Compassion Focused Group Therapy for Adult Female Survivors of Childhood Sexual Abuse: a Preliminary Investigation

The adverse effects of childhood sexual abuse (CSA) are often significant and enduring. It is therefore important to identify therapeutic interventions that can effectively minimize these effects. Compassion focused therapy (CFT) was originally developed for people with high levels of self-criticism and shame. It holds significant promise as an intervention for survivors of CSA, but has not yet been empirically tested. This study explored both the acceptability and preliminary efficacy of a CFT group intervention (CFT-SA) for adult female survivors of CSA. CFT-SA was developed and tested with adult female survivors of CSA, to determine if there was any change in outcome measure scores after participating in the 12-week intervention (n = 30) and at 3-month post-intervention follow-up (n = 25). Low attrition and high session attendance, in addition to positive participant feedback, suggested the program had high acceptability. Significant improvement was observed across all outcome variables from pre- to post-intervention (n = 30), which were maintained at follow-up (n = 25). Participants demonstrated increased self-compassion and self-reassurance, and reduced symptoms of post-traumatic stress, shame, and self-criticism, as well as fears of compassion, depression, anxiety, and stress, with medium to large effect sizes (d = .55 to 1.36). Across all measures, between 20 and 57% of participants demonstrated reliable change pre- to post-intervention, and 22 to 57% from pre-intervention to follow-up. This study provides preliminary support for the acceptability and potential benefits of utilizing CFT-SA as an intervention for adult female survivors of CSA and offers recommendations for future research. Trial Registration. Australian New Zealand Clinical Trials Registry, ACTRN12616001041448.

feeling, compassionate imagery, mindful attention, and sensory focusing. All competencies are developed utilizing qualities of wisdom, strength, and commitment. The process for cultivating the compassionate competencies is known as compassionate mind training (Gilbert, 2009).
Research has demonstrated CFT's effectiveness for a wide range of therapeutic issues (see Craig et al., 2020;Kirby, 2017;Leaviss and Uttley, 2015 for reviews). Importantly, CFT is effective in reducing shame and self-hating self-criticism (Au et al., 2017;Gilbert and Procter, 2006), which are common and lasting effects of CSA and perpetuating factors in ongoing post-trauma symptomatology (Feiring & Taska, 2005;Finkelhor & Browne, 1985;Whiffen & MacIntosh, 2005). Promising results have been found in studies investigating CFT as an intervention for survivors of interpersonal trauma more broadly, including a 6-week CFT individual therapy intervention with trauma survivors (n = 10), including survivors of adult sexual assault (Au et al., 2017), a 16-week CFT mixed-gender group intervention for participants (n = 8) with a history of complex trauma and had a diagnosis of a personality disorder (Lucre & Corten, 2013), and a 5-week CFT group program for adult female survivors of gender-based or intimate partner violence (Naismith et al., 2020). Across these studies, results demonstrated increased self-compassion (Au et al., 2017) and self-reassurance (Lucre & Corten, 2013), and reliable improvement in symptoms of post-traumatic stress disorder (Au et al.;Naismith et al., 2020), depression, stress, and anxiety (Lucre & Corten, 2013;Naismith et al., 2020), feelings of shame and guilt (Au et al., 2017;Lucre & Corten, 2013;Naismith et al., 2020), and perceptions of self-inadequacy (Naismith et al., 2020).
However, there has been no systematic investigation of CFT as an intervention for survivors of CSA. Such an intervention is worthy of investigation with this client group for several reasons. Sexual abuse survivors demonstrate lower levels of self-compassion and higher fears of compassion than people without an abuse history (Boykin et al., 2018;Miron et al., 2016). The experience of CSA can condition survivors not to trust care-giving and careseeking responses, and the resulting shame and self-blame may contribute to perceptions of unworthiness (Gilbert et al., 2012). This is problematic as low self-compassion is negatively associated with trauma symptoms (Karatzias et al., 2019;Maheux & Price, 2015;Thompson & Waltz, 2008), and fear of self-compassion can contribute to symptoms of post-traumatic stress disorder (PTSD) and depression in CSA survivors (Miron et al., 2016).
The current study explored the feasibility of a CFT groupbased intervention (CFT-SA) tailored specifically for adult female survivors of CSA. Bowen et al. (2009) offers a framework for assessing and conceptualizing feasibility studies, based on eight areas of focus. Of particular relevance to this study are the areas of acceptability and demand which assess the likelihood that a new program will be utilized and assessed as satisfactory, as well as adaptation and expansion which explore how existing and previously tested frameworks perform when applied to a different population and/ or in a new context. Limited-efficacy testing acknowledges the methodological limitations inherent in feasibility studies, but aims to assess if a new idea or program shows promise and justifies further efficacy testing. The specific aims of this study were to investigate if the CFT-SA program was an acceptable intervention for adult female survivors of CSA and if preliminary efficacy was demonstrated. Primary target outcomes related to self-compassion and fears of compassion, and post-traumatic stress symptoms. However, as CFT was initially developed for shame and self-criticism, which are prevalent issues for survivors of CSA, these variables were also considered primary outcomes to be investigated. Secondary outcomes were depression, anxiety, and stress.

