycho- logical inquiry, even less is known about its treat- ment than in the case of Vietnam veterans. However, there is a small literature on the effec- tiveness of trauma-focused psychotherapeutic ap- proaches.

Much of the therapeutic work with adults who were sexual-assault victims has been done im- mediately or very soon after the trauma occurred. Crisis intervention is widely practiced, and, while its effectiveness has not yet been systematically evaluated, there are numerous descriptions of crisis intervention programs in the literature (e.g., Bur-

gess & Holmstrom, 1974; McCombie et al., 1976). Crisis counseling typically involves one emergency room contact with the victim, and possibly a few additional follow-up contacts (Veronen & Best, 1983). Behavioral and cognitive-behavioral treat- ments within the early aftermath period have been studied and purport some success in reducing symptoms (Frank & Duffy-Stewart, 1983a,b; Turner & Frank, 1981). However, because of expected spontaneous improvement in adult victims during the first 3-month postassault period (Kil- patrick, 1984), these effects are not necessarily treatment related. Kilpatrick has found that a brief therapeutic intervention received shortly after the assault was no more effective than infrequent, regularly scheduled contact with staff members of a research project. While authors have discussed the difficulty in engaging victims early on in in- tervention procedures that involve systematic, relatively frequent, and sustained contact with a therapist (Kilpatrick et al., 1982; Veronen & Best, 1983), the amount of distress experienced post- assault and the long-term persistence of symptoms for some victims warrants further research into early intervention strategies.

More relevant to our own work and the work with Vietnam veterans are interventions that are targeted at longer-term problems. There is evidence for the effectiveness of behavioral and cognitive- behavioral approaches in reducing assault-related symptoms in these more chronic samples (Becker & Skinner, 1983; Veronen & Kilpatrick, 1983), and there are also isolated accounts of effective assault-focused psychoanalytic psychotherapy (Schuker, 1979) and trauma-focused psychotherapy in the Horowitz tradition (Donaldson & Gardner, 1985). Finally, supportive psychotherapy groups without any stated theoretical orientation also pur- port some success (e.g., Cryer & Beutler, 1980; Herman & Shatzow, 1984). Clearly the work in this area is just beginning, and we were interested in the extent to which our conclusions about the workwithVietnamveteranswouldextendtosexual- assault victims. Specifically, we were interested in how effective a group psychotherapy intervention grounded in Horowitz's work would be in reducing chronic psychological aftereffects in adult women who were victims of sexual assault.

Method

Subjects and Procedure

The subjects were 13 white, female victims, a subset of 72 women originally recruited for a

83

Susan Roth et al.

paper-and-pencil study on the aftermath of rape (Cohen & Roth, 1987). Women were recruited for the original Cohen & Roth study through newspaper and radio advertisements and through local rape crisis centers. The advertisements, which were headlined "Understanding Rape," asked women to fill out a confidential questionnaire about their assault experience, and offered participation in a free therapy group in return. All female re- spondents who described themselves as rape victims were eligible to participate in the study. Following participation in that study, women were given the option of continuing to participate in a second study by being intensively interviewed about their assault experiences. Twenty-nine of the original 72 women chose to participate in that study, and the first nine women electing to join the psycho- therapy group comprised the original treatment sample. We did not exclude incest victims (3) or women with previous psychiatric hospitalizations (2, one of whom was an incest victim). Thirteen additional women of the 30 interviewed agreed to serve as control group subjects for the present study.

Therapy subjects were asked to fill out ques- tionnaires at seven points in time: Once for the Cohen & Roth study (2Vz months prior to the beginning of group treatment), once immediately prior to the beginning of group treatment, again after eight, 20, and 28 sessions of group treatment, again at the end of group treatment (47 sessions and one year), and finally 6 months following termination. Two undergraduate members of the original treatment sample only contracted for the first eight weeks of therapy, due to the ending of the semester. Data for these subjects were not included in the analyses, leaving seven women in the final treatment sample. Two of these seven terminated after the 28th session (fifth data point) but before the end of the group (sixth data point), but nevertheless completed all questionnaires. Fi- nally, for five of the seven therapy subjects, the interviews occurred in between the first and second data points. For the remaining two therapy subjects, the interview occurred just before the first ques- tionnaire was filled out.

Control group subjects were assessed at only the first three time periods because of a high dropout rate, a procedure which unfortunately compromised the design of the study. While all of the therapy group subjects returned their questionnaires, only six (46%) of the control group subjects returned their questionnaires through the third assessment

period. These six women comprised the final con- trol sample. For five of the six control subjects the interviews occurred in between the first and second data points. For the remaining control sub- ject, it occurred just after the second questionnaire was filled out.

