The effectiveness of sexual assault referral centres with regard to mental health and substance use: a national mixed-methods study - the MiMoS Study

Ethical approval process

Recruiting people who have recently experienced violence and abuse to research studies needs to be conducted sensitively and in conjunction with those with lived experience. While there is a risk of adding an additional burden during a difficult time, there is also evidence to suggest that survivors of abuse find participating in research helpful. Survivors Voices produced a charter ‘Pain into Power’, which aims to promote the engagement of survivors in research and this survivor-led document provides a useful framework to inform the design and implementation of research. 9

There were three main risks to the MiMoS study. The LEAG was mostly concerned with protecting anonymity of participants and the risks that could be posed to participants if they lived with their perpetrator. The LEAG stressed the importance of using a range of methods of communication according to the preference of each potential participant. We also implemented a code-word scheme so that if during a phone or online conversation a participant gave us the code word we would know that they were no longer able to have a private conversation. The LEAG felt that it would be helpful to have information that signposted all participants and potential participants to a range of relevant helplines and resources. This information was on a page of our study website as well as on participant information sheets. We additionally identified that participation may exacerbate emotional distress and there may be disclosures of self-harm or suicidality. To mitigate these risks, we were very clear in our written and verbal communication exactly what the study was about and the nature of the participation. We excluded anyone who was already in acute distress (as evidenced by admission to a mental health unit and/or having been in contact with mental health crisis team). All researchers underwent training prior to data collection and this was facilitated by Concetta Perot from Survivors Voices regarding how to work with people who experience trauma and we had a robust standard operating procedure (SOP), which included detailed protocol for handling any issues that arose during recruitment and data collection (see Report Supplementary Material 1).

WP3, 4 and 5 required NHS ethical and Health Research Authority (HRA) approval. We co-produced the recruitment strategy and data collection methods with the LEAG. NHS Ethics Committee Favourable opinion was obtained on 6 December 2019 (IRAS 238440 REC reference 19/NW/0663).

Eligibility criteria

There were three subsearches within the review and each one related to the main questions and as such had specific inclusion criteria. Some papers had information that addressed more than one review question.

Question 1: approaches to mental health and substance misuse

We found 78 papers that provided information on how SARCs identify and treat mental health/substance misuse problems. These were published between 1979 and 2019, with 64 being journal articles and 14 retrieved through searches of grey literature. Based on extracted data from these 78 papers, we developed a provisional service typology for three domains of service provision as shown in Figure 4.

FIGURE 4.

Approaches to mental health and substance use.

The most common approach to identification was unstructured assessment, which included professional judgement, casual observation, medical history-taking, self-reported disclosure and one-off questions about mental health. Unstructured approaches to identification were the most reported across all service models, and especially in UK SARCs, and the equivalent model in the United States. The most common approach to in-house support was the provision of supportive or unspecified counselling. That included any type of counselling provided from the sexual assault team that was not specified as a structured psychological therapy; for example, crisis counselling, supportive counselling and unspecified counselling. Finally, referral to other services was most reported to involve signposting and/or referral without active follow-up. Overall, there was a wide variation in mental health and substance misuse service provision in SARCs. The included papers had minimal information on how drug and/or alcohol was identified or addressed.

Question 2: evaluation of interventions

The review identified five studies evaluating psychological interventions for people attending SARCs. The studies were all randomised controlled trials (RCTs) published between 2003 and 2017, and set in the United States and the UK. Four of these studies evaluated brief psychoeducational video interventions compared with standard care and one study evaluated a brief six-session ‘cognitive processing therapy’ compared with standard care. The assessed outcomes were substance use and abuse, anxiety, depression and post-traumatic stress disorder (PTSD). Video interventions were shown before or after the forensic medical examination. They included information on what would happen during the examination, as well as psycho-educational material about self-management to prevent post-assault distress and substance misuse. These RCTs provided mixed evidence of moderate quality. They had significant methodological shortcomings, including small sample sizes and high attrition rates. In summary, there is a lack of robust empirical evidence to guide interventions for mental health and drug and/or alcohol conditions provision in SARCs.

Question 3: stakeholder views and policy recommendations

The review identified 34 papers (relating to 32 studies) that reported stakeholder views on how SARCs should identify and respond to mental health/substance use for people following sexual assault. Of these, 25 were peer-reviewed publications and 9 were identified from grey literature searches. They were published between 1980 and 2018; 11 studies included staff from sexual assault services as participants, 5 included service users, 2 included carers, 1 included staff from partner agencies working closely with sexual assault services, and 13 included a mix of stakeholders. The included studies consisted of surveys, interviews and quantitative questionnaires. Of the total, 13 studies were set in the United States, 12 in the UK, 4 in Canada, 2 in South Africa and 1 in each of 7 European countries.

The main recommendations included (Table 2):

• accessibility, flexibility and continuity of care

• SARCs to systematically assess for mental health and drug/alcohol needs

• provision of in-house counselling/psychological support for both adult and child service users

• clear referral pathways and closer links with mental health/substance misuse services

• SARC staff to facilitate and encourage the take-up of follow-up services

• SARC staff to receive training in assessing and managing mental health and drug/alcohol needs; how to support people who identify as lesbian, gay, bisexual, transgender, queer (LGBTQ+) plus pansexual, asexual, non-binary, and intersex LGBTQ+ and people with learning difficulties.

A total of 10 policy documents from government and expert bodies published between 2004 and 2018 were retrieved, including 6 UK government documents, 3 European policy documents reporting on the European Union-funded Comparing Sexual Assault Interventions (COSAI) project, and 1 research project commissioned by the UK Home Office. 7 The main recommendation that consistently emerged from the synthesis of the policy documents was that there was a need to assess for mental health and substance use issues at the SARC. Older guidance suggested this should be in the form of a comprehensive assessment; however, more recent guidance recommended assessing risk of self-harm and vulnerability but did not specify how this should be done.

In the UK, there is a stepped approach to the provision of psychological treatment. NHS talking therapy services offer steps 2 and 3 of the stepped model of care, in which step 1 is delivered in primary care. Step 2 comprises low-intensity interventions for mild to moderate depression and anxiety delivered by psychological well-being practitioners, and step 3 comprises high-intensity interventions delivered by specialist therapists. Step 4 is treatment for severe and recurrent anxiety and depression, and is delivered by senior therapists. National Institute for Health and Care Excellence (NICE) guidance13 recommends that people with a diagnosis of PTSD should receive step 3 therapy within one month of the traumatic event and does not recommend low-intensity interventions due to the lack of evidence that they are effective. People with more complex and multiple mental health needs above NHS talking therapies level 3 should be referred to community mental health teams (CMHT) or acute mental health services. The role that SARCs play in providing support for survivors with mental health needs at or below this severity threshold is not specified. From the review of the policy documents, it was noted that there was a lack of guidance about how to address the drug and alcohol needs of people who attend SARCs.

European Union COSAI guidance recommendations were broadly similar to the UK SARC specifications but they included more specific guidance in directing SARCs to aim to actively engage people with mental health problems, ensuring provision of cognitive behaviour therapies (CBTs) to treat clients with symptoms of PTSD, advocating the provision of counselling support for families and close friends of the SARC client where needed, as well as access to specialist counselling support for people who identify as LGBTQ+.

Mental health background of sexual assault referral centre staff

A full set of descriptive statistics can be found in Appendix 1. The responses indicated that there is overall limited mental health experience in the core SARC skill mix. It is interesting to note that just under half reported that the counsellor had a mental health background or qualification. As seen in Table 3, this question was left blank by some respondents. It is therefore difficult to know whether a non-response indicates no mental health staff in that SARC or that those roles are not present in that SARC. The table presents the percentage ‘yes’ of those who responded to the question but also the percentage ‘yes’ of the total respondents to the survey.

Assessment of mental health and substance use needs

The respondents were asked to indicate the method (e.g. professional judgement, standardised screening tools) used to assess a set of common additional needs that survivors may present with including depression, anxiety, suicide, drugs and/or alcohol and PTSD. From Table 4, it is noted that just over one-third of SARCs reported using standardised tools for any diagnosis (38%) and this seemed to be mainly for screening for drugs and or alcohol use (24%) and PTSD (21%).

Mental healthcare provision

Only 7% (2/28) of SARCs stated that a talking therapy service was provided or co-located within the SARC (even if the provider was a different organisation). One-fifth of respondents (21%, N = 6 of 29) reported that they had access to a clinical psychologist, either in-house or co-located within the SARC (even if the provider was a different organisation) and half (46%, n = 13 of 28) reported that they had a counselling service provided within or co-located at the SARC (even if provided externally). Only 10% (n = 3) indicated that they had some form of in-house or co-located substance use service provided by an external organisation. Only one SARC stated that they had an ‘other mental health or substance misuse service’ provided or co-located within the SARC (even if the provider was a different organisation). This was indicated to be a ‘mental health nurse on an 18-hour per week contract’ and their role was to undertake assessments, manage and follow up referrals.

One question asked respondents to consider a range of follow-up services that would likely to be involved in mental health and substance use care and indicate whether there was a pathway agreement in place, whether it was informal and ad hoc, whether there were discussions in place but not yet implemented, and finally if none and no plans for formalised onward referral (Table 5). The majority of SARCs reported that they had formal pathways to child protection and counselling for children and young people, but for adults it appears to be less formalised and variable between SARCs. Just under half of the respondents indicated that they had a written referral pathway agreed to single point of access (48%) and community mental health services (48%), and only 40% to the local mental health crisis teams. Only 36% had a formal pathway to drug and/or alcohol teams and 19% had a formal pathway to talking therapy services.

Sexual assault referral centre respondents were asked to rate how satisfied they were with the level of integration with external agencies. Figure 5 indicates that SARCs are most satisfied with integration with external sexual abuse/assault therapy services, and 64% were very satisfied/somewhat satisfied with mental health services. However, satisfaction was most positive in relation to substance use services with 80% of respondents indicating very satisfied/somewhat satisfied with integration.

