Survivor‐led guidelines for conducting trauma‐informed psychological therapy assessments: Development and modified Delphi study

Abstract Background Psychological therapy assessments are a key point at which a person is accepted into a service or referred on. There is evidence of service users experiencing harm, dropping out of services and potentially experiencing poor outcomes because of inadequate assessment practices. Approaches to assessment tend to be developed by individual services, with a lack of research identifying what makes a good assessment. Methods This survivor‐led study, based in England, aimed to generate guidelines for conducting trauma‐informed psychological therapy assessments. The study was guided by a Service User Advisory Group and a Clinician Advisory Group. The study was conducted in three key stages: (i) identifying, modelling and drafting guideline content (ii) modified Delphi study and (iii) guideline finalization. Stage 1 was informed by literature reviews, qualitative research, data workshops with Advisory Groups and an expert consultation. Fifty‐nine people with relevant experiences then participated in a single‐stage modified Delphi (Stage 2). The guidelines were finalized through an analysis of Delphi open comments and a final expert consultation (Stage 3). Results The guidelines evolved through each stage of the process, and all items were deemed important by >90% of Delphi participants. The final trauma‐informed guidelines contain eight principles, including ‘focus on relationships’, ‘from systems to people’ and ‘healing environments’. Conclusions Experiential knowledge was key in generating the guidelines and conceptualizing content, with a consequent focus on areas, such as recognizing power differentials, understanding oppression as trauma and the relational aspects of assessments. Future research should focus on guideline implementation and investigate whether this impacts service user dropout, engagement with therapy, and outcomes. Patient or Public Contribution This study is an example of survivor research, with several authors, including the study lead, identifying as survivors. We consider the ways in which our identities as survivor researchers impacted the study findings.


| INTRODUCTION
In the UK National Health Service (NHS), assessments for psychological therapies-such as cognitive behavioural therapy and psychodynamic psychotherapy-are gateway points at which a decision is taken as to whether a person is accepted into service, declined or referred onwards. Despite the clear importance of the encounter, there are few evidence-based guidelines for conducting psychological therapy assessments. Instead, assessment processes are typical encounters that are typically developed by individual psychological therapy services, and led by assessors, according to the service modality and setting. Within this, approaches range from the use of highly structured symptom and risk assessment tools 1 to story-telling and meaning-making, 2 or a combination of both.
There is some evidence that service users can feel harmed by their experiences of psychological therapy assessments. 3 For instance, structured tools can feel difficult or upsetting 4,5 and overly technical assessments can result in people's stories becoming reinterpreted, pathologized 6,7 or viewed through the lens of a particular therapy modality, rather than the person's own sensemaking. Evidence suggests that assessment experiences can be so difficult that they contribute to service drop-out. 8 There is also evidence that people who are minoritized, including people from Lesbian, Gay, Bisexual and Transgender and racialized communities, are more likely to experience long-term harms as a result of therapy, 9 although there is a lack of research evidence on assessment experiences specifically. 3 Furthermore, the impact of austerity policies on mental health service provision may compound such harms by prioritizing 'demand management' 10,11 and privileging a 'production-line mentality', which undermines the human aspects of healthcare, such as practitioners' ability to convey compassion. 12 This could occur, for instance, where the focus is on extracting needed information to test eligibility, rather than creating a healing encounter between two people. Nevertheless, clear assessments can support forward therapy engagement, and partially mitigate potential iatrogenic harms-even in cases where the encounter does not result in a therapy place. 13 In contrast, clear assessments can support forward therapy engagement and partially mitigate the potential iatrogenic harms of therapy. 13 Findings from one study suggest that positive assessment experiences can be traced to positive therapy outcomes. 14 Assessments, then, are a key part of the therapeutic process, warranting careful thought and resourcing.

| Trauma and psychological therapy assessments
Systematic reviews consistently find an association between childhood trauma-such as bullying, emotional neglect and childhood sexual abuse-and adult mental health outcomes. [15][16][17] It seems appropriate, therefore, to assume that trauma can play a key role in mental distress and help-seeking, and that prior experiences of trauma may impact service users' current experiences of psychological therapy assessments.
A small number of studies have explored trauma survivors' experiences of psychological therapies, including some references to assessment processes. Rapsey et al. 4