Treatment

The treatment for this study consisted of a year- long psychotherapy group. Each weekly session averaged 2V2 hours and was led by two female therapists.l Because of the powerful nature of the trauma material, and the vulnerability of some of the group members, we advised concurrent in- dividual psychotherapy. Five of the members were in individual therapy at the start of the group, and the remaining two members began individual ther- apy during the course of the group.2

We viewed the integration of the rape trauma as a critical part of the psychotherapeutic process (see Horowitz, 1976). The hope was that some meaning could be given the events that was con- sistent with underlying assumptions about the self and world through a process of assimilation and accommodation. The expectation was that this process would be extremely difficult because of the inevitability of reexperiencing such feelings as rage, fear, humiliation, betrayal, and grief in the context of painful memories.

Elements of Horowitz's model were commu- nicated to the group members as part of establish- ing the therapeutic framework for their work. Also, the therapists conveyed to the women the belief that their responses to sexual abuse were within the normal range of response to severe trauma. The level of affect experienced by the women and the alternating cycles of flooding and numbness were less terrifying when placed in this normative context of response to stress. The ther- apists were extremely supportive (e.g., they were available for emergency contacts) in response to women's fears of being overwhelmed by negative affect. Finally, a sense of safety was also provided

1 The group was led by the third author along with Jill Morgan, two graduate students in clinical psychology. The group was observed and closely supervised by the first author.

2 We are attributing our effects to the group treatment rather than ongoing individual psychotherapy because the majority of our group members had been in individual therapy for some time at the start of the group. Thus, individual therapy was not a sufficient condition for change. We do not know, however, if it was a necessary part of the treatment.

84

Group Therapy for Sexual-Assault Victims

by the therapists in maintaining appropriate distance and differentiation from the victims, thus allaying the victims' fears of overwhelming others.

The group members supported each other's ef- forts to continue working therapeutically in the face of significant distress, and to find a better resolution or adaptation to difficulties. Working together with people with similar experiences pro- vided a common bond which served to normalize the experience and reduce the sense of isolation and alienation that is common. The common bond allowed for the identification of common issues, the sharing of coping methods, and the attainment of insight. The group also provided its members with a sense of hope by giving them the chance to witness people arrayed along a continuum of recovery. Finally, by listening to others' stories, members often became aware of important aspects of their own trauma that had to be resolved. One woman's affect might trigger associations to similar affect in another member, or similarities in different people's stories might facilitate the uncovering of previously repressed memories.

For the members of the group, the experience of feeling overwhelmed and then managing or profiting from it, and seeing other people also survive these periods of intense affect, was en- couraging and empowering. People gained faith in their own ability to survive and cope with periods of intense emotion, and began to lose their fear of the affects associated with the trauma. Similarly, the women's understanding that their response was a normal part of the recovery process seemed to potentiate their growing sense of strength and capability. These processes facilitated the women's willingness and ability to continue the difficult work of resolution.

The first 20 sessions of the year-long group represented the end of the second contract period (first period was after eight sessions), and a potential ending point for the group. While the group mem- bers were enthusiastic about recontracting for more treatment at this point, the members had done a significant amount of work on their trauma. Their focus was beginning to extend to issues which, although highly related to sexual abuse, were one step removed from the actual trauma itself (e.g., relationships, sexuality). The end of the second contract period was also psychologically delineated by asking members to reflect on the gains they perceived having made, and by the therapists pro- viding feedback. The data are thus considered in regard to two end points. In the first phase of the

analyses, improvement in psychological func- tioning is considered in regard to the first 20 ses- sions. In the second phase of the analyses, im- provement is considered beyond the 20th session and into a 6-month follow-up period.

Measures and Analyses

Four standardized measures of psychological functioning were given to subjects at all points in time, and were chosen by Cohen & Roth (1987) based on the following criteria: 1) relevance to what is currently known about psychological af- tereffects in sexual assault victims; 2) availability of reliability and validity data; and 3) availability of normative data for the purpose of comparison. The Modified Fear Survey (MFS) (Kilpatrick et al., 1981; Kilpatrick & Veronen, 1983; Veronen & Kilpatrick, 1980) was chosen as a measure of fears and phobias. The SCL-90-R (Derogatis, 1983; Derogatis et al., 1973), a symptom checklist with nine primary symptom dimensions, was cho- sen to measure somatization, obsessive-compul- siveness, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid idea- tion, and psychoticism. The SCL-90-R also in- cludes global indices of severity and extensity of distress. The Self-Adjustment Scale-Self-Report (SAS-SR) (Weissman & Both well, 1976; Weiss- man & Paykel, 1974; Weissman et al., 1978) was chosen to measure adjustment in seven role areas: work (including work in the home or as a student), social and leisure, extended family, marital, pa- rental, family unit, and economic functioning. The SAS-SR also includes an average or overall measure of adjustment. Finally, the Intrusion and Denial subscales of the Impact of Event Scale (IES) (Horowitz et al., 1979; Zilberg et al., 1982) were chosen to measure symptoms .of posttraumatic stress disorder.3

Analyses including control subjects could only be completed for the first three data points. Two x three (treatment by time) repeated measures anal- yses of variance were planned for the full scale and rape subscale of the MFS, the General Severity Index (GSI) of the SCL-90-R, the overall ad- justment measure of the SAS-SR, and the intrusion and denial subscales of the IES. To protect against chance findings, it was decided to not look at

3 Items were endorsed by subjects on a scale from 1-5, which is different from that used by Horowitz and colleagues (0, 1, 3, 5). Our means are thus not comparable to other available data on the IES.