FIGURE 5.

Satisfaction with level of integration.

At the end of the survey, a free-text question asked respondents to make suggestions about what could improve integrated mental health support. The free-text responses were grouped thematically as follows (with the number of related responses in brackets):

• SARCs having more ‘resource’ (n = 5).

• Additional training and development occurring in SARCs (n = 6).

• Additional training and development occurring together with external mental health services (n = 7).

• Improved pathways when making referrals to external agencies (n = 10).

• Improved responses when making referrals to external agencies (n = 12).

• Provision of specialist mental health services on SARC sites (n = 17).

Recruitment

Participants were recruited from four of the six case study sites. A total of 236 consents to contact were received from people who had recently attended a SARC, of which 157 (67%) were contacted by a member of the MiMoS team. Of those contacted, 78 (50%) gave informed consent to participate and completed the online screening (Figure 6). Two were excluded from the analysis as they were under 18 years of age, leaving 76 to be included in the analysis.

FIGURE 6.

Participant flow.

Sample characteristics

The sample was recruited mostly from the London SARC. However, on comparing sample characteristics from London and non-London, there were very few demographic differences except that in London:

• there were slightly fewer people who identified as white

• participants had slightly higher educational attainment, were more employed and fewer were living in temporary accommodation

• slightly lower mental health needs were identified and lower prescription of medication for mental health and alcohol/drug treatment.

Most of these differences were not statistically significant except for those living outside London who were more likely to state that they were prescribed medication for mental health or substance use difficulties (64%) than participants living in London (36%) (Pearson Chi-square p ≤ 0.05). Therefore, the sample was not disaggregated by London/not-London.

London and the south were less affected by the impact of autumn local area lockdowns compared with the north of England, and this is one of the reasons why recruitment was more successful in these areas. Most of the sample (80%) were in the 18–24 years and 25–34 years age groups. The vast majority were women (91%), with 6% identifying as male and 3% identified as trans/non-binary. In terms of sexuality, most (73%) identified as straight; 8% as gay/lesbian and 19% as bisexual. Most (65%) of the sample identified as white; 13% as Asian/Asian British and 13% as Black Caribbean, African, British.

The survey demographics were compared with national data from the 2018–19 Sexual Assault Referral Centre Indicators of Performance (SARCIP) returns. This showed that the survey was broadly representative of the overall SARC population in 2018–19, with a slightly lower proportion of people identifying as white due to the dominance of participants from London and slightly more participants aged 25–34 years.

Prevalence of mental health, alcohol and drug needs

A series of standardised screening tools were used to measure mental health, well-being and substance use needs. The moderate to severe cut-off score for the CORE-10 is above 2 and the clinical cut-off (indicating those in need of psychological therapy compared to a non-clinical population is 1). We are deliberately reporting the percentage of those scoring in the higher severity range and therefore likely to be experiencing an elevated level of psychological distress requiring specialist mental health assessment and support (Table 7).

These results showed an overall prevalence (95% CI) of:

• 76% (65–85%) with moderate–severe psychological distress (CORE-10)

• 12% (6–21%) scored between 8–15, indicating ‘increasing risk of alcohol problems’ (AUDIT-C)

• 94% (86–98%) PTSD (PC-PTSC-5)

• 87% (77–94%) were in the clinical range for low quality of life (Re-QoL-10)

• 45% (33–57%) moderate to severe health problems (Re-QoL)

• 26% (17–39%) moderate to severe drug problems (DAST)

• 63% (50–74%) personality disorder (SAPAS).

In 2018–19, 14,357 adults (aged 18 years or over) attended SARCs across England. By applying the unadjusted prevalence of mental health or substance use conditions to this total (Table 8) we estimate that, of these 14,357:

• 10,911 had moderate–severe psychological distress

• 1700 had increasing alcohol risk

• 13,548 had likely PTSD

• 12,511 had clinically low quality of life

• 6423 had moderate to severe health problems

• 3800 had moderate to severe drug problems

• 8973 had a probable personality disorder.

The prevalence of each mental health condition or substance misuse varied slightly by various demographics (age, ethnicity, sex/gender) – but disaggregated sample sizes were relatively small. When applying the disaggregated prevalence to the disaggregated national SARC population for 2018–19 (see Appendix 3), the weighted estimates show broadly similar numbers of SARC attendees with identified mental health issues, alcohol and/or drug use.

Although sample sizes are small, a cross-tabulation between outcomes shows that there is some tendency for dual diagnosis of mental health and substance misuse (Table 9). These differences were statistically significant only for 72% of individuals with moderate to severe drug misuse who also had health problems (compared with 38% of those with no/low drug misuse).

Realist methodology

The mixed-methods study was informed by an iterative programme of realist inquiry, which used realist evaluation20,21 and realist synthesis approaches22 to develop and refine programme theories and provide an increasingly evidence-informed theoretical framework to understand emerging findings.

The initial programme theories explained how the SARCs, and related services are intended or expected to achieve specific outcome patterns. As the project proceeded, further theories were developed and refined to encapsulate the diversity of experiences and explain what works, for whom, in what circumstances and why. The programme theories were developed and refined iteratively using tools such as logic models, if–then statements and context, mechanism and outcome configurations (CMOCs) and supported by data from a range of sources including WPs 1–5.

While each WP produced its own discreet outputs, an iterative approach was taken; each stage was informed by the findings of the previous stages, as well as within each stage (e.g. qualitative interview schedules were modified as patterns were identified, and theories were developed which were tested at the subsequent interviews).

We used a comparative realist evaluation approach comprising detailed investigation of the relationship between organisational processes and contexts,23 which generated a holistic understanding of the phenomenon under investigation. 24 Using this methodology, the study sought to explain and understand current SARC service provision by considering how variations in service provision and processes interplay with important aspects of individual service users, thereby influencing how mechanisms work and produce different outcome patterns. 21 This was achieved by developing, refining and testing hypotheses (programme theories) that described how various outcome patterns (O) were observed depending on whether the contexts (C) allowed specific mechanisms (M) to operate. These hypotheses are called CMOCs. A simple example could be the following:

• Mechanism (resource) – identification of need and referral to external mental health services.

• Context – previous positive/negative experience of mental health services.

• Mechanism (response) – service-user decision to take up the referral.

• Outcome patterns – attending/not attending mental health services, which influences longer-term expected outcomes such as improved wellbeing and reduced risk of further assault.

A CMOC incorporating these elements could be constructed as follows. It is expected that identifying a mental health need and subsequent referral to external mental health services (M) will result in improved wellbeing (O). However, if the service user has had previous adverse or unsatisfactory experiences of these services (C) and as a result decides not to take-up the referral (M) then the anticipated desirable mechanism (decision to take-up the referral) will fail to ‘fire’. Not only will the anticipated benefits of attending these services not be realised (O), but the service user might experience adverse outcomes such as despair or lack of hope (O), thereby potentially leading to worsening of psychological state and higher risk of further assault.

In this CMOC, the key elements that influence outcome patterns (O) are the referral process mechanisms (M) and whether they operate as expected, which is dependent on the context (C) of previous experiences of referral to these (or similar) services.

Realist review

To generate initial programme theories that would then be refined and tested in the case studies and other WPs, a realist review was undertaken. This used outputs identified in the WP1 systematic review. In line with realist methodology,20 the review aimed to create an initial framework for understanding how SARCs respond to mental health/substance misuse while identifying gaps in knowledge and key areas of interest. The review aimed to address the following question:

• What are the theoretical explanations relating immediate and longer-term outcomes to the identification and treatment of mental health and substance misuse problems in SARCs?

The review was based on international literature. However, the findings were prioritised for relevance to the UK sexual assault service models.

Methodology

The review protocol was registered on PROSPERO (CRD42020182808) and the review follows the RAMESES publication standards for realist reviews. 22

The realist review included papers from the systematic review and built on its searches. 7 We updated the search in the PsycINFO and MEDLINE databases (20 March 2020) using broad search terms and synonyms for SARCs. The search was limited to humans but there were no language restrictions. We also consulted experts in the field to identify additional key papers. The purpose of the review was to develop a theoretical framework and initial programme theories that were specific to the SARC setting, which would then be further refined through the research activities of the other work streams. 25 Therefore, a single comprehensive search, specifically focused on SARCS, rather than iteratively developing additional theoretically driven searches was considered appropriate for the purposes of the programme as a whole.

All retrieved records were independently screened for relevance by two reviewers in a web-based review management system (Covidence, Melbourne, Victoria, Australia). Any disagreements were resolved by discussion and, where necessary, after discussions with a third researcher. A data extraction schedule was drafted, piloted and further developed. We extracted data using Microsoft Excel^®.

A separate row was used for each theory, whether this was a complete theory or fragment, rather than for each paper. This approach ensured that theory-rich papers did not result in theory-dense outputs and allowed more precise categorisation and comparison of theories. Data were extracted in the form of form of context (C), mechanisms (M) and outcomes (O), and included ‘if–then, leading to’ statements. This approach ensured that relevant details were extracted for theory development and the relationships between C, M and O were articulated. Training sessions and workshops were held for the team to develop skills and knowledge in realist synthesis. However, where members of the data extraction team were less comfortable with describing theoretical relationships, the ‘if–then’ statement column was completed later, by more experienced members of the team using the extracted data and with reference to the original sources.

Extracted information also included theory description, position in pathway, intervention, expected outcomes, for whom, in what setting; generalisability for UK SARCs, usefulness, plausibility, relevance and trustworthiness of the theory. We extracted data for all included papers, found both from the original and the updated search. Most members of the review team extracted data and were involved in discussions around theory generation.

Microsoft Word^® was used to collate extracted data according to their position in the pathway. The extracts were then compiled into themes and subthemes that were relevant to these pathway positions. These documents were then reviewed and the contents synthesised into further refined ‘if–then’ statements and CMOCs. Regular discussions among the review team facilitated theory generation. During this process, consideration was given to associations with existing mid-range theories or the development of bespoke theories.

