Is Sharing One’s Personal Story of Victimization Preferred? Incarcerated Women’s Perspectives on Group Treatment for Sexual Trauma

Abstract

Treatment preferences are an important part of evidence-based practice and have been shown to affect treatment outcomes. In this two-part study, incarcerated women were asked about their preferences for two versions of a trauma-focused group treatment: one that requires sharing their personal memory of sexual victimization (Sharing Required) and one that does not (Sharing Not Required). Study 1 enrolled 88 non-treatment seeking women who evaluated the treatments based on descriptions of the groups. Study 2 was a partially randomized patient preference trial with 85 treatment-seeking women who either agreed to be randomly assigned to one of the two therapy groups or declined randomization and instead were directly assigned to their preferred therapy. Participants in Study 1 evaluated the Sharing Not Required condition more favorably (p < 0.001, Cohen’s d = 0.39). However, the results were affected by PTSD symptom severity as those above the clinical cutoff for probable PTSD evaluated both group treatments more favorably than those below the cutoff (ps < 0.05, Cohen’s ds ≥ 0.46). Study 2 found no significant difference between the proportion of participants who chose Sharing Required, Sharing Not Required, or had no personal preference, and the results did not differ by PTSD symptom severity (ps ≥ 0.70). Outcomes suggest that a variety of forms of trauma-focused therapy may be acceptable to incarcerated women, including those that involve personal narration of trauma memories and those that do not.

1. Introduction

There are several evidence-based treatments and practice guidelines available for how to treat trauma-related disorders such as posttraumatic stress disorder (PTSD) (e.g., Forbes et al. 2020). Guideline-endorsed treatments are a critical pillar of evidence-based practice, which also involves considering clinical expertise and patient factors such as their strengths and values, their cultural context, and their preferences when making treatment decisions (APA Presidential Task Force on Evidence-Based Practice 2006). Patient preference is also critical to trauma-informed care, which emphasizes the importance of patient empowerment and choice (SAMHSA 2014). Indeed, a growing number of studies show that preferences affect treatment outcomes, suggesting that preferences should be carefully considered (Swift and Callahan 2009; Zoellner et al. 2019).

In a systematic review of 41 studies with trauma-exposed samples, patient preference for therapeutic approaches that involved narration of trauma memories was highlighted (Simiola et al. 2015). For example, in eight of those studies, given the choice of Prolonged Exposure (PE) or medication, participants preferred PE, with a common rationale being the importance of talking about the traumatic experience in therapy. Similar reasons were found in another study that compared preferences for five different evidence-based psychological treatments in a treatment-seeking PTSD sample (Schwartzkopff et al. 2021). Participants generally preferred trauma-focused treatments (e.g., exposure therapy, CBT, EMDR) over trauma-informed treatments (e.g., psychoeducation and coping skills; referred to as stabilization by the authors). More participants also preferred exposure therapy (33%) and CBT (27%) than EMDR (18%) and psychodynamic psychotherapy (18%), with one of the most frequently provided reasons for choosing exposure therapy being the belief that “confronting” the trauma is important (p. 6). Researchers have investigated several factors such as demographic variables, trauma history, and symptom severity that could affect individual treatment preferences, but the results are inconclusive (Schwartzkopff et al. 2021).

Studies have, however, more consistently linked treatment preferences to more favorable outcomes. In a recent meta-analysis of 23 studies using either doubly randomized preference trials (DRPTs) or fully randomized preference trials (FRPTs), there was a small but statistically significant effect of matching someone’s treatment preferences with improved treatment outcomes (Delevry and Le 2019). When analyzing outcomes from the DRPT studies (n = 6), researchers found that being randomly assigned to having a choice in stating one’s preference also had a small and significant effect on treatment outcomes, suggesting that both preference and choice matter. In one of those DRPT studies, results showed that preference and choice influenced treatment outcomes in a sample of PTSD patients (Le et al. 2018; Zoellner et al. 2019). More specifically, treatment preference affected symptom changes and quality of life outcomes from pre- to post-treatment, treatment completion, general adherence to treatment, and maintenance of treatment gains two years later. Choice had a beneficial impact on outcomes, but not as much as treatment preference did.

