Drug Use in Street Sex worKers (DUSSK) study: results of a mixed methods feasibility study of a complex intervention to reduce illicit drug use in drug dependent female sex workers

Objectives The majority of female street-based sex workers (SSWs) are dependent on illicit drugs and sell sex to fund their drug use. They typically face multiple traumatic experiences, starting at a young age, which continue through sex work involvement. Their trauma-related symptoms tend to increase when drug use is reduced, hindering sustained reduction. Providing specialist trauma care to address post-traumatic stress disorder (PTSD) alongside drug treatment may therefore improve treatment outcomes. Aims to (1) evaluate recruitment and retention of participants; (2) examine intervention experiences and acceptability; and (3) explore intervention costs using a mixed methods feasibility study. Setting Female SSW charity premises in a large UK inner city. Participants Females aged 18 years or older, who have sold sex on the street and used heroin and/or crack cocaine at least once a week in the last calendar month. Intervention Female SSW-only drug treatment groups in a female SSW-only setting delivered by female staff. Targeted PTSD screening then treatment of positive diagnoses with eye movement desensitisation and reprocessing (EMDR) therapy by female staff from a specialist National Health Service trauma service. Results (1) Of 125 contacts, 11 met inclusion criteria and provided informed consent, 4 reached the intervention final stage, (2) service providers said working in collaboration with other services was valuable, the intervention was worthwhile and had a positive influence on participants. Participants viewed recruitment as acceptable and experienced the intervention positively. The unsettled nature of participant’s lives was a key attendance barrier. (3) The total cost of the intervention was £11 710, with staff costs dominating. Conclusions Recruitment and retention rates reflected study inclusion criteria targeting women with the most complex needs. Two participants received EMDR demonstrating that the three agencies working together was feasible. Staff heavy costs highlight the importance of supporting participant attendance to minimise per participant costs in a future trial.


INTRODUCTION
Most female street-based sex workers (SSWs) in the UK use heroin and/or crack cocaine. [1][2][3] Their drug dependency keeps them entrenched in a 'work-score-use cycle', 4 5 which contributes to the morbidity and social instability typically seen in this group. 6 Despite their drug treatment needs, drug dependent SSWs have poorer outcomes from drug treatment services compared with other service users, 7 8 sometimes due to stigma associated with their street sex work. 9 Previous SSW-focussed interventions aiming to reduce drug use have used educational, 10 11 substitute prescribing-based 12 13 and psychological 14 approaches but none robustly demonstrated a positive effect in reducing drug use. 15 Poor mental health is a significant problem among SSWs. [16][17][18] Many have experienced Strengths and limitations of this study ► The novel intervention integrates a trauma-focussed treatment approach in order to reduce drug use in a challenging drug treatment population. ► The intervention was delivered by specialists, reflecting the skills and experience required to appropriately manage the complex needs of the study population. ► Patient and public involvement formed an important part of the study methodology and informed each stage. ► Recruitment took place over several months within an agency trusted and used daily by the study population to allow familiarisation with the researchers and multiple opportunities to participate. ► This feasibility study design and methodology was not able to examine intervention effectiveness or cost-effectiveness.
Open access multiple adversities in early life and during their involvement in sex work, 5 16 19 which exposes women to further risk of significant trauma. 16 17 Consequently, many SSWs are affected by post-traumatic stress disorder (PTSD). 16 17 Trauma symptoms, which often recur when drug use is reduced, may motivate a return to drug use. 20 Individual trauma-focused therapy alongside drug treatment may provide the best outcomes for reductions in drug use. [21][22][23] However to date, there is no robust evidence to demonstrate the impact of an integrated trauma-focussed treatment approach in reducing drug use among female drug dependent SSWs.
In collaboration with SSWs and service providers, and informed by existing research, 9 15 we developed a novel intervention, to simultaneously address the unique and complex combination of drug use and PTSD in female drug-dependent SSWs. The intervention proposes an integrated care pathway through an innovative multiagency partnership. 24 We report here the results of the Drug Use in Street Sex worKers (DUSSK) feasibility study, which aimed to (1) evaluate the recruitment and retention of SSWs to the intervention; (2) examine the experience and acceptability of the intervention for participants and service providers; and (3) explore costs to service providers associated with the intervention.