85

Susan Roth et al.

SCL-90-R and SAS-SR subscales unless global measures showed significant effects. Likewise, one-way repeated measures ANOVAs for the treated subjects alone, and including the fourth time period, were planned for the same six mea- sures. This provides for an evaluation of treatment effects for the first 20 sessions. Finally, one-way repeated measures ANOVAs for the treated subjects for times 2, 4, 5, 6, and 7 were also planned for the same six measures. This last series of analyses provides for an evaluation of treatment effects beyond the 20th session and over a 6-month follow- up period.

Results

Description of Therapy and Control Subjects

Table 1 provides a description of the subjects in the current study in regard to demographics, assault characteristics, and postassault behaviors. The data were collected in the context of the Cohen & Roth (1987) study. These data are presented to draw a picture of our subjects in terms of char- acteristics relevant to the description of sexual- assault victims. Interestingly, the therapy and control subjects seem to differ on level of force used, percent of subjects assaulted by a relative, and prior sexual assaults, variables that have been related to poorer outcome (see, e.g., Cohen & Roth, 1987). In fact, on all 22 measures of psy- chological functioning (including SCL-90-R and S AS subscales) except one (economic functioning on the SAS), the direction of mean differences between therapy and control subjects indicated poorer adjustment of subjects who elected to join the psychotherapy group.4

Analyses Including Control Subjects

For the analyses including control subjects, the treatment by time interaction and the time main effect for IES Intrusions were the only two sig- nificant effects [F(2,22) = 5.39, p < .05 and F(2,22) = 23.63, p < .0001, respectively]. The means for these effects appear in Table 2. As can be seen, while both therapy and control subjects appear to have improved between the first and second questionnaires with regard to symptoms of posttraumatic stress, the decrease in intrusions was greater for control subjects. Since the second questionnaire was filled out by all subjects right around the starting date of the group psychotherapy

(latest date was after two meetings), these effects are not likely to be attributable to the treatment. It is quite plausible that these results are due to the differential impact of the interview or the in- volvement in research on therapy versus control subjects, and/or that the rate of change is slower for the more highly symptomatic therapy subjects. It is worth emphasizing, however, that no time, treatment, or treatment by time effects were found for the remaining five measures.

Change in Therapy Subjects Over 20 Sessions

While there was no indication from the above analyses of any positive effects of therapy by the eighth session, by the twentieth session therapy subjects had improved on several of the indices of psychological functioning. In the ANOVAs there were significant time effects for the full- scale MFS [F(3,18) = 4.92, p < .05], the rape subscale of the MFS [F(3,18) = 7.53,/? < .01], the overall adjustment measure of the SAS-SR [F(3,18) = 4.75, p < .05], SAS social and lei- sure functioning [F(3,18) = 6.8, p < .01], SAS economic functioning [F(3,18) = 4.64, p < .05], and the intrusion subscale of the IES [F(3,18) = 4.05, p < .05] .5 The means for these six measures across the four data points can be seen in Table 3. There was not a significant time effect for the SCL-90-R GSI, and thus analyses were not per- formed on any of the SCL-90-R subscales.

In each of the instances of a significant time effect, planned comparisons using two-tailed t tests for related means were performed. Time 1 was compared with Time 4 to evaluate overall improvement. Time 2 was compared with Time 4 to evaluate improvement over the course of the 20 group sessions. Finally, Time 1 was compared with Time 2 to evaluate nontreatment-related im- provement.

While the pattern of means is similar in five out of six cases represented in Table 3, the strongest case can be made for positive effects of the psy- chotherapy experience in regard to rape fears, social and leisure adjustment, and IES intrusions. In each case, there was significant (p < .05) improvement from Time 1 to Time 4 [t(6) = 6.28, 3.3, and 2.52, respectively], and either a marginally significant (p < .10) or significant (p < .05) improvement from Time 2 to Time 4 [f(6) = 2.19

5 For the SAS work and family subscales, data were available for only five subjects. For the SAS marital and parental sub-

4 These data are available from the authors upon request. scales, there were not enough subjects to perform the analysis.