Few studies have looked at preferences within the context of group therapy (c.f. Kracen et al. 2013). Although researchers have found larger effects for individual trauma-focused therapy compared to group therapy (Sloan et al. 2013), there remain many reasons to offer the latter. For instance, the group context can help address common trauma-related outcomes such as issues with trust and social isolation (Sloan et al. 2012). Sharing one’s experiences with others can also help normalize these experiences (ibid). Moreover, there are settings, such as hospitals and prisons, where service demand greatly exceeds the available resources. In these cases, group therapy may be the only viable option.

Prisons may have a particular need to incorporate therapy focused on recovery from traumatic stress into their rehabilitative programs. Most incarcerated women have experienced multiple types of trauma, especially interpersonal and sexual victimization, oftentimes beginning in childhood (Karlsson and Zielinski 2020). They have multiple mental health needs, many commonly associated with trauma exposure and the sexual trauma sequalae (ibid). Studies have shown that untreated trauma-related issues such as PTSD increase the likelihood for relapse (Kubiak 2004) and criminal recidivism (Zlotnick et al. 2009) among women. Despite this, there is a lack of research on the effectiveness of evidence-based trauma-focused treatments in carceral settings (Harner et al. 2015). Moreover, no studies have investigated incarcerated populations’ preferences for different types of psychological treatments. Two recent studies explored their preferences for medications for substance-related issues (Kaplowitz et al. 2021; Russell et al. 2022), but no studies have investigated their preferences for psychological treatments broadly or trauma treatments more specifically. Investigating incarcerated populations’ treatment preferences is also important because some studies have shown that incarceration can worsen mental health (Goomany and Dickinson 2015), highlighting the need for acceptable and appropriate treatments.

To begin to address this gap, we conducted two studies that investigated treatment preferences among incarcerated women for two types of group treatments. Both treatments included some level of exposure but differed in whether group members were expected to share their personal story of sexual victimization (henceforth referred to as Sharing Required) or not (henceforth referred to as Sharing Not Required). In the first study, a general sample of incarcerated women (not treatment seeking) were asked to read and evaluate the descriptions of the two types of group treatments, making hypothetical choices about the treatment options. The second study was a partially randomized patient preference trial with a sample of treatment-seeking incarcerated women who were making actual decisions about what treatment to enroll in. Participants were given the choice of being randomly assigned to one of the two types of group treatments, but they could decline. If they declined randomization, they were asked to choose one of the two groups to join. Participants in Study 2 were recruited from a facility in a different state than Study 1.

Based on previous research suggesting a preference for trauma-focused treatments with the common rationale of needing to talk about one’s traumatic experience(s) (Schwartzkopff et al. 2021), we predicted the following:

H1. 

Participants would prefer to share their personal stories of victimization.

H2. 

PTSD symptom severity would affect treatment preferences such that those above the clinical cutoff for probable PTSD would prefer the Sharing Required group compared to those below the clinical cutoff.

2. Study 1: Perceived Preferences from a Non-Treatment-Seeking Sample

In the first study, a general sample of incarcerated women that were not treatment seeking were asked to read and evaluate the descriptions of two types of group treatments, one that requires sharing one’s personal story of sexual victimization (i.e., Sharing Required) and one that does not require sharing but includes listening to others’ stories of sexual victimization (i.e., Sharing Not Required).

2.1. Materials and Methods

2.1.1. Participants and Procedure

Participants in the first study were a non-treatment seeking sample of 94 incarcerated women from the general prison population. Six participants were excluded from the analyses based on inconsistencies in their answers that could not be resolved (e.g., reported a sexual trauma history but chose answer option “Not applicable because I have never experienced sexual violence/assault” when evaluating the group treatments), resulting in a final sample of 88 women. Most self-identified as White (94%); two women identified as Black, one as Hawaiian, and two as biracial (White and Asian-American; White and Black). Average age was 37.88 years (range: 21 to 63). Most participants (n = 82) had at least one child and 80% had previously been in therapy. Most women did not have experience of being in a group treatment similar to the ones described in the study (81 and 86% reported no previous treatment experience similar to the group descriptions). There were no differences between participants who were above and below the clinical cutoff for probable PTSD on any demographic variables.