METHODS
Study design, setting and eligibility Detailed methods are described in the published protocol. 24 This mixed methods feasibility study took place in a UK inner city setting. Females aged 18 years or older, who sold sex on the street at least weekly in the last calendar month and used heroin and/or crack cocaine at least once a week in the last calendar month, were eligible to participate. 24 The intervention was delivered at SSW charity premises, which supplied support, health and advocacy services. 24

Recruitment
The recruitment target for this feasibility study was 30 women. 24 Local organisations that SSWs were known to access were provided with study promotional materials. One of three researchers (JMK, NJ, SR) attended an SSW support and advocacy charity, at least twice a week, to directly approach potential participants; alternatively, interested SSWs could telephone researchers. This approach meant that women were potentially approached and counted as contacts several times during recruitment. Eligibility screening was conducted face-to-face or via telephone. Women gave fully informed, written consent to participate in the study and baseline data were also collected at the time of recruitment. To maintain safety and confidentiality, each participant provided details of acceptable ways in which to be contacted. Screening data were retained and remained anonymised for those not recruited.

Patient and public involvement (PPI)
Women with experience of street sex work and drugdependency took part in focus groups and one-to-one discussions with NJ to inform study design, processes, documentation and intervention development. On commencement of the study, a group of women who were ineligible for recruitment were approached for involvement in PPI. They addressed challenges with recruitment, participation and adherence issues (described below) and suggested solutions, which were implemented. For example, they recommended changes such as provision of sandwiches to improve attendance.

The intervention
The intervention consisted of SSW only drug treatment groups, targeted screening for PTSD symptoms (oneto-one clinical interview and PTSD Checklist (PCL5)) 25 and, if positively diagnosed, one-to-one eye movement desensitisation and reprocessing (EMDR) therapy, all delivered by female staff through a collaboration between three service providers (National Health Service (NHS) trauma services, the SSW and drug treatment charities). The intervention was designed so participants were initially invited to attend a weekly 'Getting started' drug treatment group to reduce fear and anxiety about engaging in a group setting and get used to the format and level of disclosure expected. Participants were to progress to a 'Preparation for recovery' drug treatment group which focused on people's barriers to motivation for change, examining pros and cons of drug use and exploring triggers for using drugs to enable participants to achieve a level of drug use stability. As stated in the protocol 24 and in line with provider's usual care, participants were perceived as demonstrating drug use stability by exhibiting evidence of life/drug use stability such as engagement and functioning in the group, positive interaction with group facilitators and regular opioid substitution therapy (OST) by the group facilitators. When group facilitators judged participants were achieving drug use stabilisation, and they had attended three sessions consecutively, they were offered screening for PTSD symptoms by a female clinical psychologist. Those experiencing PTSD symptoms were invited to attend five PTSD 'Stabilisation' group sessions, facilitated by the same female clinical psychologist, to equip participants with the skills to selfsoothe and reorientate in preparation for the one-to-one EMDR treatment. Once all 'Stabilisation' group sessions had been completed, the clinical psychologist assessed participants for readiness for one-to-one EMDR sessions and if eligible, participants progressed to a course of 12 sessions with the clinical psychologist on a weekly, or fortnightly, basis. Trauma treatment (PTSD screening, stabilisation groups and one-to-one) ran in parallel to the drug treatment groups; figure 1 (red boxes) shows the planned flow of participants through the study.
The intervention proceeded as described in the protocol paper 24 with the following changes (bulleted next):  Figure 1 Flow of participants through the Drug Use in Street Sex worKers study. PTSD, post-traumatic stress disorder; SSWs, street-based sex workers; EMDR, eye movement desensitisation and reprocessing therapy ► The number of sessions continued beyond those initially planned due to delays in PTSD screening (see below). ► Poor attendance, the participant's unstable behaviour and the intervention running for longer than planned (due to the retendering of drug services) affected the drug group facilitators' ability to deliver structured content. An art worker was included in four sessions to maintain participant's interest and engagement with sessions. ► PTSD screening was halted from the 29th session onwards, due to the limited remaining study time to complete the intervention, meaning the drug group became closed and only included PTSD screened participants.
Screening for PTSD ► Delays occurred to both the organisation of the PTSD screening and PTSD stabilisation group set up due to lack of capacity and delays in NHS trauma service staff recruitment. Therefore, some participants had long gaps between recruitment and PTSD screening and subsequent stabilisation group sessions.
PTSD 'Stabilisation' group ► Incentives of £10 vouchers per session were offered for attendance (mandatory sessions).
One-to-one EMDR therapy for PTSD ► Some sessions were scheduled in a private rented room (due to availability issues) in a local community centre and not at the SSW charity premises. ► Two participants were offered weekly one-to-one extended stabilisation sessions (eight maximum) with the clinical psychologist rather than EMDR therapy as this was deemed the most appropriate treatment.