Participants
Forty-seven women recruited from the current client group of four different sexual assault services in Queensland, Australia, consented to participate in the CFT-SA group program (McLean, 2018). Participants were invited to participate in the study if they (a) were at least 18 years of age, (b) had experienced CSA, (c) reported ongoing symptoms and impacts resulting from this experience, (d) were considered therapeutically appropriate for a group program by their counsellor and themselves, and (e) had literacy and language levels suitable for written activities, completion of measures, and group participation. Participants were excluded from the study if they (a) reported an acute sexual assault experience within the past 3 months, (b) were of current suicide risk, (c) reported untreated psychosis, and/or (d) reported active substance abuse which would impact participation.
Of the 47 women initially recruited, a total of 36 commenced the program. Participant numbers varied between the four groups and ranged from seven to 11 participants per group. Twenty-five participants attended the follow-up session which occurred 3 months after the group program finished; however, two of the participants did not complete all measures at follow-up resulting in varied sample size across measures.
Age of participants who completed the program (where provided, n = 29) ranged from 18 to 65, with a mean age of 41.07 years (SD = 13.34). The majority of participants across groups were Caucasian, with three participants identifying as indigenous, Filipino, and Salvadorian respectively. All had experienced CSA, and nine had experienced sexual abuse in both childhood and adulthood. Of the 30 participants who completed the program and measures, 76% (23 out of 30) met criteria for PTSD at intake (based on PCL-5; Weathers et al., 2013). Additionally, 73% (22 out of 30) demonstrated moderate to extremely severe symptoms of depression, 83% (25 out of 30) for anxiety, and 70% (21 out of 30) for stress at intake (based on DASS-21; Henry & Crawford, 2005). Participants varied in the number of prior counselling sessions they had attended in an effort to address these presenting concerns and other impacts relating to the experience of CSA (ranging from one to over 100 sessions, M = 29.25 sessions).

Development of Program
The CFT-SA program was developed and facilitated by the first author, who is a registered psychologist, with over 20 years clinical experience working with survivors of interpersonal trauma. Additionally, the facilitator has undertaken extensive training and supervision in CFT, including with its founder, Professor Paul Gilbert.
At the time of the current study, there was not a standard CFT group program protocol. However, as outlined above, there are six competencies associated with the first psychology of cultivating compassion (engagement with suffering), and six competencies associated with the second psychology (the action required to alleviate and prevent suffering) (Gilbert, 2015). There are also a number of key psychoeducation components as well as breathing, imagery, and mindfulness practices utilized in CFT to help develop these competencies. It is beyond the scope of this paper to discuss each of these aspects of compassionate mind training (refer to Gilbert, 2009;Lee & James, 2012 for further details). However, as summarized in Table 1, the CFT-SA program incorporated opportunities to develop each of these competencies, and utilized the key CFT concepts and practices across the 12-week program.
CFT-SA was also developed consistent with traumainformed principles of practices Stavropoulos, 2012, 2019;SAMHSA, 2014). All content and practices were considered in the context of the experience of CSA to help participants understand current threat-based responses and provide a rationale for cultivation of a compassionate mind response. Titration and individualized adaptations of practices were a key consideration, particularly in acknowledging and normalizing fears, blocks, and resistances to breath and body-focused practices in this client group (Rothschild, 2017;Treleaven, 2018).

