‘Beyond places of safety’ – a qualitative study exploring the implementation of mental health crisis care innovations across England

Background Mental health acute and crisis care consumes a large share of mental health budgets internationally but is often experienced as unsatisfactory and difficult to access. As a result, there is an increasing move towards developing innovative community crisis services, to improve patient experience and relieve pressure on inpatient and emergency services. This study aims to understand what helps and hinders the implementation of innovative mental health crisis care projects in England. Methods Using a qualitative approach, 18 interviews were conducted with crisis care service managers exploring their experiences and views of the development and implementation of their service developed with support from an English national capital funding programme. A framework analysis was conducted informed by implementation science. Results Key facilitators to implementation of innovative crisis services included bottom-up development, service user involvement, strong collaborative working, and leadership and management buy-in. Key barriers that affected the projects implementation included the complexities of crisis care, workforce challenges and resourcing issues. Conclusion There is a recognised need to improve, update, and innovate current crisis care offers. Results from this study suggest that a range of models can help address the heterogenous needs of local populations and that new approaches can be implemented where they utilise a whole-systems approach, involving service users and relevant professional stakeholders beyond mental health services in planning and developing the service.


Introduction
Mental health crisis care consumes a large share of mental health budgets internationally but is often experienced as unsatisfactory and difficult to access.Crisis care encompasses a range of pathways that provide people facing a mental health crisis with urgent access to care [1,2].Highly salient aims for such services are often to prevent admission to inpatient care, which are costly and often associated with poor patient experience [3] and potential harms including loss of rights and freedoms, stigma institutionalisation and development of unhelpful coping strategies [4][5][6].Diverting people from emergency department attendance, also often experienced as aversive and unhelpful, is another key aim for mental health crisis services.The availability of a range of models of crisis care is also helpful in offering service user choice, in attempting to pre-empt severe crises by earlier intervention, and in ensuring psychosocial and supported selfmanagement strategies are available.
In England, a national survey conducted by the Care Quality Commission in 2021 found that over a quarter of mental health service users did not know how to access help in a crisis, and just over half did not feel they received the help they needed [7,8], despite a longstanding national initiative to establish crisis resolution and home treatment teams (e.g., teams aiming to provide intensive home support to people who would otherwise need to be hospitalised) across the country [6].Furthermore, wide variations in quality and availability of care have been reported despite the nationwide availability of crisis resolution teams [7].With involuntary admissions under the Mental Health Act rising steadily [8], and a rise in referrals to crisis care service observed following the pandemic of 11% per catchment area there is an increasing demand facing existing mental health services and the availability of a range of approaches, especially offering alternatives to hospitalisation and emergency department attendance, is valued.
Considering these challenges, improving access, quality, and choice in mental health crisis care has been the focus of several UK policy initiatives in the past decade [9,10].Both local and national initiatives have resulted in the implementation of a range of innovative service models designed to improve quality and effectiveness of care and flexibility and integration in local crisis care systems.These have included innovations such as safe havens, crisis cafes, specialist crisis assessment services, walk in crisis centres 24-hour crisis lines, and mental health decision units [11].Such developments have been supported in national policy, but there are considerable gaps in evidence regarding the clinical effectiveness and cost-effectiveness of mental health crisis care, the integration of crisis care across inpatient care, post-discharge transitional care, and Community Mental Health Teams/ intensive case management teams, and innovative models have been implemented with little evidence [12].
A component of these national policy developments was the launch in 2018 by the Department of Health and Social Care (DHSC) in England of a call for applications for capital funding for innovative projects to improve local crisis care systems: the "Beyond Places of Safety" (BPOS) initiative [13].Fifty projects were commissioned, which varied in scale, focus, and intended outcomes.Some were delayed due to the COVID-19 pandemic, but by late 2022, all were up and running.The models commissioned included a variety of innovative approaches to improving access, quality, flexibility, and choice.This offers the opportunity to the explore further the changing landscape of crisis care pathways, understanding what helps and hinders innovative crisis projects in implementation and achieving their aims.

Aims
The main aim of this study is to understand what helps and hinders the implementation of innovative mental health crisis care projects in England.