Data collection methods
Sample size and quantitative data collection A formal sample size calculation was not conducted as the aim was to assess feasibility. 24 Self-reported levels of illicit drug use, involvement in SSW, completion of PTSD Checklist PCL5 25 and demographics were collected at the time of consent. Attendance registers were taken at the start of each group or one-to-one session by the facilitator(s).

Qualitative data collection
With participants' verbal consent one non-participant observation of a drug group was conducted to understand delivery, examine interactions and intervention experiences, with brief notes taken during the group. 26 In-depth semistructured interviews were conducted with participants and service providers either face-toface or by telephone. Participants were interviewed after intervention completion or study drop out. Consent to contact participants regarding interviews was sought at recruitment. Additional written or audio recorded verbal, informed consent was obtained prior to all interviews.
Interviews explored views and experiences of the intervention and how to improve acceptability. Participants received a £20 high street shopping voucher for taking part. Most service provider interviews were conducted at the end of the intervention period and also sought to understand operational issues, interagency working and intervention delivery.
Economic data collection Resource use information (2018 £GBP) was collected prospectively by the agencies. Data collection focused on four categories: staff time, facilities, travel (provider funded staff and patient transport) and materials.
Non-attendance was dealt with as follows; if no participants arrived after 45 min for a 'getting started' session, staff left and were only costed for the time that they spent waiting. Staff delivering 'trauma screening', '1-1 s' and 'stabilisation groups' were costed for all sessions booked, regardless of non-attendances.

Data analysis
The integration of qualitative and quantitative data used the established 'following a thread' technique 27 where key themes were traced using all data sets.

Statistical analysis
Descriptive statistics (using Stata v.14) were used to monitor recruitment and retention via CONSORT statement 28 style flow charts, and to examine participant demographics, questionnaire responses and attendance.

Qualitative analysis
Interviews were conducted by JMK and NJ, and the audio files transcribed, anonymised and checked for accuracy. QSR NVivo v.10 software was used to perform inductive thematic analysis 29 using constant comparison techniques. 30 31 A preliminary coding framework was developed by JMK and discussed with the multidisciplinary research team, JH and NJ to ensure credibility and confirmability.

Cost analysis
Costs to the service provider were examined and summarised separately for each of the four intervention components. Staff costs were calculated using salaries and oncosts or generated using standard unit cost data available for health and social care professionals. 32 Facility costs were calculated based on similar space rental options. Total cost, total cost per eligible participant, total cost per session held, and total cost per session per eligible participant were calculated for each intervention component. Cost data were tabulated using principles of heat-map methodology, where colour lightness is used to communicate the magnitude of different costs. 33 34

RESULTS
Recruitment was from November 2017 to March 2018 with the intervention delivered until December 2018.