Implementation
CFT-SA was facilitated at four separate sexual assault services. Counselling staff within each service provided participant information and consent forms to clients who met the eligibility criteria. Expression of interest forms were forwarded to the facilitator who followed up via phone with each participant to provide further information, clarify eligibility to participate, and organize a pre-intervention individual face-to-face meeting. Meetings were scheduled for within 2 weeks prior to group commencement and provided the opportunity to complete the intake and screening form, provide any further information or clarification about the group to the participant, obtain written consent, and complete pre-intervention measures.
Groups were facilitated by the first author and co-facilitated by an experienced sexual assault counsellor from each service. Sessions occurred once per week for 2 h over a period of 12 weeks, as well as a 3-month follow-up session. Each group received the same content and dosage and the co-facilitator of each group offered a fidelity check by completing a checklist of session content, practice, and process each week, consistent with the session outline.
The structure of each 2-h weekly session followed a consistent format as much as possible: (1) initial grounding and soothing rhythm breathing practice; (2) check-in, with review of previous week's homework tasks, opportunities to reinforce key elements of CFT-SA in lived experiences, and identify and work through fears, blocks, and resistances; (3) introduction to session topic (psychoeducation); (4) experiential practice linked to topic; (5) debrief of practice (with further opportunity to identify and work through fears, blocks, and resistances, and personalize practices if needed); (6) homework tasks; and (7) review and brief end-of-session check-in and reflection.
Participants were provided with manual content for the relevant session each week. Audio recordings (produced by the author) of the breathing and guided imagery exercises practiced throughout the program were provided to participants at the start of session one. Practice scripts were adapted from a range of available CFT resources (Gilbert & Choden, 2013;Kolts, 2016;Lee & James, 2012).

Compassionate Engagement and Action Scales
The Compassionate Engagement and Action Scales (CEAS; Gilbert et al., 2017) is a 39-item scale measuring compassionate competencies related to engagement with distress and suffering, and action to prevent and alleviate distress. The scale has three subscales relating to the different flows of compassion: compassion received from others, compassion provided to others, and self-compassion. Only the self-compassion subscale was used in the current study (CEAS-SC). The CEAS-SC has demonstrated good internal consistency, α = 0.90 for action, and the engagement factor broken down into a further two-item emotional sensitivity factor and 4-item engagement with suffering factor, α = 0.77 and α = 0.72 respectively (Gilbert et al., 2017). This study was interested in the total self-compassion score as a combined measure of compassionate engagement and action, and internal consistency for the current sample (n = 36) was α = 0.72; ω= 0.72 (engagement) and 0.84 (action).

Fears of Compassion Scales
The Fears of Compassion Scales (FCS; Gilbert et al., 2011) is a three-part self-report measure identifying common fears relating to offering compassion to self (13 items), receiving compassion from others (15 items), and offering compassion for others (13 items). The latter part was omitted from this study as the primary outcome was focused on fears relating to receiving compassion (from self and others) which have been found to be more strongly correlated with poorer mental health outcomes than the compassion for other subscales (Kirby et al., 2019). Gilbert et al. (2011) reported internal consistency as α = 0.85 to α = 0.92 for fears of compassion towards self, and α = 0.87 to α = 0.85 for receiving compassion from others in a sample of therapists and university students respectively. The current study demonstrated α = 0.92; ω = 0.92 for fears of self-compassion and α = 0.90; ω = 0.90 for fears of compassion from others.