An example of the development of synthesised CMOCs is provided below. This example is positioned at the ‘pre-SARC’ point in the process pathway.

Findings

The initial programme theories were developed from the study proposal and discussions with the study team. For instance, the prevalence of unmet needs indicated problems with the start of the pathway, such as access, referral, and case-finding issues. Rates of mental health service-use, PTSD and substance use are also known to be high in users of SARC services. However, it was not clear how well these needs were being assessed or met during first contact or subsequently (e.g. through referral pathways or follow-up). A process/logic model was developed, which included assumptions about what good-quality care would look like, possible decisions at key points in the pathway and important questions for the study to answer.

Initial hypotheses were developed, which assumed that attending a SARC would be a good thing for those with mental health and/or alcohol/drug use needs. In theory, attending a SARC means that:

• Survivors can have their mental health and/or alcohol/drug needs assessed; can be referred to talking therapies or drug and alcohol services where needed; can be supported to access other services to have their healthcare needs met, which might otherwise cause stress; that is, SARC attendance provides a point from which formal support services can be accessed.

and

• In the immediate aftermath of assault, survivors can gain reassurance that they will be believed; they can have negative appraisals about blame challenged by SARC staff; can regain a sense of control and safety. That is, SARC attendance might reduce risk for developing/exacerbating MH/SU difficulties through informal (trauma-informed) ways that survivors are treated, which increase sense of safety.

These two processes may also interact – survivors are more likely to disclose difficulties and take up services where they have experienced trauma-informed care. The synthesis aimed to develop initial programme theories that begin to describe how these formal and informal processes impact on outcomes for people with histories of mental health problems and substance use.

The findings were grouped into three time points to reflect the journey of service users through support services and to highlight places in this journey that can facilitate or impede improved outcomes. Within each section is a theory statement, developed from CMOCs that provide an explanatory interpretation of the synthesised literature. The sections are titled as followed:

1. Pre-SARC – this includes awareness of the SARC service, what is known about its purpose, how it is accessed (via police, third party service, self-referral).

2. Within the SARC – first attendance and initial impressions, process and procedures, assessments and methods of assessment, the forensic examination (consent, capacity, being informed, psychosocial support, physical environment), follow-up support, appointments, signposting and information giving.

3. Post SARC – how referrals to external agencies are conducted, follow-up support (e.g. for legal processes), satisfaction with support, gaps in onward service provision.

These are presented below and in summary form in Table 10.

Barriers to access

Important barriers to accessing services are related to lack of understanding and awareness about the existence of services, what they provide and how they are organised. Even when SARC services were known about, fear and anxiety about the consequences of contacting the services could be a significant barrier, especially when considering the prevalence of mental health problems and post-traumatic stress in this cohort. The relationship with the justice system could also be off-putting for certain potential service-users, for instance those with substance use problems.

This lack of awareness applies to the general population as well as organisations and agencies that might be in a position to signpost or refer service-users. There was a lack of evidence about the best ways to organise and deliver awareness-raising activities and where responsibility should lay.

There are further concerns that specific, vulnerable groups might face additional barriers to access, and that (partially as a consequence of low levels of exposure to or identification of these groups) SARC staff might be lacking in skills or knowledge about providing appropriate services.

Identifying and responding to mental health and substance-use needs

The literature suggests that consistent, systematic assessment processes are beneficial, both directly to identify and respond appropriately to unmet needs and indirectly to raise the profile of mental health and substance use problems in the cohort, thereby creating a more supportive culture.

Training in trauma-informed care and the promotion of trauma-informed interactions are also viewed as important elements in the provision of appropriate care. These can reduce distress, prevent re-traumatisation, improve long-term outcomes and potentially improve access through reputation and prior experience, in the case of repeat victimisation.

Service organisation

Multidisciplinary teams could have many advantages, including speed, quality and variety, coordination and continuity of care. A critical area that would benefit from further exploration is the integration or differentiation of roles. Potentially, this could be addressed by initiating a set of core competencies and responsibilities, such as FMEs being trained to deal with emotional needs. However, some elements of role clarity are important to maintain to reduce confusion and variability of service provision.

While relationships with non-co-located services can expand the services available to SARC attendees and provide access to specialist services, maintaining these relationships and positive communication requires resources and ongoing commitment.

There were several ways in which the systematic collection and use of route data could be used to improve and monitor services. For instance, in identifying training needs, providing tailored care, and addressing underrepresentation of certain groups.

Post sexual assault referral centre services

As discussed earlier, attendance at further appointments can be influenced by experiences at the initial appointment. Key mechanisms in the literature associate good outcomes with positive experiences, leading to service-users’ trust, recognition of their own needs and feeling empowered. Practical support can also be important to help coordinate follow-up support, particularly when service users have difficulties in arranging their own care through experiencing distress. There can be significant barriers in accessing mental health services, particularly regarding availability of appropriate services. This can lead to additional pressure being experienced by SARC staff members, particularly when the SARC does not have appropriate resources for mental health support.

Identification of data sources

Service managers and personnel were asked for their consent to support the identification of eligible documentation during WP2 (national survey of SARCs). During the case study site visits, SARC staff and/or researchers identified eligible materials through hand searches and SARC service intranet searches. In addition, internet searches identified documents such as Care Quality Commission reports and published journal articles related to the sites.

Data extraction

Data from all relevant documentation were extracted by a researcher using a standardised extraction form developed by researchers and based on SARC national service and NHS England guidance. This form was used to extract data on any policies, procedures and guidelines related to mental health and substance use (e.g. assessment protocols, referral pathways, intervention strategies). A second reviewer (another member of the research team) independently extracted data from a random sample of 20% of documents to verify intersubjectivity of decision-making, eliminate ambiguity and improve communicability of key concepts. All data were extracted into a Microsoft Excel spreadsheet.

These data were then used to create a ‘journey map’ for each site, providing a condensed and chronological representation of how SARC pathways identify and respond to mental health and substance use needs.

Programme theory development

The journey maps were used to refine initial programme theories identified in the realist review and through stakeholder consultation. This was done by examining evidence relating to the complex patterns of contexts, mechanisms and outcomes;21 for example, identifying potential blocks or aids to receiving appropriate support. Journey maps were also used to inform staff and service user interviews in WP4 by identifying any points of clarification needed about SARC processes and any opportunities for emerging theories to be tested. This process was supported by the LEAG, who reviewed one of the journey maps and suggested questions for staff based on this review.

Journey maps

Journey maps for each of the case study sites were created using a realist methodology and has subsequently been published. 8 These journey maps were used to orientate the researcher to the emerging context–mechanism–outcomes and inform the topic guides for each focus group and interview that was subsequently conducted.

Sample and setting

A purposive sample of service users across the six identified SARC case study sites was recruited to ensure diversity with respect to gender, age, ethnicity, mental health status and timing of assault (recent or non-recent). The case studies were chosen to represent the three types of SARC models of working with mental health and/or drug and alcohol use that were identified from a cluster analysis of variables from the WP2 national audit of SARCs.

Survivor inclusion:

• aged 18 years or over

• able to read and understand English or there are suitable confidential translation services available

• had an identified mental health or substance use need to take part in the interview.

Exclusion:

• lacks capacity to provide informed consent

• where participation is deemed to significantly increase risk to self or others.

SARC and partner agencies’ staff were also purposively recruited across the six study sites to ensure representation within the workforce; for example, within SARCs to ensure representation from management, FMEs and nurses, independent sexual violence advisors (ISVA), and within external agencies to ensure representation from relevant third-sector partners including sexual and domestic violence sector services and sexual violence counsellors, NHS mental health services and local substance use treatment providers. Where SARCs provided in-house or co-located therapeutic services, these staff were also approached.

Ensuring rigour

Several researchers undertook the coding. Therefore, prior to coding, researchers undertook training to identify key features in the data (i.e. contexts, mechanisms, and outcomes), with all researchers then coding the same transcript and comparing findings to assess intercoder agreement. Additionally, a code book was developed, which included descriptions of all the codes and which was regularly updated, ensuring that all researchers were using the codes consistently. Similarly, at the stage of developing theories, all researchers explored the data within one theme/node and then met to discuss findings. This process ensured that all researchers were confident in the method of analysis and had opportunities to discuss any challenges. The research team met regularly during the analysis process to discuss emerging theories and consider additional questions that may help to test and refine these theories.

In addition to weekly researcher meetings, three whole-team meetings were held (October 2019, September 2020, May 2021) to discuss emerging findings and plan the next stages of data collection and analysis. These meetings included co-applicants, researchers, and lay representatives. Members of the LEAG contributed both in person and by email correspondence. These meetings helped to ensure that the study aims were being met and that data from other WPs were being considered alongside the data from WP4.

Consensus across the study team workstreams was sought at key points in the investigation; for instance, development of initial theories to be refined and tested and construction of the coding framework. This was achieved through a combination of workshops, meetings and discussion documents circulated by email.

Recruitment and demographics: sexual assault referral centre staff and partner agencies

A total of 26 interviews and focus groups were conducted with SARC staff, representing 31 individuals, and 17 interviews and focus groups were conducted with staff from partner agencies, representing 20 individuals. See Table 11 for demographic details. Interviews were conducted with staff and partner agencies from all six SARC sites and were purposively selected to represent the range of roles at each SARC based on the information obtained via the national survey (WP2) and documentary analysis (WP4).

Within sexual assault referral centre analysis

In this next section, data are presented on the within-case analysis from SARC staff and partner agency care providers in each of the six case study sites. Four central themes generated from the programme theory have been identified where there is strong and consistent data emerging:

1. Access to the SARC and initial engagement.

2. Identification and assessment of mental health and alcohol/drug use needs.

3. Care and treatment for mental health and psychological distress that is provided within the SARC service.

4. Care and treatment that is provided by partner agencies.

Between-case analysis

In WP2, a cluster analysis was performed on the mental health-related variables and three clusters of SARC respondents emerged. While this was useful in informing the purposive sampling of the SARCs for the case studies, there are some considerable differences between the SARCs even within the same cluster as well as between clusters. This next section focuses on the crosscutting similarities as well as some divergencies that have been identified with reference to the programme theories identified in the realist review.