Recruitment process
Approximately 400 flyers and nine posters were distributed. Potential participants were spoken to by three researchers at 37 3 hour 'drop-in' sessions at the SSW charity. Of 125 contacts made with women, 84 declined screening. Reasons for decline included being too busy (n=13), not interested (n=11) or reporting not currently street sex working (n=10). Fourteen approaches were reported as repeat approaches, with seven reporting previous recruitment and four previous screening (as the researchers recruited on different weekdays and screening data was anonymous). Figure 1 details the flow of screened and recruited participants through the intervention.
Of 41 women screened, 11 were eligible and consented to participate, 3 were eligible but unable to give consent (2 were too distressed and 1 had health issues preventing participation). Of the 27 ineligible women, two did not fully complete the screening questions. The main reasons for exclusion related to ineligible frequency of drug use and/or sex work. Table 1 shows the range of days since responders last street sex worked or had taken heroin/ crack by those recruited and not recruited.
Of 11 participants consenting to be invited for qualitative interviews, 5 were uncontactable. Seven interviews were conducted with six participants (six face-to-face and one via telephone); four interviewed participants received all components of the intervention. Ten service provider interviews were conducted with representatives from the drug treatment service (n=4), the trauma service (n=2), and the SSW charity (n=4). Table 2 details quotes from the interviews to support the main results.

Acceptability of recruitment
The recruitment process was described as acceptable by most participants and service providers. Face-to-face recruitment was experienced as confidential, with participants reporting receiving clear explanations of the study. Recruitment over 5 months within the SSW charity offered multiple opportunities to participate and was acceptable. Most service providers interviewed reflected that if study inclusion criteria were broadened to include less regular drug use, participants with more lifestyle stability could have been recruited, and whom may have found the intervention easier to engage with. However, some felt that creating groups with diverse levels of drug use could have negative consequences for example, risk of relapse for those who had reduced their drug use.
Group attendance All 11 consented participants were invited to attend drug treatment groups. However, participants attendance varied throughout the study, with participants sometimes arriving late or leaving early (table 3). Four attended 30%-76% of sessions compared with 7 attenders who attended only 0%-18% of sessions. The five most frequent attenders were invited to PTSD screening of which four attended and were all found to have symptoms of PTSD. All four PTSD screened participants attended 20%-100% of the stabilisation groups with the clinical psychologist. At the end of the stabilisation groups, two participants were deemed suitable for EMDR therapy by the clinical psychologist and two were offered extended stabilisation sessions. All four participants attended some one-to-one sessions (table 3) but missed at least two consecutive trauma treatment appointments with the clinical psychologist and had to withdraw from the sessions. However, all participants were referred to further mental health services and one participant was also referred to mainstream drug services.

Facilitators to attendance
Service providers across all partner agencies sent reminders to participants, which were described as helpful and appreciated by participants. One SSW charity service provider played a vital role in encouraging attendance through reminding participants to attend, arranging transport (taxi, bus or driving participants to sessions) and helping participants prepare for the intervention. Provision of sandwich lunches before the groups was seen by service providers and participants as helpful for encouraging attendance, facilitating a relaxed start to groups and supporting concentration. Vouchers were also viewed as encouraging attendance by participants and service providers.

Barriers to attendance
The unsettled nature of participant's lives was perceived as an attendance barrier and was underpinned by problematic drug use, poor adherence to OST, sex work, tiredness and poor mental health. Arguments between participants, a lack of readiness to confront issues with drugs and trauma and an absence of social support were described as making attendance difficult. Delays to screening referral and trauma treatment were also reported to negatively affect participants' motivation.
Experience and acceptability of the intervention Initial impressions All participants perceived the intervention as valuable and welcome. Common reasons given by participants for taking part included the opportunity for 'change', greater stability and valuing the opportunity to combine mental health and drug treatment. Service providers viewed the intervention as a novel opportunity for (1) SSWs to receive mental health treatment while continuing to use drugs and (2) to address the barriers to mainstream drug treatment. Some service providers highlighted the challenge for participants to process trauma while continuing to use drugs and potential risks of sex working when receiving trauma treatment.