PTSD Checklist for DSM-5-Civilian Version
Τhe PTSD Checklist for DSM-5-Civilian Version (PCL-5; Weathers et al., 2013) is a 20-item self-report measure assessing Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (American Psychiatric Association (APA), 2013) symptoms of post-traumatic stress disorder (PTSD). Items represent PTSD diagnostic criteria of intrusive or re-experiencing symptoms, avoidance behavior, negative alterations in cognition and mood, and alterations in arousal and reactivity. Scores are available for each diagnostic subscale as well as a total score, out of a possible 80. The total score was used in the current study as an indicator of overall post-traumatic stress symptom severity. PCL-5 has shown excellent internal consistency, α = 0.94, in both a trauma-exposed sample of college students (Blevins et al., 2015) and the current sample (Cronbach's alpha and omega).

External and Internal Shame Scale
The External and Internal Shame Scale (EISS; Ferreira et al., 2020) consists of eight items, four relating to internal shame, and four relating to external shame. Ferreira et al. (2020) reports that the EISS demonstrated good internal consistency in a Portuguese sample, α = 0.89 for global shame, α = 0.80 for external shame, and α = 0.82 for internal shame. The current study was interested in global shame scores which also demonstrated good internal consistency, α = 0.85; ω = 0.82.

Other as Shamer Scale
The Other as Shamer Scale (OAS; Allan et al., 1994;Goss et al., 1994) is a self-report measure of external shame consisting of 18 items measuring how people think others evaluate them. The scale reported high internal consistency in both a transdiagnostic clinical population, α = 0.93 (Cuppage et al., 2018), and in the current sample (Cronbach's alpha and omega).

The Forms of Self-Criticizing/Attacking & Self-Reassuring Scale
The Forms of Self-Criticizing/Attacking & Self-Reassuring Scale (FSCRS; Gilbert et al., 2004) is a 22-item scale measuring self-criticism and ability to offer self-reassurance. Three subscales provide measures of "inadequate self" (9 items), "hated self" (5 items), and "reassured self" (8 items). "Inadequate self" and "hated self" are separate indicators of self-criticism with "inadequate self" focused on internal disappointment and a sense of inadequacy in response to failure and setbacks, and "hated self" defined by anger and disgust towards the self (Gilbert et al., 2004). The scale has good consistency: α = 0.91 for "inadequate self," α = 0.87 for "hated self," and α = 0.85 for "reassured self" in a clinical population of mixed diagnoses (Baião et al., 2015), and α = 0.86, 0.82, and 0.81 and ω = 0.87, 0.83, and 0.82 respectively in the current sample.

Depression Anxiety and Stress Scale
The Depression Anxiety and Stress Scale (DASS-21; Henry & Crawford, 2005) is a shortened version of the DASS (Lovibond & Lovibond, 1995), consisting of a 21-item selfreport questionnaire with three subscales providing measures of depression (7 items), anxiety (7 items), and stress (7 items). The DASS-21 has shown good internal consistency in a clinical sample seeking treatment for a range of anxiety and mood disorders, α = 0.96 for depression, α = 0.89 for anxiety, and α = 0.93 for stress (Brown et al., 1997). In the current study, α = 0.92, 0.74, and 0.86, and ω = 0.93, 0.78, and 0.86 respectively.

Completion of Measures
Measures were completed at three time points. Pre-intervention measures were obtained within a 2-week period prior to group commencement, post-intervention measures were obtained at the end of the final (twelfth) week of the group program, and follow-up measures were obtained at the start of the 3-month follow-up session to avoid potential priming effects generated by the session itself.

Evaluation Form
Participants were asked to complete a program evaluation form at the end of the twelfth week, which allowed for general comments regarding what they found most and least valuable, recommended changes, and topics and practices they found least and most helpful, as well as how valuable they found the CFT-SA program overall, quality of the manual content, audio material, facilitation of the program, and whether or not they would recommend the program to other survivors of CSA.

Data Analyses
Preliminary efficacy was assessed via a one-way repeated measures analysis of variance (ANOVA) to determine if there was any change in outcome measure scores after participating in the 12-week CFT-SA intervention and at 3-month post-intervention follow-up. Effect sizes were also calculated. Additionally, reliable change indices (RCIs) were calculated for each participant on all outcome measures to measure change pre-to post-intervention and from pre-intervention to follow-up. Clinical change scores were also calculated for measures with available clinical cut-offs (PCL-5 and DASS-21).