Participants were recruited from a minimum-security prison (e.g., convicted of non-violent crimes, mainly drug-related, and placed in a facility with the lowest level of security) for adults in the state of Kentucky (USA) between September 2016 and March 2019. All participants were able to read and speak English. There were no other exclusion criteria. General announcements were made in common gathering areas in the prison as well as over the speaker system. Those interested in learning more about the research project met the research team in an announced location. All participants who came to the announced location consented to be part of the study. The study was described as being about “incarcerated women’s experiences and thoughts related to traumatic events such as unwanted sexual experiences” as well as wanting to find out about “preferences for treatments that could help someone who has experienced traumatic events.” A statement clarified that participants were not expected to have experienced a traumatic event in order to qualify for the study. Those who agreed to participate completed consent forms prior to completing the study packets. Part of the packets were instructions to read a brief statement about “unwanted sexual experiences such as sexual assault or sexual abuse” that included behaviorally specific examples and prevalence rates among women in community and in carceral settings (Appendix A). Participants were also instructed to read a description of two potential types of group treatments to address consequences of unwanted sexual experiences. Both treatment descriptions were identical in their intended purpose (i.e., to target the sequelae of sexual victimization), length of story sharing (i.e., 20–30 min), and number of words. They only differed in whether participants were expected to share their personal stories of sexual victimization (i.e., Sharing Required) or not during treatment (i.e., Sharing Not Required). The descriptions were counterbalanced between participants who answered questions about their preferences for the treatments. At the end, participants were debriefed and asked again to consent or decline participation in the study. Only participants who consented at both time points were included in the study.

2.1.2. Measures

Trauma Exposure

To assess trauma exposure, participants completed the trauma checklist from the Posttraumatic Diagnostic Scale (PDS) (Foa et al. 1997). The 12-item checklist provides descriptions for potentially traumatic events and respondents indicate whether they have experienced this event. The checklist includes three items about sexual violence (sexual assault by a family member or someone you know, sexual assault by a stranger, and sexual contact when you were younger than 18 with someone who was 5 or more years older than you). We excluded the item about imprisonment. A total score of exposure to different traumatic event types is calculated by summing the number of items respondents endorse having experienced. Total scores range from 0–11, with higher scored indicating having experienced a greater diversity of traumatic events.

A modified version of the Sexual Experiences Survey Short Form Victimization (Koss et al. 2007) was used to assess for additional sexual victimization experiences. The scale includes seven different kinds of sexual experiences described in behaviorally specific terms (e.g., “Someone had oral sex with me or made me have oral sex with them without my consent”) with five follow-up questions about the context (e.g., “by threatening to physically harm me or someone close to me”) and questions about the frequency of the experience in the past year and since age 14. The original scale also includes one global question about ever being raped. In the current study, we included the seven behaviorally specific items and the global item about rape. For each item, participants were asked to indicate whether they had this experience and whether it happened to them as a child, adult, or both. We did not include the context-based questions. Similar to the PDS, the sexual victimization items were summed together based on how many items participants endorsed, for a total score ranging between 0 (no sexual victimization) to maximum score of 8 (endorsed all sexual victimization items).

PTSD Symptoms

Participants completed the PTSD Checklist for DSM-5 (PCL-5; Weathers et al. 2013). Answer options range from 0 (not at all) to 4 (extremely) and are summed to compute a symptom severity score; higher numbers indicate more severe PTSD symptoms. A clinical cutoff score of 33 has high sensitivity (93%) and specificity (72%) (Wortmann et al. 2016). Internal consistency reliability of the PCL-5 was 0.95 in Study 1. Three participants were missing an answer on one or two items on the PCL-5. Their mean scores were created based on the items answered and then the mean was multiplied by a number of items on the full measure (Mazza et al. 2015).

Treatment Preference

After reading the descriptions of the two treatment groups, participants were asked to rate each of them in terms of interest (i.e., “how interested would you be to participating in GROUP A?”), willingness (i.e., “how willing would you be to participating in GROUP A?”), and helpfulness (i.e., “how likely do you think a group such as GROUP A is to help you and other women who have experienced sexual assault?”). Answer options ranged from 1 (Not at all interested/willing/helpful) to 5 (Extremely interested/willing/helpful). Participants could also choose an option “Not applicable because I have never experienced sexual violence/assault”. Since the three items were significantly correlated with each other (ps > 0.62), they were combined into one variable by calculating the average of the three items. Participants were also asked to respond to the following hypothetical question: “Which of the two groups would you rather participate in?”.