Design
We employed an exploratory approach utilising qualitative methods.We conducted semi-structured one-to-one interviews with leads in new crisis care projects funded under the BPOS initiative project, aiming to explore their experiences and views of the development and implementation of their crisis care service.

Participants
The BPOS scheme funded 50 mental health crisis care projects across England, run by NHS, Local Authority, and voluntary sector providers, ranging from substantial capital-funding of new services to smaller schemes to improve or renovate existing facilities.For our study the inclusion criteria required that the funded crisis care projects had completed their BPOS funding end of year monitoring forms (a yearly form to update the funders as to the activity, achievements, and evaluation, expenditure) to ensure that the work had started and had progressed to a position where project leads were able to reflect on the implementation process.This excluded small scale projects that funded purchasing of equipment, projects that had been delayed, or projects that had withdrawn from receiving the funding.All eligible projects were invited to contribute to the study.All participants were staff members with a management and/or clinical lead role in the project.

Sampling and recruitment
A member of the evaluation team (UF) contacted the project leads for all eligible crisis care schemes to inform them about the study.Where project leads had changed, or where resource required, project teams selected the most appropriate individual to take part in the evaluation.Participants were invited to take part in an interview by contacting the lead contact (UF) who arranged an interview with a member of the research team (NL, NA, RS, RA, MSc) at a time and date that suits the interviewee.

Measure
Topic guides were co-produced and agreed upon within study meetings with members of the Mental Health Policy Research Unit research team, experts by profession and Lived Experience Working Group (LEWG) representatives which includes people with lived experience of mental health problems, using mental health services, and/or caring for people with mental health problems.Informed by implementation science, the topic guide consisted of key questions including (a) what the service was aiming to achieve and whether they think they achieved it, (b) the implementation and sustainability of the project; (c) the barriers that affected the implementation of the project; (d) the facilitators for implementing the project; (e) perceived impact and unintended consequences of the service innovation; and (f ) advice for other crisis care services implementing such services.In addition, we also collected brief information from participants about their work role and nature of the service they were working in, e.g., NHS Trust or third sector organisation.Participants were also asked to send the evaluation team any available reports or summary data which they were happy to share regarding changes in service use and patient flows or service outcomes, which could help the researchers understand the role of the project within its local crisis care system.

Procedure
We conducted semi-structured interviews online via Microsoft Teams between December 2021 and April 2022.Participants were provided with a letter of support from the crisis care scheme lead at the Department of Health and Social Care and were able to read an information sheet and complete an online consent form in advance of the interview.The researcher reconfirmed consent at the start of the session with recorded verbal consent taken from all participants.Interviews lasted on average 60 min.Audio recordings were transcribed by an external company and checked for accuracy by a member of the research team, who also anonymised all identifying information about specific people and organisations.All transcripts were allocated a unique ID number.Data were stored on a secure server with restricted access, with files only made accessible to the lead author and authenticated research staff.

Analysis
We conducted a framework analysis [14] informed by two implementation science frameworks.Using a Micro-Soft Excel file, our coding framework was created to use Proctor et al. 's taxonomy of implementation outcomes [15] to record projects' evaluated or perceived acceptability, appropriateness, feasibility, adoption, penetration, implementation cost, fidelity, and sustainability.For analysis of the barriers and facilitators to implementation, the Consolidated Framework for Implementation Research [16] comprises five major domains including Intervention Characteristics, Implementation Processes, Outer Setting, Inner Setting, and Characteristics of the Individual.
Each domain consists of a number of constructs that reflect the evidence base of types of factors most likely to influence implementation of interventions [17].This initial stage was undertaken by seven members of the study team with each researcher charting data summaries onto the framework for each of the interviews they had conducted (UF, RA, NL, NA, RS, MS, XH).Researchers were provided with coding guidance on the domains including definitions and examples that was discussed and refined as a team prior to the commencement of coding to ensure a consistent process was followed when coding each transcript.The lead author double coded the first transcript alongside each researcher to compare coding and ensure consistency in coding approaches.Subthemes within each broad deductive theme from our initial framework were then derived inductively through further coding and collaborative discussion within the research team, inclusive of Lived Experience Researcher colleagues in the team (PN, CD, KP, CL).
Quotes selected to be included in this manuscript reflected patterns in the data.Selections were made based on discussions within the research team during iterative rounds of analysis.Where necessary, discrepant examples were used where there were distinct differences to be illustrated to ensure the data presented accounts for the overarching barriers and facilitators but also the outliers.This was salient given the range of different BPOS funded project types.