86

TABLE 1. Demographics, Assault Characteristics, and Postassault Behavior of Subjects

Variable Therapy Subjects Control Subjects

Age Marital status

Never married Married Divorced Separated

Socioeconomic status (SES) 1 (highest) 2 3 4 5 (lowest)

Age at assault Years since assault Type of assault (circled all that applied)

Intercourse Attempted intercourse Other sexual act

Level of force used (circled all that applied) Verbal threats Restraint Violence Had weapon Used weapon Other

Relation to assailant Stranger Seen before Date/friend Boyfriend Ex-boyfriend /ex-husband Husband Relative

Prior sexual assaults Yes No

Age at prior assaults Childhood Adulthood Both

Reported to police Yes No

Anyone told about assault Yes No

When others were told Less than 1 hr. later 1 h r . - l day 2 - 7 days 8-30 days 1 month-1 year More than 1 year

Sought professional help (emergency room, rape crisis center, gynecologist or family doctor, psychiatrist, or psychologist) Yes No

29.7 (4.06)

29% 29% 29%

28.6 (6.78)

33% 33% 33%

14%

14% 67% 33%

Note: Numbers indicate means and standard deviations (in parentheses) or percent of sample (« = 7 for therapy subjects, 6 for control subjects). SES was evaluated using the HoUingshead two-factor index (Myers & Bean, 1968).

14% 0%

0% 0% 29% 17% 71% 50% 0% 33% 0% 0%

21.6(8.47) 20.2(5.59) 8.1(5.72) 8.4(5.92)

71% 83% 14% 17% 43% 0%

86% 50% 71% 67% 57% 33% 43% 17% 14% 0% 0% 33%

43% 33% 14% 33% 0% 17% 0% 0% 0% 0% 0% 0%

43% 17%

57% 33% 43% 67%

50% 50% 0% 0%

50% 50%

43% 50% 57% 50%

86% 100% 14% 0%

71% 17% 0% 17% 0% 50% 0% 0% 0% 0% 29% 17%

Susan Roth et al.

TABLE Therapy

Subjects

Therapy Control

2. Means and Standard Deviations on the IES Intrusion Subscale for (« = 7) and Control (n = 6) Subjects for the First Three Data Points

Time 1

2'/2 Months Prior to Group

30.29 (5.44) 25.57 (3.29)

Time 2

Group Starts

24.43 (7.01) 15.33 (7.09)

Time 3

After 8 Sessions

26.29 (6.43) 16.67 (4.78)

(marginal), 2.19 (marginal) and 2.65 (significant), respectively]. Regarding these three measures, nontreatment related improvement (Time 1 vs. Time 2) was marginally evident for rape fears [t(6) = 2.34, p < .10].

For the remaining three cases represented in Table 3, there is no indication of positive effects of the psychotherapy experience. For the full- scale MFS, while there is significant improvement from Time 1 to Time 4 [r(6) = 4.05, p < .01], there is no improvement from Time 2 to Time 4, and marginally significant improvement from Time 1 to Time 2 [t(6) = 2.14, p < .10], indicating a nontreatment related effect that was maintained. For the overall SAS score, none of the planned comparisons were significant. Post hoc compar- isons indicated a significant decrease in symptoms from Time 3 to Time 4 [/(6) = 7.12, p < .001], turning around a potentially deleterious treatment outcome. For overall SAS, the three control group means for Times 1,2, and 3 are 1.97, 1.80, and 2.01, respectively, and show no evidence of an increase in symptoms. Finally, for SAS economic, none of the planned comparisons was signifi- cant. Post hoc comparisons indicated a significant increase in symptoms from Time 1 to Time 3 [t(6) = —2.63, p < .05]. Control group means

for Times 1, 2, and 3 are 2.50, 2.14, and 1.67, respectively, and again show no evidence of an increase in symptoms even though the interaction was not significant in the 2 (treatment) X 3 (time) analysis.

Change in Therapy Subjects through 6-Month Follow-Up

The above analyses indicated that by the 20th session, therapy subjects had shown treatment- related improvement on rape fears, social and leisure adjustment, and intrusive symptoms. In the case of rape fears and intrusive symptoms, the gains were maintained through the follow-up period. Gains on additional indices were also evi- denced beyond the twentieth session. In the anal- yses of variance for times 2, 4, 5, 6, and 7, there were significant time effects for the full-scale MFS [F(4,24) = 2.97, p < .05], the rape subscale of the MFS [F(4,24) = 3.43, p < .05], SCL-90- R depression [F(4,24) = 4.90, p < .01], and the intrusion subscale of the IES [F(4,24) = 3.16, p < .05]. There was a marginally significant effect for SCL-90-R GSI [F(4,24) = 2.57, p < .10]. The means for these five measures across the five data points can be seen in Table 4. There was not a significant time effect for the overall adjustment