2.1.3. Data Analysis

Whether participants preferred to share their personal histories of victimization or not (H1) was tested using paired sample t-tests comparing participants’ evaluations of the Sharing Required and the Sharing Not Required group treatment descriptions. Whether PTSD symptom severity affects treatment preferences (H2) was tested by using independent t-tests to compare participants with probable PTSD to those without probable PTSD on their evaluations of the two group treatment descriptions. Lastly, chi square analysis was used to compare percentages of participants preferring one or the other group treatment descriptions. We had a maximum of two participants with missing data for the main analyses. All continuous variables were normally distributed.

2.2. Results

2.2.1. Trauma Exposure in Sample

All participants were trauma exposed and reported on average that they had experienced 5.35 different categories of traumatic events (range: 2–9). The most endorsed traumatic event types were interpersonal violence items: 85% childhood sexual victimization, 76% sexual assault by a family member or an acquaintance, and 73% physical assault by a family member or an acquaintance. All participants endorsed experiencing at least one form of sexual violence with an average of 5.02 different sexual victimization types (range: 1–8); 87.5% of participants endorsed the global question about being raped at some point in their lives. Women classified as having probable PTSD reported significantly more categories of traumatic events than women without probable PTSD (M = 5.77, SD = 1.79 vs. M = 4.93, SD = 1.69; t (86) = 2.27, p = 0.026); the difference between these groups on sexual assault victimizations was marginally significant (M = 5.43, SD = 1.86 vs. M = 4.61, SD = 2.06; t (86) = 1.96, p = 0.054).

2.2.2. Hypotheses Testing

Evaluations of the Group Treatments: Total Sample

Contrary to our hypothesis that participants would prefer to share their personal stories of victimization (H1), participants rated the Sharing Not Required group (M = 3.53, SD = 1.10) more favorably than the Sharing Required group (M = 3.06, SD = 1.19), t (85) = 3.64, p < 0.001, Cohen’s d = 0.39. However, there was no significant difference in the proportion of participants who indicated they would choose to participate in the Sharing Not Required group (55%) versus the Sharing Required group (45%), χ² (1, N = 87) = 0.93, p = 0.335.

Evaluations of the Group Treatments: Probable PTSD vs. Non-Probable PTSD

Half of the sample (n = 44) were at or above the clinical cutoff of 33 for probable PTSD (Wortmann et al. 2016). Although we predicted that those with probable PTSD would prefer the Sharing Required group (H2), we found that they rated both groups more favorably compared to the non-probable PTSD group (Sharing Not Required group: M = 3.77, SD = 1.15 vs. M = 3.27, SD = 1.01; t (84) = 2.14, p = 0.035, Cohen’s d = 0.46; Sharing Required group: M = 3.49, SD = 1.27 vs. M = 2.66, SD = 0.96; t (86) = 3.46, p < 0.001, Cohen’s d = 0.74). See Figure 1 for a visual representation of these results. There was no difference between the probable (50% Sharing Required and 50% Sharing Not Required) and non-probable PTSD groups (39% Sharing Required and 59% Sharing Not Required) when they were asked to choose one of the two group treatments, χ² (1, N = 87) = 0.96, p = 0.326.

To clarify the findings connected to H2, we conducted post hoc analyses to compare the evaluations of the two group descriptions separately for the probable PTSD and non-probable PTSD groups. Results from paired samples t-tests showed that the non-probable PTSD group rated the Sharing Not Required condition (M = 3.27, SD = 1.01) more favorably than the Sharing Required condition (M = 2.66, SD = 0.96), t (41) = 3.56, p < 0.001, Cohen’s d = 0.55. However, there was no difference between the ratings for the probable PTSD group (M = 3.77, SD = 1.15 vs. M = 3.49, SD = 1.27; t (43) = 1.61, p = 0.12, Cohen’s d = 0.24). See Figure 2 for a visual representation of these results.

2.3. Discussion

Contrary to our prediction that participants would prefer to share their personal stories of victimization (H1), when analyzed in aggregate, participants in Study 1 preferred the Sharing Not Required to the Sharing Required group. Both groups were rated modestly favorable: the Sharing Required group mean was around the mid-point of the scale, while the Sharing Not Required group mean was approximately one-half standard deviation units above the scale mid-point. This suggests that, on average, non-treatment seeking incarcerated women have a slightly more favorable view of trauma treatments that do not require personal disclosure in a group context.