interviewed male childhood
sexual abuse survivors about their experiences of therapy and found a range of structural barriers to accessing therapy, including the need to undergo assessments. The study also found that failures to ask about childhood sexual abuse, despite this being the primary reason for seeking therapy, contributed to people feeling that their assessors were not listening.
Asking about trauma in assessments can also cause harm. For instance, in research exploring psychological therapy services for people diagnosed with personality disorders, Crawford et al. 18 found that assessments were experienced as traumatic where trauma was discussed without adequate support being offered afterwards. In their review of the barriers and facilitators to trauma survivors using mental health services, Kantor et al. 19 found that survivors were often reluctant to enter psychotherapy for fear of re-experiencing trauma.
Qualitative research undertaken by some of the authors found that experiences of trauma can shape all aspects of psychological therapy assessments, regardless of whether trauma is disclosed. 3 Through interviews with service users, the study found that while assessments are typically a time of crisis and emotional turbulence for people seeking therapy, assessment encounters can at times be experienced as tick-box, administrative exercises that leave people feeling like a number to be processed through a system.
Where this occurs, an opportunity to create a healing encounter SWEENEY ET AL. | 2819 supportive of ongoing service engagement is missed. The research also found that assessments can compound trauma where people are desperate yet feel that the assessor, who is essentially a stranger, has the power to decide whether help is received or not, potentially reinforcing a sense of shame, worthlessness and hopelessness. These damaging experiences, coupled with the dominance of assessor-led assessment encounters, suggest the need for assessment processes that are able to meet the needs of people with trauma histories.

| Trauma-informed approaches and psychological therapy assessments
Trauma-informed approaches have arisen in part due to evidence that trauma experiences fundamentally impact survivors' worldviews, relationships and ways of engaging with services and staff. 20,21 Consequently, in a trauma-informed service, treatment or therapy is delivered in ways that ensure survivors can engage without experiencing or compounding harms. This approach has been summarized as the 'four R's': A program, organization or system that is traumainformed realizes the widespread impact of trauma and understands potential paths for recovery; recognizes the signs and symptoms of trauma in clients, families, staff and others involved with the system; and responds by fully integrating knowledge about trauma into policies, procedures and practices, and seeks to actively resist re-traumatization. 22,p.9 In fully integrating knowledge about trauma into service provision, 'choice, trustworthiness, collaboration and empowerment' must be facilitated. 23 Doing so can enable trauma survivors-who may constitute a significant majority of people seeking psychological therapy-to access and remain in services, 20 preventing drop out and revolving door services. This is particularly important given the often high 'did not attend' rates in psychological therapies 24 and the potential role of assessments within this. 8

| Survivor-led approaches to research and guideline development
While outcome measures and guidelines have historically been developed by clinicians and academics, recent years have seen growing recognition of the need to include survivor perspectives to centre the needs and priorities of service recipients. 25 Service user and survivor inclusion in research range from tokenistic consultations through to coproduction and control of the entire research process. 26 Supportive arguments include the ethical and democratic-such as the need for those who receive services to have a role in shaping them-to the ecological and epistemic-such as the unique validity of knowledge rooted in experience. [27][28][29] Survivor research-which lies at the control end of the spectrum of involvement-contends that first-person experiential knowledge is accessed through an exploration of subjective, lived experiences of phenomena. 30 Reflection on points of connection and disconnection enable a shift from the 'I' to the 'we', 31 potentially leading to a form of 'deep experiential knowledge'. 32 This can create different ways of understanding phenomena, sometimes challenging taken-for-granted knowledge and practices. Reflecting on points of dis/connection and situating these within wider institutional, societal and theoretical levels of understanding 33 can also mean that survivor research shines a light on the power relations, inequities and intersecting axes of oppression and advantage that shape our lives. 34

| Ethics and survivor research
The study was led by the first author, an experienced trauma survivor researcher (A. S.) who has undergone multiple psychological therapy assessments and was guided by two Advisory Groups, a Service User Advisory Group (SUAG) of people with direct experience of undergoing psychological therapy assessments and a Clinician Advisory Group (CAG) whose members had experiences of overseeing, delivering and undergoing psychological therapy assessments. The Advisory Groups were established specifically for this study programme and met approximately twice a year across the 6 years of the study. Advisory Groups contributed to the study design, participated in data workshops and some members chose to contribute as coauthors (K. K. and A. F.). Our approach to engaging with Advisory Groups was influenced by Shimmin et al.'s 39 account of trauma-informed, intersectional patient and public involvement, such as utilizing techniques for discursive reflection that create opportunities to consider issues of identity and marginalization. 39 The research programme was run according to the Ethics of Survivor Research outlined by Faulkner. 28 This meant that as well as prioritizing ethical principles such as clarity transparency and respect, we also worked within a participatory ethos, which aimed to challenge the power asymmetries between researcher and researched, foreground the experiential knowledge of survivors and commit to those service changes most valued by participants. We established a small ethics working group, drawn from the SUAG, to ensure that qualitative research was conducted ethically and safely. This included, for instance, developing a distress protocol, offering an advance copy of interview questions, setting up any questions relating to trauma disclosures carefully before and at interviews and reiterating the right not to respond. Ethics approval was granted by Camberwell and St Giles Research Ethics Committee (18/LO/0077).