TABLE 3. Means and Standard Deviations for Therapy Subjects (rc = 7) for Four Data Points for Measures of Psychological Functioning with Significant Time Effects

Measure

MFS All fears Rape fears SAS-SR Overall Social and leisure Economic IES Intrusions

Time 1

2Vi Months Prior to Group

300.71 (63.41) 117.29(22.29)

2.18 (.40) 2.29 (.50) 1.57 (1.05)

30.29 (5.44)

Time 2

Group Starts

267.71 (74.53) 105.00 (27.08)

2.16 (.42) 2.11 (.67) 1.43 (1.05)

24.43 (7.01)

Time 3

After 8 Sessions

273.71 (67.03) 108.43 (29.49)

2.41 (.41) 2.53 (.43) 2.57 (1.59)

26.29 (6.43)

Time 4

After 20 Sessions

249.57 (64.98) 95.43 (22.83)

1.84 (.30) 1.63 (.54) 1.86(1.35)

21.60(6.74)

88

Group Therapy for Sexual-Assault Victims

TABLE 4. Means and Standard Deviations for Therapy Subjects (n = 7) for Five Data Points for Measures of Psychological Functioning with Significant Time Effects

Measure

MFS All fears Rape fears SCL-90-R GSI Depression IES Intrusions

Time 2

Group Starts

267.71 (74.53) 105.00 (27.08)

1.17 (.69) 1.85 (1.18)

24.43 (7.01)

Time 4

After 20 Sessions

249.57 (64.98) 95.43 (22.83)

1.04 (.50) 1.49 (.79)

21.60(6.74)

Time 5

After 28 Sessions

234.57 (64.48) 87.29 (24.03)

.60 (.41) 1.05 (.87)

18.43 (5.60)

Time 6

End of Group after 47 Sessions and One Year

234.71 (66.69) 87.57 (24.23)

.84 (.65) 1.18 (.97)

20.71 (7.30)

Time 7

Follow-up

221.29(65.88) 83.29 (24.43)

.76 (.80)

.88 (.83)

17.00 (7.80)

measure of the SAS-SR, and thus analyses were not performed on any of the SAS subscales except social and leisure adjustment. While subjects had shown improvement by the twentieth session on social and leisure adjustment, the analysis of var- iance for times 2,4, 5, 6, and 7 was not significant.

In each of the instances of a significant time effect, planned comparisons using two-tailed t tests for related means were performed. Time 2 was compared with times 5 , 6 , and 7 to evaluate improvement beyond the twentieth session and into the follow-up period. Likewise, Time 4 com- pared with times 5 , 6 , and 7. Finally, Time 6 was compared with Time 7 to evaluate change in the follow-up period.

For the full-scale MFS, there is evidence of treatment-related improvement by the end of treatment which is maintained in the follow-up period [t(6) = 2.80, p < .05 for 2 vs. 6; t(6) = 2.41, p < .10 for 2 vs. 7]. Likewise, for rape fears, gains are evidenced at the end of treatment and at follow-up, although there is no evidence for additional improvement beyond that found by Time 4 (twentieth session) [f(6) = 3.31, p < .05 for 2 vs. 6; t(6) = 2.61, p < .05 for 2 vs. 7]. For the SCL-90-R GSI, there is improvement after 28 sessions [f(6) = 2.42, p < .10 for 2 vs. 5; f(6) = 3.68, p < .05 for 4 vs. 5], but no evidence for improvement by the end of treatment or during follow-up. However for SCL-90-R depression, gains are evidenced by session 28 and maintained through follow-up [?(6) = 3.05, p < .05 for 2 vs. 5; t(6) = 3.34, p < .05 for 2 vs. 6; r(6) = 2.42, p < .10 for 2 vs. 7]. Finally, for IES intrusions, gains that were seen after 20 ses- sions are also evidenced at session 28 and beyond, with an indication for additional gain in the fol-

low-up period [f(6) = 2.23, p < .10 for 2 vs. 5; r(6) = 2.46, p < .05 for 2 vs. 6; f(6) = 2.32, p < .10 for 2 vs. 7; t(6) = 2.59, p < .05 for 6 vs. 7].

Summary and Conclusions Taking the findings as a whole, there is evidence

for nontreatment-related improvement for control subjects on IES intrusions, and for therapy subjects on rape fears and the full-scale MFS. However, whereas therapy subjects did not improve on in- trusive symptoms without treatment, they did im- prove with treatment by the twentieth session. This improvement was maintained throughout treatment, and there was evidence for additional improvement in the follow-up period. Also, rape fears continued to improve for therapy subjects with treatment, and improvements evident after 20 sessions were maintained through follow-up. In fact, by the end of treatment there was evidence for treatment-related improvement that was main- tained through follow-up on the full-scale MFS. Furthermore, for SCL-90-R depression, gains were evident by session 28 and maintained through follow-up. Finally, there is evidence for an initial deleterious effect of treatment on symptoms, and no evidence for stable change in social and leisure adjustment.