We also predicted that participants with probable PTSD would prefer to share their personal stories of victimization in comparison to those below the clinical cutoff (H2). However, when comparing participants above and below the cutoff for probable PTSD, results showed that those with probable PTSD evaluated both Sharing Required and Sharing Not Required treatments more favorably than the non-probable PTSD group. Furthermore, the probable PTSD group showed no preference for the Sharing Not Required versus Sharing Required groups, while the non-probable PTSD group preferred the Sharing Not Required group. These findings suggest that treatment attitudes may be different in people who are suffering from clinically elevated psychiatric symptoms of traumatic stress versus people who are generally functioning well. Given the fact that these treatments are designed for people who are elevated in psychiatric symptoms, our finding that both groups were evaluated comparably among more distressed participants is noteworthy.

3. Study 2: Actual Enrollment Decisions of a Treatment-Seeking Sample

The second study was a partially randomized patient preference trial with a sample of treatment-seeking incarcerated women who were making decisions about what treatment to enroll in. As for Study 1, the options included two different group treatments, one that requires sharing one’s personal story of sexual victimization (i.e., Sharing Required) and one that does not (i.e., Sharing Not Required).

3.1. Materials and Methods

3.1.1. Participants

The second study included a sample of 95 treatment-seeking women recruited from a minimum-security facility in a different state than Study 1. Ten participants had previously received at least one version of the group treatments included in the study and were therefore excluded from these analyses. This resulted in a final sample of 85 participants who participated in 10 different groups. Most participants self-identified as White (85%; n = 72); the remainder identified as Native American (n = 6), Black (n = 4), Latina (n = 1), and other (n = 2). Average age was 33.69 years, with a range between 19 and 57. Most were the mothers of at least one child (92%) and two-thirds had been in therapy before (n = 57).

3.1.2. Group Treatments

The Sharing Required version of treatment was a full implementation of Survivors Healing from Abuse: Recovery through Exposure (SHARE) (Karlsson et al. 2015, 2020). SHARE is a brief exposure-based group treatment for incarcerated women with sexual abuse histories that was adapted from evidence-based PTSD treatments to fit the needs of incarcerated women. It consists of eight 90 min sessions and the main focus is on imaginal exposures that are conducted in the group setting, with one woman sharing her story of personal victimization at the time. Other participants listen and provide feedback after each individual exposure if the person sharing agrees to receive such feedback. Each participant completes one imaginal exposure during the course of treatment. If there is additional time, then participants discuss common themes related to sexual victimization, such as issues with trust, power and control, and generational transmission of trauma. No out of session assignments are expected. This version of the treatment has been provided at the specific facility since January 2012.

The Sharing Not Required treatment was developed for the purpose of being able to compare standard SHARE to a modified SHARE, one that included all components but without the requirement to share one’s personal story of victimization. It consists of the same number and length of sessions as well as the same psychoeducational information about sexual victimization and its sequalae. However, instead of sharing personal stories about victimization, participants listen to audio recorded imaginal exposures created by the research team. Based on previous experience running SHARE groups, the team developed scripts for five stories about sexual victimization that were recorded as an imaginal exposure session similar to those that tend to be shared in treatment. The stories are narrated by members of the research team and clinical psychology doctoral students. Participants in the Sharing Not Required condition are informed that the narratives are developed and recorded for the purpose of this study and that all voices belong to actors.

3.1.3. Procedures

Participants were recruited from a minimum-security prison in the state of Arkansas (USA) between January 2017 and October 2018. Group leaders regularly informed the prison population during mandatory gatherings at the facility about the possibility of participating in the group treatments. Potential participants were encouraged to enroll if they experienced any sexual victimization at some point in their lives and if they perceived having some psychological difficulties associated with these experiences. Members of the correctional staff also regularly recommended women with sexual abuse histories to sign up for the group treatments. All participants were able to read and speak English and had no major cognitive impairments. There were no other exclusion criteria. The two treatment groups ran simultaneously, and the group leaders (two per group) were female clinical psychology doctoral students supervised by a licensed clinical psychologist (third author and one of the developers of SHARE). Prior to the start of groups, participants met with group leaders to examine the study protocol. Women could choose to partake in the group treatments without being part of the research study. Those interested in being part of the research study were given the choice to be randomly assigned to one of the two group treatments or they could choose which one they preferred to enroll in. Participants started the group treatments approximately one week later. All participants who attended the information session about the study protocol prior to starting the group initiated treatment. No participants withdrew from the study.