Service providers' views on the intervention
The drug group facilitators described enjoying delivering the groups and building good relationships with participants. However, they described some drug sessions Open access as intense and difficult to manage due to participants' distress, accounts of trauma and chaotic behaviour. The need for appropriate support and supervision of facilitators was highlighted as a requirement to manage these challenges.
A clinical psychologist suggested that without the 're-traumatising' effects of street sex work, the effectiveness of the trauma processing in the trauma treatment might be enhanced. Service providers also proposed extending the stabilisation work to develop the effectiveness of the trauma treatment and recommended the intervention offer alternatives to EMDR to suit individual participants' needs.
Service providers said working in partnership with other specialist services to deliver the intervention was valuable, there was mutual respect and good communication   Open access between staff. It was suggested it would have been useful to have collaborative, regular case-review meetings between the services to assess the progress and needs of the participants and enhance the communication channels.
Participants views on the intervention Participants described generally positive intervention experiences. They described forming meaningful relationships with the drug group facilitators and clinical psychologist. They liked that the groups were female and sex worker only, they knew other participants already and could speak openly about, and relate to, one another's experience of trauma, drug use and street sex work. Participants also valued that the intervention was delivered at SSW charity premises, which was liked for its familiarity, safety, comfort, convenience and freedom from judgement and shame. The day of the week, time and frequency of sessions, drug group session length and group size were also acceptable to most participants and service providers. These factors overcame some of the barriers participants highlighted to attending mainstream drug services.

Strengths of the intervention
Through the intervention, participants reflected on their need to address their trauma and drug use. Some acknowledged that they were not ready to address their trauma but aspired to this in future, having had positive experiences of therapy during the intervention. Participants attributed improved well-being, coping strategies and perceptions of self-worth to the intervention. One participant was seen less on the SSW outreach van (an indicator of sex working) and, significantly, stopped using her working name, signifying 'taking back ownership of who she is'. One drug group facilitator felt the flexibility of the study intervention was able accommodate participants' unstable lives and level of trauma that would have prevented them from complying with the rules of conduct in mainstream drug services and thus prevented them from receiving treatment to address their needs. They also felt that an additional positive feature of the intervention was participants being able to discuss their sex work, due to the female SSW only membership of the groups. Participation in the intervention was described by SSW charity service providers and two participants as supporting and empowering the participants to engage with clinical and support services to address their needs. Another participant felt she used the SSW charity services less now because she needed less support.

Cost analysis
The total cost of the intervention was £11 710, with staff costs being the largest component (table 4). The most expensive component of the intervention was the 'Getting started' sessions (which totalled £6842). Despite having the second to lowest subtotal cost across the intervention (£1014), the stabilisation groups had the highest cost per session held (£203). Although the one-to-one sessions had the lowest cost per session held (£103), the larger number of sessions at this point resulted in this section having the highest cost per eligible participant (n=4, £724). Trauma screening had the lowest cost per eligible participant (n=5, £191)

Fidelity
The intervention was broadly delivered as intended incorporating suggested planned changes to the protocol; it was more flexible and less linear than originally planned. 24 Delays in PTSD screening meant that there was only a single drug group, which continued for longer than originally planned.

Summary of findings
This study used a mixed methods approach to investigate the feasibility and acceptability of a novel, complex intervention to reduce illicit drug use in female drugdependent SSWs. We demonstrated that drug-dependent SSWs could maintain attendance at female SSW-only drug group sessions and the integrated trauma-focussed treatment approach, in a trusted and supportive environment, with intensive support from SSW charity staff. Recruitment was lower than anticipated, with four participants PTSD screened and whom met criteria for PTSD. They progressed through to the final stage of the intervention; all four participants were ultimately able to access mental health services and one began the process of accessing mainstream drug services. Participants and service providers mostly experienced the recruitment process, the intervention and delivery mechanisms (especially the SSW-only environment) positively. Managing SSW trauma disclosure proved challenging for drug group facilitators

Open access
and non-clinical staff and resulted in the recommendation that there is additional training and support for staff in future studies. The need for intervention refinement, for example, extending stabilisation sessions, was suggested to provide additional support prior to trauma treatment. Attendance and adherence barriers primarily related to the issues the intervention sought to address, namely problematic drug use, sex work, and poor mental health, rather than the acceptability of the intervention itself. The total cost of the intervention was £11 710, with staff costs dominating.

Strengths and limitations of this study
Strengths of our study were that recruiting researchers also conducted the interviews, which may have facilitated a rapport with participants and supported more open and honest reflections. Second, all aspects of the intervention were delivered by specialists which was necessary for this high-risk group with multiple complex comorbidities. Third, PPI formed an important part of the study and informed each stage. Fourth, recruitment over several months with multiple approaches allowed SSWs many opportunities to take part. This approach took account of the rapidly changing lives of SSWs resulting in changing eligibility status as well as allowing time for them to become familiar with the researchers. Finally, recruitment within a trusted agency may have had a positive influence on recruitment.