In-house psychological support and external mental health and drug and alcohol support

Sexual assault referral centre 5 has an in-house psychological therapies pathway, which is the most comprehensive service within the six SARC case study sites. This is led by a clinical psychologist who is an expert in trauma. There are two benefits to this service. First, the person will receive an expert triage of their psychological needs and timely access to appropriate support, either in-house counselling delivered by two local non-government organisation (NGO) rape counselling services, or in-house trauma therapy delivered in partnership with the local mental health provider. The second benefit of this service is the impact that the clinical psychologist has to the ethos of the SARC team by promoting the trauma informed care approach in team meetings and other informal discussions. The input of the clinical psychologist also supports the SARC staff in managing the vicarious trauma they may experience. This was also seen as a benefit in the other SARCs that had in-house provision (SARCs 1 and 6).

Whilst we have training in understanding the role of vicarious trauma, our nursing colleagues won’t. And I think it’s just really important that we are always just checking in with that really, in a very informal basis. We don’t provide clinical supervision but I think that those discussions go some way to just normalising some of the experiences that our nurse colleagues may be having.

(SARC 5 clinical psychologist)

However, there are also some challenges too. Despite the comprehensive in-house offer, there are some people who are identified as having too complex needs for the in-house service, and they are then referred to the mainstream mental health services. There are gaps in provision for people who are too complex for the in-house offer yet not crossing the threshold for secondary care. The other barrier is long waiting lists even if the person crosses the threshold (NICE recommends that trauma therapy should commence within four weeks of the event). The only other SARC with clinical psychology is SARC1 (London). As in SARC 5, the presence of the clinical psychologist impacts on the assessment and triage of the person who presents at the SARC, and every case is reviewed in a multidisciplinary meeting. SARC1 does not, however, offer the same level of in-house support as SARC5 and will refer to partner agencies for trauma therapy and other mental health support. SARC1 also report challenges in referrals due to high thresholds and long waiting lists and even stated that they offer some form of support to people they feel have high needs rather than leave them with no support at all. There was a suggestion that low level and immediate support may be helpful and potentially prevent a longer-term psychological problem. SARC6 has an in-house counselling service, which can offer supportive interventions that can help stabilise trauma reactions but does not offer trauma therapy and does not work with people with severe or complex needs; however, there is no formalised psychological assessment at SARC6 to inform decisions about whether someone needs higher-intensity therapies.

In the rest of the SARCs in the study, the crisis workers offer most of the emotional support at first contact, the main examination and via some form of follow-up calls and appointments; the nature and purpose of the ‘follow-up’ is inconsistent across SARCs. In some SARCs, the crisis workers are involved alongside the FME in undertaking a mental health assessment, including risk assessment as well as enquiry about drug and alcohol. Crisis workers and FMEs come with a variety of backgrounds, and few have a mental health practitioner background and/or training. Without the presence of a mental health expert (such as in SARCs 5 and 1) they report a lack of understanding of what types of mental health and other psychological support is available and how to refer. There is also a lack of use of standardised tools to screen for mental health and substance use needs. Use of standardised screening tools can be helpful in facilitating referrals to external agencies, as well as providing a consistent way of identifying needs, as variability of assessment of risk and mental health between practitioners has been mentioned.

There is no evidence of any support offered in relation to substance use. SARCs will either signpost the person to their local drug and alcohol service or the person must self-refer. In some examples, SARC1 reports a positive relationship with local drug and alcohol services and if they refer someone to them, the drug and alcohol service will see them, if possible, on the same day and will fast track them into treatment rather than go on a waiting list. The drug and alcohol partner agencies interviewed do not report a close relationship with the local SARC and also do not routinely ask about sexual assault.

Survivor interviews recruitment and demographics

A total of 12 service user interviews were undertaken between September 2020 and April 2021. Of those approached from the prevalence study participation, 22% consented to be interviewed.

It was challenging to recruit people who attended a SARC for the reasons outlined in Chapter 7. Most of the people recruited for the qualitative interviews were drawn from those who had previously participated in the prevalence study (WP3) and recruited from four of the six case study sites. Owing to under-recruiting in WP3, there was a much smaller pool for WP4. Survivors were not recruited from SARC 5 or 6 despite efforts to promote the recruitment by the MiMoS team, the services were under strain due to impact of COVID-19 and not able to undertake the promotion of WP3 or 4.

There were 10 female, 1 male and 1 trans male recruited. Most had self-reported mental health issues (Table 12).

Analysis of survivor interviews

Onward support and care

A survivor (female 1 SARC3) who had panicked during the forensic examination recalled that the SARC staff and examiner were concerned about her immediate well-being, and as she was under the care of the local mental health services, the SARC called the care coordinator and waited for a call back. However, the survivor recalls that:

They [mental healthcare coordinator] did call back and said send her home, obviously that is why they had to say okay we have done our bit, we don’t think you should be going, you know our options are tied now … they were good like in the way … they worked out I was under them [mental health service] and yeah tried to do their best.

(Female 1 SARC3)

While many remembered discussions about their support needs, they found it hard to take in and retain information. Some remember being given a pack of information to take home, and how they were not able to start to digest the information until a few days later:

I didn’t take it all in … I just took the leaflet from her … weren’t til a few days later that I sat with the leaflet that’s when I took in more.

(Female 2 SARC5)

They made sure I knew there would be options of different types of therapies available, like group sessions or counselling or whatever and I’d get a say in what that looks like. There, obviously, is waiting lists for these things.

(Female 1 SARC1)

Another survivor recalls leaving the appointment with a ‘stack of A4’ with lots of information. The SARC was also able to connect her to a women’s project and she was at the end of eight weeks of group therapy at the time of the interview, and that she expected to receive one to one therapy from the same organisation after the groups ended. This survivor was unable to access the in-house support that SARC1 usually offers because of the limited service running at the time of the pandemic. She was asked if she would have preferred this support to be offered by the SARC or externally:

I think I would have preferred that a little bit because obviously you kind of … like I had got along with people that I had worked with so if I got it [psychological support] from there I would feel a bit more comfortable, yeah.

(Female 2 SARC1)

She also said that she was grateful that the SARC had facilitated the referral to the women’s counselling service as she saw self-referral as a bit daunting.

One survivor at SARC1 stated that it was hard to think about what sort of continuing support would be useful at the time of the SARC appointment, as she was still processing what had just happened to her. At the time of the interview, she was still waiting to hear the outcome of referrals for support but stated that the delay was not such an issue as it had given her time to process what had happened.

Survivors mentioned receiving follow-up calls from the SARCs. In this conversation, survivors recalled that SARCs were able to check on progress of referrals or whether any other support needs had arisen. This is also useful as many survivors talked about not being able to think clearly about what support they may need at the time of the main SARC appointment, so having follow-up contact also allows a chance to consider what they need after the acute reactions had subsided.

Survivors mentioned that they also appreciated the follow-up calls from the SARCs because it helped them feel that there was someone who genuinely cared about them.

yes they are quite good at just you know every now and then they’d give you a ring just to see how I was and that I am doing ok. Which you know at least makes it seem like they care.

(Female survivor 1 SARC1)

In contrast, another survivor spoke about how the lack of a follow-up call after visiting SARC5 was very disappointing:

It’s very disheartening, very disheartening to think that you have gone to a service and you feel deflated on the situation that’s happened to you, and they haven’t checked on you necessarily mentally, emotionally, and obviously physically but it’s kind of tough.

(Female 3, SARC5)

One survivor talked positively about being referred to an ISVA and that she was able to access specialist sexual violence counselling as a result.

this is one of the best things to come out of the SARC … they referred me to that [ISVA] … And then through my ISVA I’ve accessed counselling, like specialist counselling around sexual violence …

(Female 1 SARC4)

She only had to wait five weeks to access the specialist counselling and was offered up to 30 sessions. Owing to COVID-19 restrictions, the counselling was being offered via Skype. However, despite the short waiting time, she felt that it would have been good to have had it when the incident actually happened ‘because I was struggling with quite a few things at that time’. The survivor also mentioned exactly why specialist counselling was important as she had been worrying that she had imagined the incident and trying to rationalise it in her head, but the counsellor had normalised her reactions as part of the trauma response.

speaking to my counsellor about that, she said that’s a very normal response to trauma, thinking did it actually happen …

(Female 1 SARC4)

One survivor was referred to an external agency for rape and sexual assault counselling and when she received a follow-up call from the SARC, she was asked by the SARC if the external agency had been in touch and they hadn’t, so, she was advised to go to the website and self-refer (and after that, she had a contact within three days). The survivor mentioned that someone else had helped to complete the online self-referral. In addition, the SARC also helped her to get other mental health support which she found very beneficial:

I don’t think to be honest I would be here today. I wouldn’t be how I am today. I can go out now on my own and feel comfortable but before [the services helped her] I couldn’t go out at all.

(Female survivor SARC5)

it definitely does help going to counselling that is for sexual assault and rape … because they know how to word things so that you are not going to get offended … they’re brilliant they know what they are doing.

(Female survivor 1, SARC5)

One survivor had been receiving therapy at the in-house provision at SARC5 and found it contrasted significantly with previous help she had received:

literally, every sort of type of healthcare professional that I have ever spoken to or any referral places that I have been advised about or whatever, they have never wanted to help. They don’t seem to listen, and they don’t get what is going through your head. But with [SARC] they listen to you as a person not what you are struggling with, and they treat you as an individual which you are. Then they work out how to help you based on what you personally need rather than what they think everybody suffering with the same thing needs.