Demand and Acceptability
At the time of recruitment, there were five specialized sexual assault services within the geographic region accessible to the facilitator. Four of these services agreed to offer the CFT-SA program within their service, recruit eligible participants, and offer co-facilitation resourcing. All services were able to successfully recruit an adequate number of participants to proceed with the group, with 47 participants across services expressing an initial interest in participation. An attrition rate of 11% was observed with four out of the 36 participants who ultimately commenced the group leaving prior to completion. Of the 32 who completed the group, 29 (91%) attended between 10 and 12 sessions, with 15 (47%) attending all 12. A further two participants attended nine sessions, with one participant attending seven.
Twenty-one out of 32 participants (66%) completed end of program evaluation forms. Reasons for non-completion were not provided by the remaining participants. Sixteen out of 21 (76%) rated the value of the program as 10 out of 10, and 19 (90%) said they would recommend the program to others (two participants left this question blank). Sixteen participants rated the quality of the manual 10 out of 10 and a further five participants rated it 9 out of 10. Twenty participants (95%) indicated 10 out of 10 satisfaction with facilitation. The majority of comments relating to what participants found most valuable, and about their experience more generally, indicated high levels of satisfaction. Example comments included, "Life-changing"; "transformative, easy to understand concepts that could be practically applied"; "I am able to look back and remember without completely collapsing with grief and anger"; "It has changed my life"; "A great experience, I would recommend this program to others"; "excellent course, beautiful, made me feel great and accepted"; "I really needed this program to further my journey of recovery".
What could potentially be perceived as the more challenging topics also demonstrated high levels of acceptability, with the "Shame and Compassion" and "Compassionate Response to Trauma Symptoms" modules being endorsed as the most helpful by the greatest number of participants who completed the evaluation forms (66% and 62% respectively).

Efficacy
Prior to analysis, data was examined for missing items, normality, and presence of outliers, with no anomalies identified. Data was analyzed on the complete data set available for the three time points (n = 23-25). As there were variations in participant numbers across time frames, a betweengroups ANOVA was conducted to determine if there were any significant differences at pre-intervention between participants who completed the twelve-session CFT-SA program and provided follow-up measures (n = 25; follow-up completers), participants who completed the program but did not provide follow-up measures (n = 5; group completers), and participants who did not complete the program (n = 6; non-completers). Results suggested the presence of significant differences between the groups on OAS and EISS pre-intervention scores. Examination of post hocs revealed that group completers scored significantly higher on the OAS at pre-intervention (M = 58.2, SD = 12.33) compared with follow-up completers (M = 37.32, SD = 14.07; p = 0.122). Examination of post hocs for the EISS revealed the significant omnibus result is driven by the significantly higher mean in group completers (M = 27.4, SD = 3.91) than follow-up completers (M = 21.32, SD = 6.3) (p = 0.038); however, this difference is no longer significant when corrected for familywise error with Tukey adjusted post hoc analysis (p = 0.093). As an additional check, further analysis utilizing t-tests to compare pre-and post-intervention scores for all participants who completed the 12-week program (n = 30) confirmed that the omission of participants who did not provide follow-up measures in the analysis did not change the significance of results or overall effect sizes. Furthermore, reliable and clinical change scores from pre-to post-intervention were calculated and reported on the full sample of participants who completed the group (n = 30).
ANOVA results indicated a significant time effect for all outcome variables (see Table 2). Follow-up pairwise comparisons revealed significant effects from pre-to postintervention (p < 0.001) and pre-intervention to follow-up (p < 0.05) for post-traumatic stress symptoms, self-compassion, fears of compassion, global shame, external shame, inadequate self, hated self, reassured self, depression, anxiety, and stress, but no significant effect from post-intervention to follow-up, suggesting that change was maintained over this period. Medium to large effect sizes ranging from 0.56 to 1.6 were observed for each measure using Cohen's d, whereby effect sizes of 0.2 are considered small, 0.5 medium, and 0.8 large (Cohen, 1992).