The major findings, then, are illustrated in Fig- ures 1,2, and 3. Figure 1 shows the change in IES intrusions across the seven time periods for the treatment group and also for two individual subjects. The individual subject graphs illustrate the variability among subjects which existed more generally. This variability argues against averag- ing across subjects (see Barlow et al., 1984) and for distinguishing among subjects or trauma char-

89

Susan Roth et at.

35-

3 0 -

25-

20-

15

10-

o— — Subject Da

m • Treatment Group

o——- Subject D1

T-1 T-2 T-3 T-4 21/2 Months Group After 8 After 20

Prior to Starts Sessions Sessions Group

T-5 T-6 T-7 After 28 End of Group 6-Month Sessions After 47 C O M O W _ U D

Sessions and r o M O W UP 1 Year

Figure 1. IES intrusion treatment group means and individual scores over seven time periods.

acteristics that affect responsiveness to treatment. Nevertheless, the average intrusion score over time illustrates not only a statistically significant treat- ment effect, but one of considerable magnitude as well; Intrusion scores can range from 7 to 35, and dropped 13 points from Time 1 to Time 7.

Figures 2 and 3 show the change in fear (MSF full scale score) and depression, respectively, across the seven time periods for the treatment group, with various comparison groups indicated. As can be seen in Figure 2, the posttreatment average fear level is within normal limits and has dropped 79 points (scores can range from 120-600). Fi- nally, in Figure 3, while depression drops 1.2 points (scores can range from 0-4) from a level

as high as in a psychiatric outpatient sample, it still does not reach normal limits by the end of treatment.

We have thus established in a preliminary way some positive effects of a year-long group psy- chotherapy intervention for sexual assault victims. Since our sample is small and heterogeneous, and subject variability on our measures is high, our results certainly require replication. Also, while we have ruled out as many alternative explanations as possible (see Barlow et al., 1984; Kazdin, 1981), our conclusions rest on analyses that do not include control-group comparisons. We are relying on informal observations of group process in attrib- uting change to our treatment (see our treatment

90

Group Therapy for Sexual-Assault Victims

description in Method section), and we view these informal observations as a prelude to systematically documenting the change process during group in- tervention in the future. Finally, while some effects seem established prior to the end of treatment, it is not clear what the effect would have been of ending the group earlier. For example, if victims were aware that treatment was to end after twenty- eight sessions, there might have been an increase in depression after twenty sessions. Treatment effects for depression, then, would not have been established by the end of treatment.

In concluding, we emphasize that the recovery process was clearly painful for the women in the

group, as is reflected in our data in the initial worsening of symptoms. While the group members were at times ambivalent about the treatment, they seemed to value the process of trauma res- olution enough to withstand what was perceived as necessary pain. In the words of one of the incest victims as encouragement to another mem- ber:

You feel it a little bit and then you're over it. You have to feel it [the pain of incest] a little bit to get over it somewhat. Like you have a big bag of pain and you open it up and you close that bag back and there's not as much in there. Pretty soon you're going to feel so much better because there's less to bear. It hurts even more to hold it in, I think, because it's

310-

200-

290-

280-

270-

260-

250-

240-

230

220

210-

200-

Kilpatrick 11984) Victim Sample, 6-21 Days Postrape, n = 139

Cohen and\Roth 11987) Victim Sample, Average 8 Years Postrape, n=72

• MIIIIMIIIMII IIMIIIMIIIIIimiM • IIIMMIIIIIIMIM

Kilpatrick 11984] Nonvictim-Matched ControlN^Sample, n=173

T-1 T-2 T-3 T-4 T-5 T-6 T-7 2 V2 Months Group After 8 After 20 After 28 End of Group 6-Month

Prior to Starts Sessions Sessions Sessions After 47 FOMOW-UD Group Sessions and """** u»*

1 Year

Figure 2. MFS overall means for the treatment group (n = 7) over seven time periods, and for various comparison groups.

91

Susan Roth et al.

2.4-

2.0-

1.6-

1.2

0.8

0.4

Kilpatrick 119841 Victim Sample, 6-21 \ Days Post rape, n=142; and Derogatis 119831 PsychiatricVoutpatient Sample, n=577

Cohen and Roth (1987] Victim Sample, A v e r a g e \ 8 Years Postrape, n=72

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • I M I M I I I M I

Derogatis 119831 Female Nonpatient Sample, n = 122 ••••••••••IIII•••••!••••••••••••Illllllllllllllll

T-1 T-2 T-3 T-4 T-5 T-6 T-7 21/2 Months Group After 8 After 20 After 28 End of Group 6-Month

Prior to Starts Sessions Sessions Sessions After 47 Follow-UD Group Sessions and r o i l o w U P

1 Year Figure 3. SCL-90-R depression means for the treatment group (n = 7) over seven time periods, and for various comparison groups.

just going to erupt, and then you do get overwhelmed if you don't start to practice making it just a little bit lighter.