| Study setting
The study setting was England, United Kingdom, which has publicly funded talking therapy services available via the NHS, with access partially dependent on regional availability. NHS services range from primary care counselling to secondary and tertiary psychotherapy and specialist services in tertiary sites. The United Kingdom also has a national psychological therapy service, Increasing Access to Psychological Therapies (IAPT), which primarily offers interventions based on cognitive behaviour therapy to large numbers of people. Multiple therapy modalities are also available in the private sector.

| Stage 1: Identifying, modelling and drafting guideline content
Identifying and drafting the model and content for the guidelines was a multistage process, summarized in Table 1. In Step 1, preliminary work was undertaken to review existing evidence through a systematic and narrative review. Review findings were considered in a data workshop with SUAG members (Step 2) to generate a working model of principles for good practice in trauma-informed psychological therapy assessments.
Qualitative research was then undertaken with service users and assessors who had been involved in the same psychological therapy assessment (Step 3). Participants were recruited from the third sector and the NHS (including IAPT) psychological therapy services, typically with a trauma specialism, based in a large metropolitan city. Semistructured interviews were conducted separately with service users and assessors. Thematic analysis 42,43 was conducted by the first author, a SUAG member (A. F.) and a CAG member (K. K.), all of whom had lived experience of the topic as well as qualitative expertize; the CAG member had additional experience in conducting assessments. We identified potential guideline items separately and then met to discuss the findings. A second data workshop was then held with the SUAG (Step 4) to discuss key qualitative findings and implications for guideline content.

In
Step 5, the first author reviewed findings from Steps 1-4 to familiarize with existing guideline content and further identify items for the developing model of good practice in trauma-informed psychological therapy. Potential items were extracted from all materials to date, reviewed and grouped into conceptually similar content. Evolving principles were labelled and relabelled, and items reworded and moved, until a coherent model that reflected the evidence base was achieved. Findings were then written into draft guidelines that contained several principles with (i) each principle's rationale, (ii) a series of key items (or statements) and (iii) each item's rationale.
In the final step (Step 6), seven people with expertize in (i) trauma-informed approaches, (ii) exploring intersecting experi-

| Delphi stages
Two rounds of data collection and analysis were planned. In Round 1, we anticipated that Delphi scores would inform the ranking, retention or removal of items, with changes confirmed in a second Delphi round.
T A B L E 1 Stage 1: The process to identify, model and draft guideline content

Aim
Steps Method Output Identify content for a model of trauma-informed psychological therapy assessments Step 1. Review of existing evidence a) A systematic review of service users' experiences of psychological therapy assessments. b) A narrative review of trauma-informed approaches.
Two published reviews. 3,20 Step 2. Service User Advisory Group (SUAG) data workshop A data workshop was held with SUAG members to review potential guideline content extracted from Step 1 and to arrange content into a draft model of trauma-informed talking therapy assessments.
The SUAG created an early working model of good practice principles for trauma-informed psychological therapy assessments.
Step 3. Primary qualitative research Qualitative interviews with seven service users and seven assessors who had been involved in the same psychological therapy assessment. Participants were recruited from third sector and NHS services (including IAPT), typically with a trauma specialism. Interviews were analysed thematically by the main study researcher (the first author) and members of the SUAG (A. F.) and CAG (K. K.).
Two published qualitative papers. 40,41 Step 4. Second SUAG data workshop SUAG members participated in a second data workshop to discuss key findings and implications for the guidelines.
Discussions were captured and reviewed at Step 5.
Draft guidelines for traumainformed talking therapy assessments Step 5. Assessing the evidence and refining the model The first author reviewed the work to date (systematic and literature reviews, primary qualitative research and SUAG data workshop) by labelling, relabelling and moving content to create a coherent model that reflected the evidence.
An expanded and refined model of trauma-informed psychological therapy assessments.
Step 6. First draft of guidelines The first author drafted the guidelines based on the refined model of trauma-informed psychological therapy assessments.
Draft guidelines including each principle with (i) its rationale (ii) key items and (iii) a brief explanation for each item's inclusion.
Step 7. Expert consultation to review guidelines Draft guidelines reviewed by seven people with expertize in (i) trauma-informed approaches, (ii) exploring intersecting experiences of marginalization and (iii) psychological therapy assessments. Members provided extensive written comments on the draft guidelines.
Guidelines were modified by the first author in line with feedback.