Reliable Change
Reliable change indices (RCIs) are reported in Table 3. Calculation of RCIs requires information about the test-retest reliability for each measure (Jacobson & Truax, 1991). Data from the 36 participants who completed pre-intervention measures was used for this purpose and is reported in Table 3.
Results from the RCI analyses showed that between 20 and 57% of participants demonstrated reliable improvement from pre-to post-intervention across measures. Over 50% of participants demonstrated reliable improvement on the key target outcomes of post-traumatic stress symptoms (57%), self-compassion (53%), global shame (53%), inadequate self (57%), and fears of compassion from self and others (53%). While there was a small decrease in the percentage of participants who demonstrated reliable improvement at follow-up, including for post-traumatic stress symptoms, over 50% of participants who completed measures at this timeframe demonstrated ongoing reliable improvement for global shame (54%), fears of self-compassion (54%), fears of compassion from others (50%), inadequate self (57%), and reassured self (52%). Table 3 also identifies a small minority of participants who demonstrated reliable deterioration (assessed as a change score above the reliable change index in the direction opposite to improvement) from pre-to post-intervention and follow-up. There were nine incidents of reliable deterioration across scales/subscales from pre-to post-intervention. Four (out of 30) participants demonstrated reliable deterioration on PCL-5 scores, another participant showed reliable deterioration on the "inadequate self" subscale of the FSCRS, and the anxiety and stress subscales of the DASS-21. One participant demonstrated reliable deterioration for the "hated self" subscale and another participant for the "reassured self" subscales of the FSCRS.
There were nine instances of reliable deterioration from pre-intervention to follow-up across scales/subscales. No participant demonstrated reliable deterioration on the PCL-5 at follow-up; however, two of the four participants who did demonstrate reliable deterioration on this measure   post-intervention did not attend the follow-up session and/ or provide follow-up measures. The other two participants showed improvement on this measure by follow-up albeit not within reliable positive change cut-offs. Interestingly three participants showed reliable deterioration on the measure of external shame (OAS) at follow-up, despite no indication of negative reliable change on this measure post-intervention. Two participants demonstrated reliable deterioration on fears of self-compassion, and one for fears of compassion from others. One participant showed reliable deterioration for the "inadequate self" subscale of the FSCRS, and two others for the depression and anxiety subscales of the DASS-21, all of whom had not demonstrated reliable deterioration post-intervention.

Clinically Meaningful Change
Clinical change scores are reported in Table 3. The PCL-5 has a recommended clinical cut-off for a preliminary diagnosis of PTSD of 31-33, and meaningful clinical improvement is represented by a minimum ten-point change (Weathers et al., 2013). In the current study, 23 out of 30 participants (77%) who completed the group program demonstrated scores above the clinical cut-off of 33 at pre-intervention. Thirteen of these participants (57%) demonstrated clinically meaningful change and had scores below the clinical cut-off post-intervention. An additional four participants showed clinically meaningful improvement, but remained in the clinically significant range at the end of the program. Of the 24 participants who completed follow-up measures, 18 were above the clinical cut-off at pre-intervention. Nine of these participants (50%) demonstrated clinically meaningful change and were below the clinical cut-off at follow-up. However, of those who showed clinical improvement postintervention, four returned to above the clinical cut-off at follow-up.
The DASS-21 categorizes symptom severity into normal, mild, moderate, severe, and extremely severe. In the current study, 22 of the 30 participants (73%) who completed the group demonstrated scores in the moderate to extremely severe range at pre-intervention for depression, and 25 out of 30 (83%) and 21 out of 30 (70%) scored in the moderate to Table 3 Percentage of participants who demonstrated reliable and clinical change Note. *Denominator represents number of participants with scores above clinical cut-off at intake (out of possible 30 participants who completed the program). **Denominator represents number of participants with scores above clinical cut-off at intake and completed follow-up measures extremely severe range for anxiety and stress respectively at pre-intervention. Post-intervention scores showed that 68% of participants who scored in the clinical range for depression at pre-intervention demonstrated reduced severity at the end of the group program, with 12 newly categorized as normal or mild. Sixty percent of participants who scored in the moderate to extremely severe range for anxiety and 62% of participants who scored in this range for stress at preintervention also demonstrated clinically significant change post-intervention.
Of the 22 participants who also completed follow-up measures, 17 were in the clinically significant range (moderate to extremely severe) at pre-intervention for depression, 20 for anxiety, and 16 for stress. Ten participants (58%) showed reduced severity at follow-up for depression with eight participants (47%) re-categorized to the mild to normal range.
Ten (50%) participants demonstrated clinically significant improvement at follow-up for anxiety, seven of whom fell into the mild to normal range. Two participants increased in anxiety symptom severity from post-intervention to followup. Five showed symptom reduction to the mild to normal range at follow-up. However, six participants who demonstrated clinical improvement post-intervention, returned to the clinical range for stress by follow-up.