(Female 2 SARC5)

She also mentioned that the SARC therapist made her feel that she mattered because she felt listened to and not just offered medication.

a lot of people I have seen before, even like doctors, you go there with … anxiety or depression and they think everyone suffers the same and has the same triggers and medication will work ... they treat you as a blanket rather than you are an individual.

(Female 2 SARC5)

The difference between the experiences seems to hinge on feeling that the person is not being judged, treating the person as an individual, and that the counsellor or therapist understands issues related to sexual assault (context).

One survivor at SARC1 was receiving counselling online after waiting six months (this was during the 2020 COVID-19 pandemic) and found it useful to access this online as she didn’t need to leave home and go anyway, and this felt safe: ‘in a safe space at your home … but then it can feel a bit weird trying to, you know, have these certain conversations through a screen’ (female 1, SARC1).

In contrast to the positive impact of specialist sexual violence counselling (either in-house or external agency) survivors generally talked despondently about accessing continuing support for their mental health and/or drug and alcohol needs, and the lengthy waiting lists for NHS therapies. One survivor described how he struggled to find support because he had mental health and alcohol/drug use issues and therefore did not fit the criteria for existing services due to his multiple needs:

I’ve had a real long battle with substance and mental health … so it’s all a bit of a nightmare, so we don’t get picked up by anyone really.

(Male 1 SARC5)

Others mentioned long waiting lists for NHS-provided therapies:

I’ve been on the waiting list for NHS therapy since October 2019 and no idea where that is.

(Female 2 SARC1; interview conducted in Autumn 2020)

In contrast to the language used to describe how they were treated in the SARCs and other specialist sexual violence services, the survivors talked about their perceptions of mainstream mental health including the lack of understanding and feeling judged.

I’ve had talking therapy … that was on the NHS, and I thought I am never going to go again … I just didn’t like her. I felt she was judging me, so I said I didn’t need it and walked out’ and ‘other services they [SARC] send you to that are nothing to do with what you have been through [sexual assault] like they were speaking to you, you think what the … hang on a minute, you aint a clue what you’re talking about. This ain’t how I felt … or this ain’t how it happened, you’re judging me.

(Female 1, SARC5)

I think the NHS can be really poor sometimes … I think GPs can be quite insensitive towards mental health problems, so I think [the SARC] was better.

(Female 2 SARC1)

The impact of receiving a less than positive experience from external mental health services often lead to disengagement.

Those who were already under the care of the mental health services found that the lack of a response in a crisis meant they were less likely to ask for help in the future:

that’s the story of the mental health team, you know sometimes they are good and other days they leave you to it … they [therapy group] rang them [mental health team] on the Friday to say look [name deleted] is in a bad way and literally got told we’ll contact her at the beginning of next week and then my daughter found me on the floor, and I’d taken an overdose. So, it’s kind of a bit like I won’t bother to ask for help because what’s the point you end up getting upset and let down.

(Female 1, SARC3)

And in addition,

no the crisis team aren’t very good in this area. I’ve only had one experience with them where we actually spoke, and we got on well. The rest of the time I’ve found them quite rude, dismissive.

(Female 1 SARC3)

One survivor who has bipolar affective disorder was admitted to mental health ward shortly after the visit to the SARC, but it was not described as a helpful experience, and they took their own discharge:

My CPN [community psychiatric nurse] doesn’t realise how horrible the [mental] hospital is … people were attempting to bully me … so, it’s not just the patients trying to bully you, it is often the staff. So as soon as I had the opportunity, I walked out of there.

(Trans female, SARC3)

It is not clear whether the bullying mentioned was due to transphobia but experiences of transphobia from other agencies (not specified) had been mentioned earlier in the interview.

Demographic and clinical characteristics in the Clinical Record Interactive Search

Gender, age (calculated using date of birth and date last seen) and ethnicity were extracted from structured fields. Social deprivation was measured using the 2019 Index of Multiple Deprivation (IMD), a measure of relative deprivation for neighbourhoods in England termed lower super output areas (LSOAs). IMD was derived from the LSOA associated with the patient’s most recent address, and converted into quintiles (1 – most deprived, 5 – most affluent).

Diagnoses are recorded in CRIS using the International Classification of Diseases, revision 10. Individuals are assigned a primary diagnosis when in contact with a SLaM service. As patients may have several spells of care under SLaM (defined as a continuous period of care) and therefore may have been assigned multiple diagnoses, we explored several related variables:

• all primary diagnoses given between 2000 and 2019, extracting the percentage of individuals who had been assigned each major classification

• the most recent primary diagnoses assigned to the patient, extracting the percentage of individuals who had been assigned each major classification

• the number of different diagnoses that individuals were given between 2000 and 2019.

Level of risk was assessed using a structured risk assessment tool, which is completed by clinicians during assessment if the patient is deemed to be at risk either to oneself, to others or from others. This tool specifies several risk clusters, including self-harm/suicide, violence, self-neglect, physical health, and adult safeguarding. In addition, it specifies triggers and protective factors that may be present for the individual. Under each of these seven headings, the clinician completes checklists about the presence of current and historic risk factors which may contribute towards each one (e.g. social isolation, alcohol misuse).

Completion of the structured assessment is dependent on staff behaviour and, therefore, may not identify all those with risk factors. Nevertheless, a pragmatic approach was taken, assuming that, in general, where the tool has been completed, this is due to a higher level of risk being identified.

To measure risk, the proportion of sexual assault group and controls who had a completed tool was used to indicate the likelihood of full risk assessment being undertaken. In addition, within each of the risk clusters, the percentage of individuals who had been identified as having the current and/or historic risk factor was determined to identify the level of each risk.

Demographics and clinical characteristics in sexual assault referral centre 5

The two standardised clinical datasets from the talking therapies service at the NHS trust and from the third-sector organisation were combined into a single dataset for analysis.

The following demographic and clinical variables were used in the analysis of the SARC dataset: gender (categorical variable), age at referral (continuous variable), ethnicity (categorical variable), marital status (categorical variable), relationship to assailant (categorical variable) and number of therapy sessions received by discharge (continuous variable). A site code (i.e. NHS or third sector) was added to the combined dataset so that analyses could explore demographic and clinical differences between the sites.

Sexual assault referral centre 5 dataset

All data were analysed using StataMP16 (Stata Corp., College Station, TX, USA) statistics software. Descriptive statistics of demographic data, proportions and Pearson chi-square tests for categorical data and means and standard deviations for continuous data, were calculated for the entire sample and also stratified by site (i.e. NHS NGO, counselling). Tests of associations on clinical data were calculated, using means, standard deviations, 95% confidence intervals and t-tests for all normally distributed data and for non-normally distributed data using medians and non-parametric Mann–Whitney U-tests.

A thematic analysis of the brief textual summaries of reasons for discharge from therapy was also conducted, to provide further details on peoples’ treatment journeys.

Clinical record interactive search dataset

All variables were analysed using IMB SPSS (IBM, Armonk, NY, USA) statistics software, version 26 for Windows. Pearson chi-square tests were used to compare categorical variables and Mann–Whitney U-tests were used to compare means of continuous variables.

Descriptive statistics – proportional for categorical variables and means and standard deviations for continuous variables – were calculated to describe the sociodemographic and clinical characteristics of the sample.

Main statistical analyses examined changes in psychological well-being (based on CORE outcome assessments before and after treatment) following psychological treatment from a mainstream mental health service compared with a SARC. We examined the sociodemographic and clinical characteristics of the cohorts to explore whether there are systematic differences for example, history of previous psychiatric contact such as admission. We then addressed these differences using appropriate statistical techniques.

South-east England sample characteristics

As can be seen in Table 13, both NGO counselling and the NHS supported mostly female service users; the NHS supported significantly more male service users than the NGO (p < 0.05). NGO counselling services were significantly more likely to support service users of a younger age (p < 0.001) and service users of Asian and black ethnicities (p < 0.05) compared with the NHS services. The marital status of service users was similar across the two services, as was the service users’ relationship to the perpetrator.

Assessment of psychological distress pretreatment

Of the 467 referrals that NGO counselling received between 1 April 2016 and 31 December 2020, a total of 227 (49%) completed a CORE-OM assessment prior to starting treatment. Of the 229 referrals that the NHS therapies service received between 1 July 2016 and 31 January 2020, 76 (33%) completed a CORE-OM assessment prior to starting treatment.

The 76 SARC service users in the NHS therapies pathway (M = 1.87; SD 0.66 CI 1.72 to 2.02) compared with the 227 SARC service users in the NGO counselling therapy pathway (M = 1.63; SD 0.86 CI 1.52 to 1.74) reported statistically significant higher mean levels of psychological distress prior to starting treatment t (166) = 2.5, p < 0.01.

These findings highlight that, on average, SARC service users on both the NHS therapies pathway and the NGO counselling pathway had symptoms of moderate psychological distress within the clinical range prior to receiving therapy. In addition, SARC service users on the NHS therapies pathway had higher baseline levels of psychological need (prior to starting treatment) than SARC service users on the NGO counselling therapy pathway. These observations align with the tiered model of the SARC5 therapies pathway, whereby SARC service users with greater psychological needs are triaged to the NHS therapies pathway and those with lower psychological needs are directed to the NGO counselling therapies pathway.

Assessment of psychological distress post-treatment

A total of 88/227 (39%) SARC service users on the NGO counselling therapies pathway completed a CORE-OM assessment at the end of treatment. A total of 22/76 (28%) SARC service users on the NHS therapies pathway completed a CORE-OM assessment at the end of treatment.

Post-treatment analyses revealed that overall SARC service users on the NHS therapies pathway (median 0.50; SD 0.34, CI 0.00 to 0.91) and SARC service users on the NGO counselling therapies pathway (median 0.41; SD 0.60, CI 0.00 to 1.88) reported considerable reductions in psychological distress, bringing their scores (on average) below the clinical range. No statistically significant differences were observed on levels of post-treatment psychological distress between SARC service users on the KMPT therapies pathway and the SARC service users on the NGO counselling therapy pathway (z = –0.78, p = 0.43; Mann–Whitney U-test). All 22 (100%) SARC service users on the NHS therapies pathway and 66 (75%) of SARC service users on the NGO counselling therapy pathway achieved reliable change (i.e. ≥ 0.5 mean change score before and after treatment) in levels of psychological distress post-treatment.