References

American Psychological Association (1980). Diagnostic and Statistical Manual of Mental Disorders (3rd ed.). Wash- ington, D.C.: Author.

BARLOW, D. H., HAYES, S. C. & NELSON, R. O. (1984). The Scientist Practitioner: Research and Accountability in Clinical and Educational Settings. New York: Pergamon.

BECKER, J. V. & SKINNER, L. J. (1983). Assessment and treatment of rape-related sexual dysfunctions. Clinical Psy- chologist, 36, 102-105.

BERMAN, S., PRICE, S. & GUSMAN, F. (1982). An inpatient program for Vietnam combat veterans in a Veterans Admin- istration hospital. Hospital and Community Psychiatry, 33(11), 919-922.

BLANK, A. (1982). Apocalypse terminable and interminable: Operation outreach for Vietnam veterans. Hospital and Community Psychiatry, 33(11), 913-918.

BRENDE, J. (1981). Combined individual and group therapy for Vietnam veterans. Journal of Group Psychotherapy, 31, 367-368.

BRENDE, J. & BENEDICT, B. (1980). The Vietnam combat delayed stress response syndrome: Hypnotherapy of dis- sociative symptoms. American Journal of Clinical Hypnosis, 23, 34-40.

BURGESS, A. W. & HOLMSTROM, L. L. (1974). Rape: Victims of Crisis. Bowie, Md.: Robert J. Bradley.

COHEN, L. & ROTH, S. (1987). The psychological aftermath of rape: Long-term effects and individual differences in recovery. Journal of Social and Clinical Psychology, 5, 525-534.

CRYER, L. C. & BEUTLER, L. (1980). Group therapy: An alternative treatment approach for rape victims. Journal of Sex and Marital Therapy, 6(1), 40-46.

DEROGATIS, L. R. (1983). SCL-90-R Manual—II. Towson, Md.: Clinical Psychometric Research.

DEROGATIS, L. R., LIPMAN, R. S. & Covi, L. (1973). SCL- 90: An outpatient psychiatric rating scale—a preliminary report. Psychopharmacology Bulletin, 9, 13-28.

92

Group Therapy for Sexual-Assault Victims

DONALDSON, M. A. & GARDNER, R. (1985). Diagnosis and treatment of traumatic stress among women after childhood incest. In C. R. Figley (Ed.), Trauma and Its Wake (pp. 356-377). New York: Brunner/Mazel.

FAIRBANK, J. A. & KEANE, R. M. (1982). Flooding for combat- related stress disorders: Assessment of anxiety reduction across traumatic memories. Behavior Therapy, 13, 499- 510.

FAIRBANK, J. A., GROSS, R. T. & KEANE, T. M. (1983). Treatment of posttraumatic stress disorder. Behavior Mod- ification, 7(4), 557-567.

FIGLEY, C. R. (Ed.) (1978). Stress Disorders among Vietnam Veterans: Theory, Research, and Treatment. New York: Brunner/Mazel.

FINKELHOR, D. (1984). Child Sexual Abuse. New York: Free Press.

FRANK, E. & DUFFY-STEWART, B. (1983a). Treatment of de- pressed rape victims: An approach to stress-induced symptom- atology. In P. Clayton and J. Barrett (Eds.), Treatment of Depression: Old Controversies and New Approaches (pp. 309-330). New York: Raven.

FRANK, E. & DUFFY-STEWART, B. (1983fc). Treating depression in victims of rape. Clinical Psychologist, 36, 96-98.

HENDIN, H. & HAAS, A. P. (1984). Wounds of War. New York: Jason Aronson.

HERMAN, J. (1981). Father-Daughter Incest. Cambridge, Mass.: Harvard University Press.

HERMAN, J. & SHATZOW, E. (1984). Time-limited group therapy for women with a history of incest. International Journal of Group Psychotherapy, 34(4), 605-616.

HOROWITZ, M. (1976). Stress Response Syndromes. New York: Jason Aronson.

HOROWITZ, M. J., WILNER, N. & ALVAREZ, W. (1979). Impact of event scale: A measure of subjective stress. Psychosomatic Medicine, 41, 209-218.

KAZDIN, A. G. (1981). Drawing valid inferences from case studies. Journal of Consulting and Clinical Psychology, 49, 183-192.

KEANE, R. & KALOUPEK, D. (1982). Imaginal flooding in the treatment of a posttraumatic stress disorder. Journal of Con- sulting and Clinical Psychology, 50(1), 138-140.