Discussion
The current study provided an initial investigation of the acceptability and efficacy of a CFT-SA group program for adult female survivors of CSA. High levels of acceptability were demonstrated, particularly evidenced by high attendance rates and low levels of attrition. Only four participants who commenced the program withdrew prior to completion (11%). This represents a very low attrition rate compared to other group programs with similar client groups, such as 16% for a feminist-based program (Hébert and Bergeron, 2007), 31% for a cognitive analytic therapy group (Calvert et al., 2015), and 41% for a cognitive-behavior-based program (Karatzias et al., 2016).
Preliminary efficacy has also been demonstrated by the current study, with significant change pre-to post-intervention across all outcome variables, maintained at follow-up, with moderate to large effect sizes. These results compare favorably to other studies which have utilized a group therapy format with adult survivors of CSA, with significantly higher effect sizes demonstrated by the CFT-SA intervention for reducing symptoms of trauma, depression, and anxiety compared to group programs based on CBT psychoeducation (Karatzias et al., 2014), cognitive analytic therapy (Calvert et al., 2015), and CBT-based trauma-recovery and empowerment model (Karatzias et al., 2016), which all reported small effect sizes. Reliable change indices showed that more than half of the participants demonstrated reliable improvement from pre-to post-intervention on key target outcomes such as shame, self-compassion, and post-traumatic stress.
Outcomes in relation to post-traumatic stress symptoms are particularly noteworthy given the limited focus on specific trauma memories in the CFT-SA program. Best practice guidelines into the treatment of complex trauma recommend a phased approach (Cloitre et al., 2012), defined as stage (1) safety and stabilization, (2) trauma processing, and (3) integration. While this is not necessarily considered a linear process, there is agreement that stage 1 is an important pre-requisite to trauma processing and can often generate enough symptom reduction to negate the requirement for processing altogether (Rothschild, 2017).
The CFT-SA group does not facilitate stage 2 trauma processing, yet 57% of participants demonstrated reliable and clinical change on post-traumatic stress symptoms from pre-to post-intervention and 46% from pre-intervention to follow-up. These findings may indicate that for some people a reduction in post-traumatic symptoms to non-clinical levels can be achieved without specific processing of the trauma memory. While the uncontrolled nature of this study does not allow definitive conclusions regarding the contribution of CFT-SA to these outcomes, the results provide impetus for further investigation.
However, it is also acknowledged that four out of 30 participants showed reliable deterioration in symptoms of post-traumatic stress from pre-to post-intervention. It is not uncommon for a small percentage of complex trauma survivors to experience exacerbation of symptoms over the course of treatment. For example, Larsen et al. (2016) compared post-traumatic stress symptoms in women who had experienced interpersonal violence (n = 192), at various stages of intervention, across three different treatment approaches, each with varying degrees of trauma memory processing (cognitive processing therapy with and without a written trauma narrative component, and prolonged exposure). The authors found that between 14.7 and 28.6% of participants demonstrated reliable symptom deterioration at some point over the course of intervention regardless of treatment condition. The authors were unable to identify any demographic variables, aspects of the trauma, or treatment type that predicted exacerbation of symptoms, and suggested there may be various reasons to explain this outcome, including the possibility of delayed emotional engagement with the trauma and/or other negative life events and non-treatment-related factors.
The complex nature of CSA and the number of potential individual, social, and environmental variants influencing its effects can make it difficult to identify the specific factors that influence therapeutic outcomes (Hébert and Bergeron, 2007;Taylor & Harvey, 2010;Whiffen & MacIntosh, 2005).
Nevertheless, the current study provides initial and promising indications that CFT-SA may be beneficial for at least some survivors of CSA.