SARC service users on the NHS therapies pathway on average received a greater number of therapeutic sessions than SARC service users on the NGO counselling therapy pathway. All SARC service users on the NHS therapies pathway showed a reliable improvement in CORE score post-treatment, compared with 75% of SARC service users on the NGO counselling therapy pathway. A note of caution around these findings; there was a high percentage of missing data on CORE-OM post-treatment psychological assessments, particularly in the NHS therapies pathway group (missingness around 70%, n = 22 CORE-OM post-treatment data), and this impacts on our interpretations around the data.

Thematic analysis of reasons for discharge from therapy

We conducted a thematic analysis of summary textual data from the NGO counselling therapy pathway, concerning the reasons service users did not undertake therapy or stopped therapy prematurely. The analysis demonstrated that the main reasons for not undertaking therapy were due to not being able to find the time to attend whilst balancing various other commitments in their lives (e.g. university attendance, work schedules, exams, childcare commitments) and service users not responding to contacts by the service. Another barrier was service users not feeling ready to undertake therapy. Transportation difficulties were also raised as a barrier to engagement, with service users reporting being unable to travel alone to therapy sessions, raising concerns about traffic issues or reporting that they lived in a geographically remote area which presented challenges in travelling to the service for sessions. In response to the COVID-19 pandemic, the service temporarily changed the format of their therapy sessions to remote online delivery, which may have resolved some of the barriers raised by service users in the previous sentence. However, this format change did not facilitate engagement for all, and some service users elected to wait to undertake therapy until face-to-face sessions were reinstated. Further additional reasons for service users not undertaking therapy were that some had complex presentations that required support beyond what the service could offer and that some service users were already undertaking therapy with another provider. With respect to reasons for ending treatment early, some service users decided to stop early because they felt better, and others moved outside of the catchment area. A few service users ended therapy early stating that they wanted to move on and start with a new chapter in their lives, and a few others cited problems related to family issues (including domestic abuse, death/sickness of a family member/relative).

We conducted a thematic analysis of summary textual data from the NHS therapeutic pathway, concerning the reasons service users did not undertake therapy or stopped therapy prematurely. There were less textual descriptions for this dataset, so the richness of our analysis was more limited than for the NGO counselling dataset. The analysis demonstrated that the main reason for not undertaking therapy were that service users were not suitable for the service, either due to having greater mental health needs than could be sufficiently supported by the service or by having fewer mental health needs or additional relational needs that fell outside the scope of the service. In these instances, service users with greater mental health needs were referred on to CMHTs and other NHS tier-4 mental health services and those with lower needs or additional relational needs were referred on to NGO counselling or other third-sector specialist abuse organisations. The second biggest barrier was service users failing to attend sessions or disengaging from treatment (no additional details were given to indicate the possible reasons for disengagement). Further additional reasons for service users not undertaking therapy were that a few were already engaged with a CMHT and a few refused to have an initial therapeutic assessment (no additional details were given to indicate why service users refused these assessments).

Discussion

This analysis provides important insights into the impact of in-house SARC specialist therapeutic treatments in improving clinical outcomes among SARC services users. Prior to treatment, service users accessing both the NHS therapies pathway and the NGO counselling pathway reported moderate levels of psychological distress. Post treatment, service users accessing both pathways reported considerable reductions in psychological distress. Indeed, psychological distress scores fell on average to below the clinical range in both groups, despite service users on the NHS therapies pathway reporting considerably higher levels of distress at pretreatment. These findings highlight that SARC service users on the NHS therapies pathway achieve comparable improvements in post-treatment levels of psychological distress to service users on the NGO counselling pathways. One reason for this finding may be due to the fact that service users on the NHS therapies pathway receive more therapy sessions than those on the NGO counselling pathway.

The textual data in this clinical dataset allowed us to identify the main reasons why SARC service users do not undertake therapy. One important finding was that several service users were not suitable for the service, either due to having greater mental health needs than could be sufficiently supported by the service or additional relational needs that fell outside the scope of the service. These findings highlight that SARC services support people with considerable mental health and social needs, which require input from other relevant organisations (e.g. NHS mental health services, substance use services). Another reason why some service users did not take up the treatment was that they did not feel ready to engage in therapy or that they struggled to commit to this work alongside the other commitments in their lives. The result suggests that some service users may feel the need to access therapeutic support at a later date in time and this may mean that they are only able to access therapeutic support via NHS services, which may or may not have access to specialist trauma focused treatments.

Strengths and limitations

To our knowledge, SARC5 is unique in its provision of a bespoke in-house, stepped-care talking therapies service in partnership with a local mental health NHS trust and third-sector organisations. By examining SARC5’s clinical dataset we were able to examine the impact of this in-house therapeutic service in improving the psychological well-being outcomes of SARC users. We were also able to explore differences in therapeutic treatment outcomes among service users accessing two different therapeutic pathways in this stepped-care levels in this model (i.e. the NHS therapeutic pathway and the NGO counselling pathway).

This clinical database uses a robust and validated measure of psychological well-being to examine pre and post treatment changes among SARC users. It also includes detailed demographic and clinical information about SARC clients receiving the therapies service, as well as textual summaries of reasons for discharge from therapy. Despite these strengths, we observed a high percentage of missing data on CORE-OM post-treatment psychological assessments, particularly in the NHS therapies pathway group which impacts on our overall interpretations of the data.

We originally intended to use this dataset to compare changes in psychological well-being outcomes among people who received psychological treatments from this SARC with people who received psychological treatments from a mainstream mental health service. However, when we obtained the SARC dataset it contained insufficient clinical and demographic information to allow us to determine whether the SARC clinical sample was comparable to the mainstream mental health sample (e.g. presence of mental health needs prior to treatment, history of contact with psychiatric services). In addition, the SARC dataset had a small sample size with respect to the primary outcome, which meant it was not appropriate or feasible to compare outcomes between the datasets. For this reason, we undertook a within group analysis for the SARC dataset.

Clinical record interactive search analysis

To establish a cohort within CRIS that met our inclusion criteria, we developed initial search terms and subsequent methods for validating our searches. Details of this process can be seen in Appendix 4, Box 1. Through this process, we established a sample of 265 individuals who had received psychological therapy between January 2016 and December 2019, who had completed a pre- and post-therapy self-report outcome measure, and who had a documented sexual assault, and 257 individuals with no documented sexual assault but who met the other criteria (control).

Sample characteristics

The sexual assault group were significantly more likely to be female (76%) and significantly less likely to be of white ethnic origin (62%) than the control group (55% and 70%, respectively; Table 14). There was no difference in age distribution or in mean age but the those in the sexual assault group were more likely to live in more deprived areas than the controls.

Outcomes (validated)

Analysis of CORE assessments that were validated to have been associated with a period of psychological therapy showed that both pre- and post-therapy the sexual assault group had a significantly higher mean score (2.3 ± 0.79 and 1.59 ± 0.88, respectively) than controls (1.87 ± 0.81 and 1.22 ± 0.81; Table 15). The number of weeks between pre- and post-therapy COREs, the number of sessions delivered and the average number of sessions per week were also significantly higher in the sexual assault group than in the control groups. However, the average CORE change during therapy did not differ significantly between the two groups. Overall, 62% of sexual assault group showed a reliable improvement in CORE score with an average change of –0.71 ± 0.77, compared with 58% of controls showing reliable change with an average change of –0.64 ± 0.70.

Figure 7 shows the differences in clinical severity identified by the CORE assessments pre- and post-therapy between the sexual assault and control groups. There was a significantly greater proportion of people in the sexual assault group with severe psychological distress pre-therapy (45%) than in the control group (25%), whereas, post-therapy, the distribution of outcomes had shifted towards milder clinical categories, such that a significantly greater proportion of the control group were non-clinical (48%) than the sexual assault group (29%).

FIGURE 7.

CORE measures grouped by severity.

Clinical Outcomes in Routine Evaluation Outcome Measure or CORE-10 findings were grouped into levels of clinical severity: (A) pre-therapy assessments; (B) post-therapy assessments. In Figure 7, aqua bars show the percentage of the sexual assault group and purple bars show the percentage of the control group in each clinical category. Aqua squares show the change in percentage of the sexual assault group post-therapy in each clinical category. Purple circles show the change in percentage of control group in each clinical category. Pearson chi-square comparing the two groups for severe and not-severe pre-therapy (A) was p < 0·001 and for non-clinical compared with clinical post-therapy (B) was p < 0·001.

Clinical contacts/resource use

For all measures of clinical contact, the sexual assault group had significantly greater levels of service use than the control group (see Appendix 5, Table 17). That is, those in the sexual assault group were twice as likely (36% in total) to have had more than three spells of care than those in th control group (17%); they were nearly three times as likely to have had more than four CORE assessments during 2015–20 (32%) compared with the control group (12%); first contact with services was significantly earlier in the sexual assault group; the number of psychological therapy sessions attended in 2016–19 was greater in this group (mean 28.4 ± 13.7) than in the control group (mean 19.2 ± 21.3) and 62% of the sexual assault group received more than 20 therapy sessions compared with 33% of the control group; the year of first diagnosis was earlier in the sexual assault group and they received significantly more diagnoses (mean 7.7 ± 7.7) than controls (mean 4.3 ± 4.3); overall, 45% of the sexual assault group received four or more diagnoses compared with 22% of controls.

The full list of diagnoses across all year 2000–19 (see Appendix 5, Table 18) showed that the sexual assault group was more likely (23%) to have schizophrenia, schizotypal, delusional and other non-mood psychotic disorders than the control group (15%) whereas those in the control group were more likely to have mood (affective) disorders (31%) than the sexual assault group (21%).