| Stage 3: Guideline finalization
The first principle-focus on relationships-had 17 items, 15 (88%) of which had a median score of 3 (indicating very important; see Table 2). The minimum value of WC for this principle was 91.5%.
For seven items, the median rating was very important with 100% WC, indicating that all participants rated the items in the principle as important or very important.
From systems to people, the second principle had 13 items of which 4 (31%) had a median rating of 3 (very important). The median score for the remaining items was 2 (important), and the minimum WC value was 94.9%.
The third principle was trauma-competent, supported assessors.
This principle had 10 items; 90% had a median score of 3 and the remaining item had a median of 2. The WC was 88.1% for one item and ranged from 93.2% to 98.3% for the remaining items.
The fourth principle, attending to environments, also contained 10 items. There were slightly more median scores of 2 (60%) than 3 (40%). The WC ranged from 89.8% to 100%. Two items had a median of 3 and a WC value of 100%. Conversely, two items had median scores of 2.
Understanding trauma, intersectionalities and oppression, the fifth principle, was larger than the preceding principles with 18 items. The final principle, postassessment support, contained two parts.
The first part of this principle had 20 items, making it the largest. The median score was 3 for 67% of items and 2 for all remaining items.
The WC value ranged from 84.7% to 100%. The second part of this principle, if therapy is not offered, contained five items, all with a median score of 3.

| Stage 3: Guideline finalization
As the median scores for items did not drop below 2 (important), and the WC values were consistently high, we did not undertake a second, confirmatory Delphi round. Instead, we assessed qualitative comments to select, drop and refine items. was further developed through reading key literature. 46 The final guidelines contain 8 principles with a total of 80 items (see Table 2, columns 4 and 5) (available at: https://www.kcl.ac.uk/ioppn/assets/ trauma-informed-assessment-guidelines.pdf).

| DISCUSSION
We engaged in a thorough and extensive survivor researcher-led process to develop good practice guidelines for trauma-informed psychological therapy assessments. The final guideline content was deemed important by >90% of Delphi participants, suggesting that the development process was successful in identifying key content.
The final guidelines contain eight principles, such as 'focus on relationships', 'from systems to people', 'healing environments' and 'postassessment support' (see Table 2, columns 4 and 5 for final principles and example items).
It was notable that the ambition to develop trauma-informed guidelines was consistently endorsed by study participants: There was felt to be a pressing need for assessment encounters and processes to be shaped by knowledge about trauma and its impacts.
Adopting the trauma-informed guidelines for psychological therapy assessments should enable survivors to engage safely with services, promoting healing from the outset. While incorporating the guidelines should not compromise the quality of services for those without trauma experiences, not incorporating them could mean that trauma survivors are unable to engage with services, 19 although further research is needed to explore this in more depth.
The survivor-led process, centring experiential knowledge, led to several shifts over the course of the study in the way that principles were conceptualized. First, rather than understanding assessments as encounters that are inevitably led and directed by assessors, the guidelines shifted towards an understanding of the inherent difficulty for trauma survivors in undergoing assessment and the subsequent need for assessments to be replaced with something more akin to an initial meeting. Our prior qualitative research 40  survivor perspectives. Future research should investigate whether giving more time and care to psychological therapy assessments through the implementation of these guidelines creates therapeutic encounters that are more conducive to healing, improve engagement and outcomes, and reduce service dropout.

| Strengths and limitations
The survivor-led process for generating items was robust and extensive, including primary research and literature reviews. This Limitations include that a modified Delphi approach was employed, which did not include a conventional first round wherein participants generated items (although there was a robust and extensive item development phase); we did not hold a second round due to the high consensus; and as many people hold dual identities, discrete Delphi groups were unachievable. The sampling method may have resulted in a self-selecting group of participants with a particular perspective on trauma-informed psychological therapy assessments.
Although having a four-point Likert scale without the option of a midpoint may have inflated consensus, we attempted to address this by using all qualitative feedback to amend the guidelines. Delphi participants rated 111 items and so scoring may not have been consistent across the exercise due to dipping concentration.
However, many participants offered free text comments throughout the consultation, suggesting that engagement remained high. Finally, the study is specific to a UK context; further research into the utility and relevance of the guidelines to other health service systems would be needed before adoption.

| CONCLUSION
Our survivor researcher-led study generated items for psychological therapy guidelines that were widely endorsed in a modified Delphi.
Centring experiential knowledge led to key shifts in the conceptualization of the guidelines' content, including a focus on the recognition of power, understanding trauma through oppression and emphasizing the relational aspects of assessments. Future research should investigate the implementation of the guidelines and any associated impacts on therapy engagement dropout and outcomes.

AUTHOR CONTRIBUTIONS
Overall study design: Angela Sweeney, Steve Gillard. Qualitative interviews data analysis, Phase 1: Angela Sweeney, Katie Kelly, Alison Faulkner. Delphi study design and conduct: Angela Sweeney and Sarah White. Writing the first draft of the manuscript: Angela Sweeney. All authors contributed to, reviewed and agreed to the final version of the manuscript and are accountable for their contributions.