KILPATRICK, D. G. (1984). Final report, grant number MH29602. Department of Health and Human Services, Public Health Service, National Institute of Mental Health, Bethesda, Md. (Microfilm. Available on request from author.)

KILPATRICK, D. G., RESICK, P. A. & VERONEN, L. J. (1981). Effects of a rape experience: A longitudinal study. Journal of Social Issues, 4, 105-122.

KILPATRICK, D. G. & VERONEN, L. J. (1983). Treatment for rape related problems. In L. H. Cohen, W. L. Claiborn and G. A. Specter (Eds.), Crisis Intervention. New York: Human Sciences Press.

KILPATRICK, D., VERONEN, L. & RESICK, P. (1982). Psycho- logical sequelae to rape: Assessment and treatment strategies. In D. Doleys, R. Meredith and A. Ciminero (Eds.), Be- havioral Medicine: Assessment and Treatment Strategies (pp. 473-496). New York: Plenum.

LEAHY, M. R. & MARTIN, L. C. (1967). Successful hypnotic abreaction after twenty years. British Journal of Psychiatry, 113, 383-385.

MCCOMBIE, S. L., BASSUK, E., SAVITZ, R. & P E L L , S. (1976). Development of a medical center rape crisis intervention program. American Journal of Psychiatry, 4, 418-421.

MYERS, J. K. & BEAN, L. L. (1968). A Decade Later: A Follow-Up of Social Class and Mental Illness. New York: John Wiley.

PARSON, E. R. (1984). The separation of the self: Clinical and theoretical dimensions in the treatment of Vietnam combat veterans. Journal of Contemporary Psychotherapy, 14(1), 4-55.

SCHINDLER, F. E. (1980). Treatment by systematic desensi- tization of a recurring nightmare of a real life trauma. Journal of Behavior Therapy and Experimental Psychology, 2, 53- 54.

SCHUKER, E. (1979). Psychodynamics and treatment of sexual assault victims. Journal of the American Academy of Psy- choanalysis, 4, 553-573.

SCURFIELD, R. M. (1985). Post-trauma stress assessment and treatment overview and formulations. In C. R. Figley (Ed.), Trauma and Its Wake (pp. 219-257). New York: Brunner/ Mazel.

SCURFIELD, R. M., CORKER, T., GONGLA, P. & HOUGH, R. (1984). Three post-Vietnam "rap/therapy" groups: An anal- ysis. Group, 8(4), 3-21.

SMITH, J. R. (1985). Rap groups and group therapy for Vietnam veterans. In A. S. Blank, S. M. Sonnenberg and J. Talbott (Eds.), Psychiatric Problems of Vietnam Veterans (pp. 165- 192). Washington, D.C.: American Psychiatric Press.

SPIEGEL, D. (1981). Vietnam grief work using hypnosis. American Journal of Clinical Hypnosis, 24(1), 34-40.

TRIMBLE, M. R. (1981). Post-Traumatic Neurosis. New York: John Wiley.

TURNER, S. & FRANK, E. (1981). Behavior therapy in the treatment of rape victims. In L. Michelson, M. Hersen and S. Turner (Eds.), Future Perspectives in Behavior Therapy (pp. 269-291). New York: Plenum.

VERONEN, L. & BEST, C. (1983). Assessment and treatment of rape-induced fear and anxiety. Clinical Psychologist, 36, 99-101.

VERONEN, L. & KILPATRICK, D. (1983). Stress management for rape victims. In D. Meichenbaum and M. Jaremko (Eds.), Stress Management for Rape Victims (pp. 341-374). New York: Plenum.

WALKER, J. I. & NASH, J. L. (1981). Group therapy in the treatment of Vietnam combat veterans. International Journal of Group Psychotherapy, 31, 379-389.

WEISSMAN, M. M. & BOTHWELL, S. (1976). Assessment of social adjustment by patient self-report. Archives of General Psychiatry, 33, 1111-1115.

WEISSMAN, M. M. & PAYKEL, E. S. (1974). The Depressed Woman: A Study of Social Relationships. Chicago: University of Chicago Press.

WEISSMAN, M. M., PRUSOFF, B. A., THOMPSON, W. D., HARDING, P. S. & MYERS, J. K. (1978). Social adjustment by self-report in a community sample and in psychiatric outpatients. Journal of Nervous and Mental Disease, 166, 317-326.

WILLIAMS, T. (Ed.) (1980). Post-Traumatic Stress Disorders of the Vietnam Veteran: Observations and Recommendations for the Psychological Treatment of the Veteran and His Family. Cincinnati, Ohio: Disabled American Veterans.

ZILBERG, N. J., WEISS, D. S. & HOROWITZ, M. J. (1982). Impact of event scale: A cross validation study and some empirical evidence supporting a conceptual model of stress response syndromes. Journal of Consulting and Clinical Psychology, 50, 407-414.

93