3.1.4. Measures

The same measures used in Study 1 to assess for PTSD symptoms (PCL-5; Weathers et al. 2013) and trauma exposure (PDS; Foa et al. 1997) were used in Study 2. Two participants were missing one item on the PCL-5, so their mean scores were created the same way as in Study 1 (Mazza et al. 2015). Internal consistency reliability of the PCL-5 was 0.94 in Study 2.

3.1.5. Data Analysis

Participants were categorized into three groups based on their preferences: those who had no personal preference (agreed to random assignment), those who preferred Sharing Required, and those who preferred Sharing Not Required. H1, that participants would prefer to share their personal stories of victimization, was tested by comparing the percentages between these three groups using a chi square test. H2, that PTSD symptom severity would affect treatment preference, was tested by comparing the three groups in terms of PTSD symptom severity levels using a one-way analysis of variance (ANOVA) as well as by comparing the proportions of participants above the clinical cutoff for probable PTSD using a chi square test.

3.2. Results

H1, that participants would prefer to share their personal stories of victimization, was not supported. Out of the total sample of 85 women, 40% (n = 34) had a personal preference and chose Sharing Required, 26% (n = 22) had a personal preference and chose Sharing Not Required, and 34% (n = 29) did not have a personal preference (agreed to random assignment). The proportions were not significantly different, χ² (2, N = 85) = 2.57, p = 0.277).

H2, that PTSD symptom severity would impact treatment preferences, was also not supported as those with probable PTSD did not differ from those without probable PTSD in terms of their preferences (68% for Sharing Required and Sharing Not Required and 59% for the group with no personal preference were above the clinical cutoff), χ² (2, N = 85) = 0.72, p = 0.699. The groups also did not differ in PTSD symptom severity levels, F (2, 84) = 0.21, p = 979. At pre-treatment, participants’ average PCL-5 scores were similar for the three groups: Sharing Required, M = 39.29, SD = 17.95, Sharing Not Required, M = 39.55, SD = 15.22, and no personal preference, M = 38.55, SD = 21.21.

Post hoc analyses were completed in order to compare the subsamples in terms of traumatic exposure and demographic variables. On demographic variables, those who chose Sharing Required had significantly fewer children (M = 2.06; SD = 1.54) than those who did not have a personal preference (M = 2.97; SD = 1.59; t (84) = 3.32, p = 0.041), and those who chose Sharing Required were more likely to have been in therapy before compared to those who chose Sharing Not Required (76% vs. 45%; χ² (1, N = 56) = 5.60, p = 0.018).

Information about exposure to traumatic events was collected at post-treatment from those completing the treatment groups (n = 66). All treatment completers reported a history of trauma exposure with average number of events being 5.74 (range 1–11). Similarly to Study 1, the most endorsed traumatic event types were interpersonal violence items: 64% sexual assault by a family member or an acquaintance, 57% childhood sexual victimization, and 57% physical assault by a family member or an acquaintance. There were no differences in the number of traumatic events between the three groups of participants (p = 0.655).

3.3. Discussion

In Study 2, a treatment-seeking sample of incarcerated women was making actual decisions about enrolling in therapy. Contrary to our prediction that participants would prefer to share their personal stories (H1), there was no difference between the proportions of women choosing Sharing Required, Sharing Not Required, or those who did not have a preference. When provided a choice, about one-third did not have a preference and agreed to random assignment, about one-third had a preference and chose a group treatment that requires sharing one’s personal story of sexual victimization, and about one-third had a preference and chose a group treatment that does not require sharing one’s personal story but requires listening to audio-recorded stories of sexual victimization. Their treatment enrollment decisions were not affected by PTSD symptom severity as those with and without probable PTSD did not differ in their enrollment decisions (contrary to H2).

4. General Discussion

The current studies build on previous research by asking incarcerated women about their preferences for two versions of a trauma-focused group treatment: one version that requires sharing a personal memory of sexual victimization (Sharing Required) and one that does not (Sharing Not Required). Both treatment versions require participants to engage in some emotional or imaginal exposure via listening to other women’s memories of sexual victimization. The first study was a hypothetical case where a sample of non-treatment seeking incarcerated women were asked to read descriptions of the group treatments and evaluate them. The second study, conducted at a different facility in another state, was with a sample of treatment-seeking incarcerated women who made actual decisions about enrolling in one of the two group treatments.