Appendix 5, Table 19 shows that 82% of the sexual assault group had previously received at least one full risk assessment and this was significantly higher than in the control group (65%). A higher proportion of people in the sexual assault group compared with those in the control group scored positively for all the current and historic trigger risk factors recorded. These differences were statistically significant for triggers relating to relationship/family conflicts, adverse life events, historic drug and alcohol abuse and transitions in care. Furthermore, a higher proportion of people in the sexual assault group compared with the control group scored positively for most of the self-harm/suicide risk factors recorded. These differences were statistically significant for risk relating to previous suicide attempt/accidental overdose, historic acts of deliberate self-harm, alcohol/substance misuse, current deteriorating depression with distress and significant anniversary.

Health costs

We calculated the CORE-6D preference-based index and cost of therapy for patients who had CORE-OM questionnaire responses available (Table 16). As a result, this analysis is on a smaller sample than the main analysis as not all patients have CORE-OM responses and the CORE-6D cannot be calculated from the CORE-10.

The sexual assault group had a significantly lower preference-based index both pre therapy (the control group had an index 0.082 greater; 95% CI 0.022 to 0.142) and post therapy (control group index 0.101 greater; 95% CI 0.042 to 0.160) and significantly higher costs for validated (control group cost £490 lower, 95% CI –836 to –144) and all sessions (control group cost £707 lower, 95% CI –1144 to –270).

The sexual assault group had a particularly low preference-based index. There are few instances where the CORE-6D has been used, but a recent prison sample had a CORE-6D preference-based index of 0.75 to 0.713 on enrolment to a trial for common mental health problems,1 with people in prison with a common mental health problem more like community controls than sexual assault.

Individuals who had been sexually assaulted highlighted the importance of equity considerations in economic evaluations: even though there is evidence that they benefit from therapy, potentially as much as the control group, they have a lower preference-based index and cost significantly more than controls. As a result, treatment for this group is less likely to be cost-effective unless equity considerations are included in the analysis, potentially including a weighting for patients who had been sexually assaulted.

Discussion

This study is one of the first to directly compare the psychological treatment outcomes of psychiatric service users who have and have not reported a sexual assault. Our findings indicate that service users with experiences of sexual assault appear to have greater levels of mental health care need than those without experiences of sexual assault. Despite greater levels of pre-therapy psychological distress among those with experiences of sexual assault, similar post-treatment reductions in distress are observed compared with service users without experiences of sexual assault. Our findings indicate that service users who have experienced a sexual assault may be less likely to achieve clinical recovery post psychological treatment, when compared with service users who have not experienced sexual assault. These findings suggest that further mental health care may be needed for service users with abuse histories, and that more research is required on how to evaluate the cost-effectiveness of these services.

Strengths and limitations

Owing to limitations within the SARC clinical dataset, with respect to a lack of clinical and sociodemographic information and missingness on the primary outcome (i.e. psychological distress scores post treatment), we needed to deviate from some aspects of our original protocol. We were not able to directly compare this mainstream mental health dataset to the SARC clinical dataset, as we could not determine if the two populations were comparable and were unable to conduct the health economics analysis as we had planned. We, therefore, conducted a within-group analysis of the primary research question and looking at healthcare cost within this mainstream mental health dataset.

We found that the way in which sexual assault was documented in clinical records varied considerably (e.g. a lack of specification around the type, frequency or severity of violence experienced). This meant that we were unable to extract and analyse data on some assault-related variables that we had pre-specified in our protocol.

As we had access to one of the most well-established anonymised health records system in England, we decided to include an additional add-on analysis as part of our within-group comparison. We sought to explore and compare the characteristics and outcomes of those who had received psychological therapy in a mainstream mental health service and either had or had not experienced sexual assault. We were aware that there was an evidence base comparing the clinical and sociodemographic characteristics of these two groups, but an absence of data directly comparing the psychological treatment outcomes of mental health service users who had and had not experienced sexual assault.

Our inclusion criteria were restricted to people who had attended a psychological therapy session and had a recorded CORE outcome, which suggests that our sample had at least some level of engagement with therapeutic work. Our sample is not, therefore, representative of the entire population of interest. In addition, our cohort inevitably missed people who had experienced sexual assault but who had not been asked about this by clinicians or who had not had a disclosure documented in their clinical notes. In addition, completion of structured risk assessment tools was dependent on individual staff perceptions and behaviours about service users’ presenting risks, which resulted in missing data on this variable for several service users.

Finally, we were required to undertake multiple manual validation checks on key study variables to ensure correct data classification. Although these manual checks strengthen our methodology and confidence in the data it inevitably limits the sample size that we could include for the analyses, as we could only include a number that was feasible for us to manually validate.

Pre-sexual assault referral centre: lack of prior awareness of sexual assault referral centre and its functions

Most of our survivor interviewees were not aware of SARCs prior to their assault. Most heard about the SARC from reporting to the police and the police then accompanying them to the site. One person found out from a search on the internet. The lack of awareness of the SARC and what happens at a SARC could deter people from attending because they do not know that it exists or have concerns about not knowing what will happen there. This lack of awareness also featured in the partner agency interviews. There were some misconceptions from partner agencies as to what was offered at the local SARC, including one NHS mental health interviewee thinking that SARCs offered every level of psychological therapy for sexual trauma when in fact their local SARC only offered a limited number of sessions of supportive counselling. Some partner agency informants indicated that a mental health referral for therapy may be rerouted to the SARC if there was recent sexual assault rather than offer an assessment. Some SARCs had run open days prior to the pandemic so that external agencies could attend and learn about the nature of their offer. SARC staff recalled the value of being able to run external training to partner agencies but that due to having less time and resource, because of increasing footfall, this awareness raising work often did not happen.

There is a high prevalence of sexual assault experienced by people with mental health issues; however, very few mental health services have reported a pathway to SARCs following a disclosure of a sexual assault. 44 In addition, in WP5 we identified that, over a period of 10 years in one NHS mental health trust alone, 3000 women who had received some form of psychological therapy input (as indicated by CORE-10 outcome data) had a case note indication that they had experienced sexual assault. It is not known whether the psychological therapy was because of the impact of sexual assault or other reasons and, even if it was related to the sexual assault, whether it was the correct therapy (i.e. trauma therapy). It is also not known whether they had any contact with the local SARC. A free-text search was conducted for any reference to SARCs and there was an absence of any references to attendance or referral to a SARC in the case notes which could indicate a lack of awareness of attendance on the part of the mental health staff, or the group that had experienced had not reported it officially.

If SARCs were able to promote the work of the SARC, by discussing what a SARC is and what it offers (mechanism) with general population, as well with key local organisations that work with underserved groups such as mental health, substance use and so on, then this would mean that more people were aware of their existence as well as knowing in advance what the SARC offers and therefore there would be more uptake from underserved groups (outcome) and improvements in the working relationships between SARC and partner agencies (outcome).

At the sexual assault referral centre

The survivors consistently spoke highly of the non-judgemental and compassionate care they received from initial phone contact right through to follow-up contact. This was noted across all the SARCs where survivors attended. SARC staff interviews also discussed examples of compassionate care that they delivered. The survivors appreciated working with staff who ‘got’ trauma as opposed to some less positive experiences they had at other agencies (such as NHS mental health and psychological therapies). The context of trauma informed interactions may well be boosted by the presence of a mental health professional such as the clinical psychologists in SARCs 1 and 5 who can bring this into internal training and team discussions.

If SARCs and other relevant agencies adopt a trauma informed approach (context) using trauma informed interactions such as believing the survivor, non-judgement and compassion, dispelling victim blaming myths, offering choice and control (mechanism), then this improves engagement with that service (outcome) and potentially aids recovery (outcome).

The forensic examination

The systematic review identified studies that have evaluated the impact of giving people prior information about what would be involved in a forensic examination and found that this intervention reduced distress. In the interviews, some survivors expressed that the forensic examination was the thing they most dreaded, yet they felt that the actual experience was not as bad as expected because they had been fully briefed prior to the exam by the SARC staff. The other important issue was giving the survivor choice and control both before the examination and during which is something they had not had during the assault itself. The assumption that could be made is that prior information and giving choice and control serves to prevent re-traumatisation.

If survivors are prepared in advance for the forensic examination (by a priori briefing) and given choice and control (mechanism, and have someone accompany them to provide emotional support; e.g. SARC crisis worker) by staff who understand trauma and able to use a trauma-informed approach (context) then they can cope much better with the experience and the experience is less likely to exacerbate trauma (outcome).

Assessment and response to mental health and substance use

Within the systematic review and national survey there was variations identified in the approach to assessment. Mental health assessment is conducted using professional opinion following an interview. It is also important to note that from the data including the national survey and staff interviews, many of the people who work in the SARCs including the crisis workers and forensic examiners do not have a background in mental health and so the issue of consistency and quality of ‘professional opinion’ in relation to assessment of mental health is also in question. In SARCs that had a clinical psychology presence, there was evidence of the use of screening for PTSD, as well as a triage assessment of need led by the clinical psychologist, in conjunction with the rest of the multi-disciplinary team. Having a mental health practitioner in the team also offered a secondary benefit in advising others in the team on the impact of trauma (including issues of vicarious trauma among staff) and more generally about mental health issues experienced by service users.

If survivors receive an assessment of mental health needs (mechanism) that is informed by having people with specific mental health expertise within the team (context), then their needs are more accurately identified, and they can therefore be referred to the right care at the right time (outcome).