Limitations and Future Research
As a preliminary investigation, the current study has a number of methodological limitations. The small sample size, limited cultural diversity, and lack of control mean that the results should be interpreted with caution. Future research should endeavor to recruit larger samples, with more cultural diversity, and provide a comparison group, either waitlist or active control, as well as randomization.
Another methodological consideration is that the same method of measurement in the form of self-report questionnaires was utilized in this study. This may contribute to common method bias whereby the instruments themselves can introduce variance due to factors such as participants utilizing uniform response tendencies across measures, especially if they are similar in wording and scale properties, as well as social desirability factors, and procedural issues in the way the tests are administered (Jordan & Troth, 2020). While there is no single method to address this issue, the authors refer to Podsakoff et al. (2012) and Jordan and Troth (2020) for procedural and statistical recommendations to mitigate against this potential limitation.
Future studies should control for other non-CFT therapeutic influences as participants were referred from sexual assault services where they were already engaged in treatment, and many are likely to have continued with this individual therapeutic engagement throughout the study intervention. No systematic recording of this contact was included in the current study. However, concurrent engagement in individual counselling has been shown not to generate differential within-group effects (Hébert and Bergeron, 2007). Furthermore, despite engagement in prior counselling, 76% of participants still presented with symptoms of post-traumatic stress in the clinically significant range, as well as 73% for depression and 83% for anxiety at intake. A high percentage of these participants demonstrated clinical change post-intervention, which may suggest the group intervention is contributing to therapeutic gains in a way that prior and/or concurrent individual counselling has not been able to on its own. Either way, future studies would need to control for more general factors relating to participation in a therapeutic group program to determine the unique contribution of a CFT-oriented group.
Future studies should also test for facilitator effects. The current study utilized the same primary facilitator who was also the person who developed the program. While a different co-facilitator was used for each of the four groups, it will be important to determine if similar outcomes are achieved with different primary facilitators across groups, and who ideally are not in dual researcher/ clinician role. All facilitators would still need to be welltrained and experienced in CFT, complex trauma, and group facilitation.
Finally, future studies should consider how to maintain positive change once the group has finished. While some measures maintained consistent change from post-intervention to 3-month follow-up, other measures showed some reduction in the number of participants demonstrating reliable and/or clinical improvement over time, including key target outcomes of self-compassion, shame, and post-traumatic stress. These findings suggest that additional post-intervention therapeutic support may be required for some participants to prevent relapse. Compared with other types of trauma, survivors of CSA may require longer treatment duration (Sloan et al., 2013). For many participants, CFT-SA offers an entirely different way of relating to themselves and the impact of the trauma they have experienced, and may take some time and practice to consolidate. Without ongoing support and reminders of core principles and strategies, prior default patterns of thinking and responding may re-emerge.
Despite these limitations, it is considered that the current study has demonstrated feasibility, particularly in terms of acceptability and preliminary efficacy, and capacity for adaptation and expansion of an existing framework for a new target group. It is therefore recommended that the next stage of testing CFT-SA as an intervention for adult survivors of CSA addresses the following research questions, with consideration of the methodological issues identified above: Does participation in a CFT-SA program demonstrate significantly greater improvement on key outcome measures compared with a control group (waitlist, Treatment-As-Usual, or non-CFT group program)? If further efficacy is established in a controlled experiment, a longer follow-up period (e.g., 6 and/or 12 months) is recommended to determine if positive changes are maintained over time. If a longer treatment duration is indicated, it is recommended that the program is expanded to include ongoing group or individual support beyond the 12-week duration and re-assessed for efficacy.