Open access
This feasibility study provides only preliminary information on the intervention performance and cost and does not examine the effectiveness or potential for reducing costs in other parts of the health service or wider society. Delays due to changes to service provision are likely to have adversely influenced study recruitment and retention, with delays resulting in higher service costs, but reflect the real-life issues facing multi-agency work.

Comparison with other research
This feasibility study is the first, to our knowledge, to address previously highlighted barriers to effective drug treatment for SSWs. 15 22 23 Through incorporating female SSW-only drug groups alongside an intervention with specialised trauma treatment, delivered in a female SSWonly setting by female staff 22 24 we showed how an integrated treatment approach in this complex vulnerable group could feasibly be implemented and delivered, with changes to the intervention, although at a higher than expected cost, mostly due to the delays incurred due to service retendering.
This study is the first interventional study to employ clinical staff from a specialist trauma service to deliver EMDR to address trauma symptoms as part of the drug treatment process for SSWs. These women have been found to have high levels of poor mental health, 16 18 particularly trauma, 5 16 17 which contributes to poor drug treatment outcomes. 7 8 The intervention took account of SSWs frequent experience of abuse and violence, 35 36 and recommendations for female-only trauma-focussed drug treatment interventions 22 37 38 for treatment of PTSD and long-term drug use reduction. Some of the characteristics the intervention sought to address presented as barriers to attendance and retention; however, these are common in studies trying to affect behaviour changes within vulnerable groups. 23 39 Previous SSW-focussed interventions aiming to reduce drug use [10][11][12][13][14] were unable to demonstrate strong evidence of a positive effect 15 suggesting the need for a novel approach with evidence of efficacy assessed through a robust methodological approach. The highlighted barriers to attendance, engagement and delivery of the intervention are in keeping with other studies 15 23 39 but indicate that further changes to the DUSSK feasibility study design are likely to be required in future studies.

Conclusion and implications for service provision and research
This study sought to explore the feasibility of delivering a novel complex intervention to a very challenging population with high levels of unmet need. Inclusion criteria were informed by PPI, clinical expertise and academic literature. They targeted women who were likely to benefit the most from a trauma-based intervention but whose drug dependency and chaotic lives made them challenging study participants to recruit and retain. However, all four of the participants screened for PTSD were diagnosed, revealing the unmet need for trauma treatment. Though unsurprising, the severity of trauma disclosed by SSWs proved unexpectedly challenging for service providers. 22 37 38 Further data to understand the extent and severity of PTSD in SSWs are recommended to inform service provision. Overall, the experiences described by those receiving the intervention suggest that it is an acceptable approach to reducing SSWs drug use.
The three services found the intervention valuable and were able to work together effectively despite setbacks such as changing contracts and service pressures. They also suggested more staff support for managing trauma disclosure, extended stabilisation sessions and closer working could improve intervention delivery.
Intervention costs were driven up by poor participant attendance, though staff pressures and the retendering process increased the length (and cost) of the intervention period. However, decreased SSW use of health services, the criminal justice system and impacts of criminal activity on wider society may justify its adoption if future trials demonstrate intervention effectiveness in reducing drug use.
In order to support future interventional trials in this important field where there are few effectiveness studies we recommend the following study refinements for consideration. 1. The intervention could also include women with more stability in their lives to increase recruitment and retention. 2. Regular meetings throughout the study enabling all service providers involved in intervention delivery to express concerns and seek to understand participants needs from the perspective of different professionals so there is effective multiagency support for individual participants where needed. 3. Training for all involved staff in managing the disclosure of trauma. 4. Support and encouragement for participant engagement through provision of transport to and refreshments prior to treatment sessions 5. Intervention flexibility and responsiveness in offering trauma-focussed alternatives to EMDR which may be more suitable for individual participants needs. 6. An extended trauma therapy programme, including extended stabilisation therapy prior to trauma treatment, to accommodate the complexity of SSW needs.