Based on previous research suggesting a preference for treatment models that promote talking about one’s trauma (Schwartzkopff et al. 2021), we predicted that participants would prefer the Sharing Required group, which was not the case. There could be several explanations for this finding. One reason could be concerns about disclosing one’s sexual victimization history to others. Previous research shows that most people disclose their sexual assault experiences to someone at some point, more commonly to an informal source (Ullman 2023). Negative reactions from people who hear these disclosures (such as not being believed or being blamed for the assault) have a detrimental impact on future disclosures and psychological wellbeing (ibid). Therefore, people might prefer groups that do not require participants to disclose their trauma memories to others, especially to peers versus just a mental health professional. This could be related to the group modality and/or possibly a preference for having a choice about disclosing or not. Part of trauma-informed care is to create an environment where trauma survivors have autonomy and a voice, which includes being able to choose (SAMHSA 2014). In terms of the group modality possibly impacting preferences, prior research evaluating the preference people have for PTSD treatments with imaginal exposure focuses on individually administered treatments rather than group treatments (Schwartzkopff et al. 2021). Individual therapy may be perceived as a “safer” place to engage in imaginal exposure given the comparable privacy associated with this mode of intervention. It is possible that describing imaginal exposure in the context of a group setting offers potential participants more pause. Future research comparing treatment preferences with and without imaginal exposure might include the manipulation of treatment type (group vs. individual).

Another reason for the difference between our results and those of prior studies is that PTSD and anxiety symptoms are commonly maintained through avoidance (Foa et al. 2019), which could mean that an increase in anxiety related to the thought of sharing one’s story could be avoided by selecting the “not sharing” option. While this may work to decrease anxiety in the short term, avoidance maintains or worsens anxiety in the long term. Moreover, research finds that the anxiety associated with engaging in imaginal exposure dissipates once treatment begins (ibid).

There are also some potentially unique challenges in a carceral setting that could affect women’s thoughts regarding disclosure of information about sexual violence in a group setting. Trust issues are common among trauma survivors (McCann et al. 1988) and may be even more common among trauma-exposed incarcerated women given their extraordinarily high rates of interpersonal violence exposure (Karlsson and Zielinski 2020). Confidentiality is challenging in a group treatment, especially when that group treatment is inside of a prison since participants may share living spaces and have daily interactions with other group participants. Unlike group treatments administered in community settings, group treatments in residential settings such as prisons have a high likelihood of participants engaging in multiple interactions or having multiple relationships with other group members. This may create a general reluctance among potential participants to disclose personal memories of trauma.

Additionally, the results from the current study suggest that PTSD symptom severity can impact treatment preferences. In the first study with a non-treatment seeking sample, those with probable PTSD rated both treatment descriptions more favorably than people without PTSD, and all ratings were above the mid-point of the scale (interest, willingness, and helpfulness). Furthermore, people with probable PTSD rated both Sharing Required and Sharing Not Required group descriptions comparably. In the second study with a treatment-seeking sample, there were no differences in participants’ enrollment decisions based on their PTSD symptom severity. Interestingly, though, as much as PTSD symptomatology did not impact decisions made in the treatment seeking sample, previous therapy experience did. Among those who had a personal preference, participants with probable PTSD who selected to enroll in Sharing Required were more likely to have been in therapy prior to the study than those who chose to enroll in Sharing Not Required.

There are several possible explanations for these findings. Symptom severity might make one more inclined to want help and toward the belief that any sort of treatment would be helpful. People with more symptoms might also feel a greater need to share their experiences with someone (i.e., to discuss openly what happened). Symptom severity and a possible diagnosis of PTSD might also be indicative of differences in trauma history, possibly longer time since first and/or index trauma, more cumulative trauma, and/or more severe trauma experiences. It is also possible that symptom severity is indicative of a more complex symptom presentation, which was not captured by the current study. This could be a mix of symptoms commonly referred to as the sexual trauma sequalae and could include complex PTSD (Cloitre et al. 2009). Findings from Study 1 (hypothetical case) showed that the probable PTSD group reported exposure to more categories of traumatic events, but not necessarily more types of sexual victimization. More cumulative and/or chronic trauma exposure could also mean additional experience trying to cope in various ways, maybe with mixed or limited results. Most participants in Study 1 had experience with therapy but few with the kinds that were described in the study. In Study 2, where participants made actual decisions about enrolling in group treatment, PTSD symptomatology did not impact their decisions. However, those who declined randomization and selected the Sharing Required group were more likely to report previous experience with therapy, possibly indicating that even dissimilar therapy experiences could increase the desirability and/or acceptability of talking directly about their traumatic experiences. Indeed, prior experiences beginning to discuss traumatic experiences, building coping skills, or other therapeutic processes could set a positive foundation for exposure-based group therapy, even in prison—a potentially challenging setting for therapy.