There was limited evidence of the use of screening tools, and some suggestion that screening tools were too ‘black and white’ and felt impersonal like a ‘tick-box exercise’. Some responders did feel that using standardised screening tools could support referrals, serving as objective evidence of level of need. In WP3 a suite of mental health screening tools was administered in the form of an online survey to those who had presented to the SARC after a sexual assault no earlier than a week after attendance and no later than 6 weeks of attendance. Of those who consented to participate, all completed the online questionnaires (see Chapter 3). The respondents were broadly representative of the population of people who attend SARCs nationally. Most of the participants had previous contact with mental health services prior to attendance at the SARC. Most (94%) scored positive on the trauma checklist. It was estimated from 2017–18 SARC service use data that of 14,357 people who used a SARC over one year; 13,584 were likely to score above threshold for PTSD. NICE guidance indicates that where someone has been assessed as having PTSD after a trauma then they should be offered trauma therapy within a month. However, based on the findings from the review, the national SARC survey and interviews in the case studies, routine screening for PTSD is not currently routine practice in SARCs.

We only found two SARCs that screen for PTSD and the south-east SARC has a specific triage system where the person is discussed as a team and placed either in a NGO counselling service or if more complex, referred to the therapy service provided by the local mental health service (WP5) From the analysis of the routinely collected data at this service, we noted that the levels of severity of baseline symptoms (CORE) were higher for those who referred to the mental health provider service compared with the third sector supportive therapy, but the two groups of survivors did equally well as a result. While there are limitations of analysing routine data from a small data set it is a promising finding and suggests that future research is needed to evaluate the benefits of global screening for PTSD and stepped care interventions for survivors of sexual trauma.

If SARCs adopt the use of standardised screening tools for PTSD (context) as a part of their assessment (mechanism) then this would ensure that people were effectively identified (outcome), provided consistency of practice across SARCs in the identification of PTSD (outcome) as well as identifying those who require evidence-based therapies and therefore improve their mental health (outcome).

In terms of continuing psychological support, many people who came through the SARCs often only accessed in-house counselling, external counselling or were referred to a variety of mental health care. Some SARCs can and do facilitate referrals to relevant external agencies as they have good knowledge of the range of services available, as well as their referral criteria. However, in some SARCs who cover a wide area that also involves several mental health and substance use care providers (such as in Yorkshire, London and Manchester), it is less clear and transparent about how to refer when each service has a different offer and routes into treatment. SARC staff noted that it is sometimes easier to connect a person to mental health support. We found that in SARCs where there was a lack of mental health expertise in the team, this impacted on how effectively and confidently the team identified and responded to mental health and drug/alcohol needs. It was also noted that when a SARC offered more than simply signposting people to external services (such as facilitating a referral and checking in on progress and outcomes of the referral) that this lead better engagement with that service as survivors often felt overwhelmed with information or too daunted to instigate a self-referral.

IF SARCs have good relationships with external agencies (context) and therefore more clarity around referral criteria and processes, then the referral process (mechanism) would run effectively and therefore survivors would receive more timely and appropriate aftercare.

Nature of the needs of people who attend sexual assault referral centres

It is firstly important to state that the trauma response to a sexual assault is normal and expected and can manifest in people in a myriad of ways. This trauma response should not be immediately pathologised and ‘diagnosed’ as there is a risk of imposing potentially damaging and stigmatizing psychiatric labels on people. In many cases of trauma, a significant percentage of people do not go on to develop PTSD or other mental health conditions. 43

Previous research has tried to identify predisposing risk factors for PTSD and other mental health issues following a sexual trauma but failed to find specific factors that correlate with this and Dworkin and colleagues43 suggest that it is more useful to explore the multilevel social context of survivors and the impact of these on recovery. Many survivors experience feelings of shame and self-blame after an assault in part due to internalising societal ‘rape myths’ which leads to a reluctance to seek help or report the assault. SARC services have shown that their trauma informed approaches have a powerful and positive impact on the experience of the survivor and their future recovery. Indeed, the survivors in the interviews talked about the importance of being believed, as well as being told that they were not to blame for what had happened. These seem to be really key messages that promote the start of recovery. Some of the survivors had also had contact with mental health services and described receiving a less trauma-informed approach including feeling judged and blamed for the assault and stated that it made a real difference if they worked with someone who really understood sexual assault and its impact including psychological therapies and/or counselling delivered at a SARC or counselling delivered by a sexual violence service specifically. An accurate assessment of need can only be useful if it opens the door to someone getting the kind of help that will support them to move forward with their life.

In terms of thinking about mental health needs, it is important to distinguish between those who attend a SARC who have prior mental health needs as well as those with no prior mental health needs. We have identified a third group who struggle to find the right care at the right time. They present with mental health and/or alcohol/drug issues needs that are perceived by SARCs (if they provide in-house counselling) as too ‘complex’ for the in-house support (if it exists), and not meeting the threshold for secondary mental health or NHS psychological therapies, or they face long waiting lists to get help. There were examples of SARCs struggling to access crisis mental health support if they had immediate and significant concerns about the safety of a person and often the only option was to take someone in acute distress to the local emergency department.

In WP5, those with a history of assault presented with higher levels of psychological distress at start of therapy as well as scoring higher on risk assessments especially in relation to self-harm and suicidality. However, the group with a history of sexual assault did just as well in treatment as the comparison group. While it is unclear how many people were accessing therapy in relation to the sexual assault, it does indicate that while those who have experienced sexual assault experience significant mental distress, accessing therapy can improve their symptoms. However, owing to having more complex needs and severity of distress at baseline, the people in the sexual assault group tended to require more sessions of therapy. Any assessment of cost-effectiveness should address this to ensure that there is equity in comparing different groups.

Models of care

The cluster analysis performed using data from the survey served to identify three ways of working in relation to the responses to the mental health questions in the national survey. This informed the sampling strategy and ensured we had SARCs that represented the three clusters. However, it was clear from the richer case study data that the clusters lacked specificity and accuracy. The survey was completed typically by a SARC manager who may not have all the specific details on how mental health was managed at the micro levels. Overall, it became clear from the case studies that a more sensible grouping of SARCs were ones that had in-house clinical psychology (SARCs 1 and 5) which meant that survivors were reviewed and triaged in discussions led by an experienced mental health professional (clinical psychologist). Survivors appreciated receiving therapy in a setting with expertise in trauma related to sexual assault. The second model of care was one SARC (SARC6) that provided in-house counselling. This was provided by trained counsellors who were not necessarily mental health practitioners. Unfortunately, we were unable to recruit survivors from this SARC, so we are unable to present the survivor perspective on that service. The third group of SARCs did not have any mental health provision in the team, and variable relationships with relevant external agencies. It was notable that where there was a lack of mental health expertise in the skill mix, the staff identified this as a deficit and one SARC (SARC2) is planning to recruit a mental health nurse to work in the SARC and be able to take on the liaison role with the local mental health services.

Contexts and mechanisms facilitate identification, care and treatment for mental health and substance use issues within and beyond sexual assault referral centres

Several contextual issues were identified: the presence or not of a mental health professional in the SARC, the quality of relationships with external agencies and capacity within the local therapy provision. Contextual issues identified clustered around the presence of a mental health professional within the team. This had the benefit of helping assess need and triage survivors according to that need, developing the ethos of trauma-informed care in the team by role modelling, informal teaching and within multidisciplinary team discussions. Another important contextual issue was the quality of external relationships with external agencies. If there were positive communication channels (usually a single point of contact) then it facilitated the referral process. Where these relationships did not exist, the referral process became confusing and fragmented and often led to the survivor not getting their needs met. The third important contextual issue was whether there was local capacity in the systems to be able to provide timely and appropriate therapies. Waiting lists and limited sessions were mentioned by both SARC staff and survivors as contextual issues that inhibited recovery and added to the stress felt by both the SARC staff (who felt responsible for the continuing well-being of the people that they saw but also felt that it was beyond their role and expertise to offer continuing mental health support) and for the survivors who struggled to advocate for themselves due to feeling overwhelmed by what they were going through, and felt rejected by the system if they didn’t meet the referral criteria or experienced a long wait for treatment.

Impact of COVID

As with all other NHS research, the MiMoS study was impacted in 2020–21 by the COVID-19 pandemic. Specific details of the ways in which aspects of the study have already been addressed. As the data were collected during unprecedented social change (lockdown and social distancing measures) and there were changes to how services can be delivered (remote and home working, use of telephone and online appointments) the data may be additionally impacted by these issues.

Validating that the sample met our criteria

The presence of sexual assault terms in the free-text fields did not necessarily mean that the individual had experienced sexual violence. For example, these terms were present in cases where the individual was a perpetrator, or where another family member had been a victim of sexual assault. Therefore, a random sample of 520 sexual assault cases (approximately 20%) were manually checked to establish whether the classification was correct.

Of the 520 sexual assault cases, 26 cases were excluded where the notes/correspondence could not be accessed (too long or corrupted) and five cases were excluded where they had been extracted due to a typographic error (e.g. ‘neccsarry’, ‘hecsays’, ‘medicsal’). Of the remaining 489 sexual assault cases, 305 (62%) were correct. Owing to the substantial number of false positives (38%) and the fact that it was not feasible to manually check all 2555 cases, the 305 validated cases were used as a representative sample for further analysis.

One hundred cases without a documented sexual assault term were also checked for accuracy. Of these, 98 (98%) cases were correct. Owing to the high percentage of correct classifications, further checks were not undertaken on this sample. Excluding the two cases identified as incorrect, a random sample of 305 individuals (including the 98 checked for negative validity) were selected from the 2456 non-sexual-assault group to act as a control sample.

Our selection criteria included those cases that had more than one CORE measure, one of which was a last session/follow-up assessment and had a face-to-face psychological therapy attendance within the specified dates. However, these data needed further interrogation to identify whether the CORE measures represented pre- and post-therapy scores related to the psychological therapy attendance. For example, in some cases, CORE assessments were recorded on dates that were in a different year to the documented therapy attendance. Therefore, further manual validation was undertaken on the 610 cases to identify those cases with an episode of therapy that had associated outcome measures. During this validation, the number of therapy sessions attended within the pre- and post-CORE date range was counted, providing a measure of dosage.

Of the 610 cases, 522 (85%) had identifiable episodes of therapy with associated CORE measurements – 265 sexual assault and 257 no sexual assault. This therefore represented our final sample for subsequent analysis.