4.1. Clinical Implications

Researchers have noted that there is a lack of trauma-focused treatments in carceral settings (Harner et al. 2015) despite evidence suggesting a great need, especially among women (Karlsson and Zielinski 2020). Studies have also shown that some incarcerated women experience prison as being safer than their communities and are more likely to receive access to treatment while incarcerated than when in the community (Goomany and Dickinson 2015), suggesting prisons are important contexts for offering women access to evidence-based trauma-focused treatments. Data from our team’s work over the past decade also indicate that prisons are feasible places to implement trauma-focused treatments, including exposure-based therapy groups. Indeed, we have successfully implemented and evaluated an exposure-based group treatment for incarcerated women with sexual victimization histories (Survivors Healing from Abuse: Recovery through Exposure, SHARE) via more than 50 groups and over 300 women across two minimum-security facilities to date. Result show significant declines in PTSD, depression, and anxiety symptoms with large effects from pre-to posttreatment (Karlsson et al. 2015, 2020). Treatment effects persist months after completing the treatment (Karlsson et al. 2022).

Our findings from the current study support the appropriateness of offering a variety of trauma treatments, including those that incorporate exposure. Our work also supports the utility of shared decision making, which is part of evidence-based practice (APA Presidential Task Force on Evidence-Based Practice 2006). This includes discussing the content and the purpose of the various treatments as well as the various levels of evidence for these treatments. Moreover, when offering exposure-based trauma treatments, especially if offered in groups, one should take extra care to set a therapeutic tone that is attentive to possible concerns participants may have. Our experiences with implementing and evaluating SHARE are useful to consider when providing a trauma-focused group treatment to incarcerated populations (Zielinski et al. 2016). During the first session of SHARE, participants agree on group norms and ways to interact as well as rules around confidentiality including the consequences for breaching confidentiality. The main goal is to create a sense of safety and trust that provides the basis for being able to share one’s memory of sexual victimization (ibid). Lengthy discussions around confidentiality includes determining how to balance being able to receive support from each other between sessions if preferred while maintaining confidentiality. Group members develop various strategies for this including agreeing on a code word to use when wishing to receive support. During the imaginary exposure sessions, one participant at the time shares her memory of sexual victimization while everyone else listens attentively. After sharing one’s story and processing with the group leaders, the person sharing can choose to receive feedback from the other participants. Feedback is instructed to be supportive and therefore not judgmental or advice-giving.

4.2. Limitations and Future Directions

Limitations of this study include the fact that it was conducted at two sites with limited diversity. Specifically, both sites were a minimum-security prison with a predominantly White prison population, in which most women are incarcerated due to drug-related charges. In general, the characteristics of the samples and the prison facilities impact the generalizability of the results from the current study. More research is needed with other types of samples (e.g., ethnic and racial minorities) and in other types of facilities (e.g., higher-security categorizations, other geographical regions). Additionally, our study focused on comparing preferences for two versions of a trauma-focused group treatment. Additional research should explore preferences for other types of treatments, including other types of psychological treatments such as various evidence-based models in various modalities. Future studies could also benefit from investigating the impact of treatment preferences on outcomes in samples of incarcerated women.

5. Conclusions

To our knowledge, this is the first study investigating incarcerated women’s preferences for different types of psychological treatments, in this case two versions of a trauma-focused group treatment. Both groups included listening to other women’s stories of sexual victimization, but the groups differed in whether it was required or not required to share one’s own story of sexual victimization. In a hypothetical case, with non-treatment seeking women, participants preferred the group where sharing was not required. However, results were impacted by PTSD symptom severity as those above the clinical cutoff for probable PTSD rated both treatments more favorably compared to those below the clinical cutoff. Those with probable PTSD also rated the two treatments equally favorable. In an actual case where women were making decisions about enrolling in a treatment, a similar number of participants chose one of the three options: to share one’s personal story, to not have to share one’s personal story, or to agree to be randomly assigned into one of the two (no personal preference). Their choices were not affected by PTSD symptom severity. This study highlights the importance of providing a variety of trauma-focused treatments to incarcerated women and involving the women in a process of shared decisions making when contemplating treatment options.