Research team
As part of this evaluation, a co-production group was established at the outset of the study and group meetings occurred approximately every quarter; work on coproducing this manuscript was conducted via video-calls and emails with all manuscript authors.Co-production work ran throughout the course of the study (e.g., via coproduction of the study protocol, interview topic guide and analyses).The lead author (UF) provided training on analysing qualitative research with the lived experience members of the team.One health researcher (BLE) led on setting up the study, including obtaining ethical approvals/amendments and establishing data management systems.The lead author (UF) coordinated the overall study progress, including coordinating co-production group meetings, leading on recruitment processes and procedures and coordinating interview and analysis procedures.

Ethics
This study was reviewed by the Research Director of Noclor, the North London Research Consortium, which provides research governance services for several London NHS Trusts, who confirmed it met Health Research Authority (HRA) criteria [18], to be classed as a service evaluation and formal ethical review was not required.Participants were given written information regarding the purpose of the study, aspects of confidentiality (and their limits), and the extent and limits of confidentiality including publication of anonymised quotes.Staff could decline to take part or withdraw at any stage.

Project characteristics
Of the 50 projects funded through the BPOS scheme, there were 33 completed projects that met the inclusion criteria for the evaluation (60%); the remaining projects had withdrawn or had not yet completed works.Of the projects that met the inclusion criteria and were eligible to participate (n = 33) a total of 18 project managers took part in an interview (55%).Of the 15 projects that did not take part this included where service managers could not be contacted, or where services had experienced high staff turnover thus staff involved in the implementation of the service were no longer available.The majority of included projects were from NHS providers (n = 10) and the remaining eight were from the third sector or nonprofit organisations.

The nature and aims of included projects
The types of projects included within the evaluation involved innovative or novel approaches such as creating or improving local crisis cafes or safe haven provision (n = 7), creating online or digital places of safety such as personalised apps (n = 3), and the development of multiagency hubs or pathways to improve the support offer for those in crisis (n = 2).In addition, there were projects funded to improve existing spaces such as rebuilding or enhancing existing assessment suites or places of safety within mental health services.These included s136suites which are facilities for people who are detained by the police under Sect.136 of the Mental Health Act to provide a place of safety for a mental health act assessment and care (n = 4), and the creation or improvement of A&E places of safety including creation of dedicated assessment suites (n = 2).Of the 33 projects included, only one provided specific crisis care for young people, while the remaining 32 provided care for that was designed for adults of a working age.
Participants reported two main purposes for the projects funded through this scheme outlined by the services included within this evaluation.The first was to improve safety within acute services for individuals experiencing mental health crisis, e.g., improving s136 suites, or providing a specialist assessment suite in A&E intended to be more appropriate and safer for those in a mental health crisis.The second was to reduce use of acute services, by diverting from A&E attendance, acute admissions, or police detentions under the Mental Health Act by offering alternative, immediately accessible crisis care provision or support and safe spaces.An overarching aim of all initiatives was to improve the local crisis offer, e.g., to improve access to care or patient experience for people using services.

Understanding the barriers and facilitators to implementation
Table 1 highlights the key facilitators and barriers associated to these domains that are further discussed and exemplified in the following section.

Characteristics of individuals a. Staff acceptability and belief in the service
It was reflected that positive attitudes and buy-in from wider members of the service staff were important for ensuring the completion and reported success of the service as these elements enhanced referrals and engagement in supporting and working in (or with) the service.

It was the staff group that really sort of drove the fact that they wanted to continue doing it. (Project ID 10) I think it's just the passion, the dedication, the motivation, really, of staff. And it's just going that extra mile really. (Project ID 25)
Training wider service staff across the existing organisation was found to facilitate this buy-in and positive staff attitudes to the new project as it allowed opportunities to build staff confidence and belief in the approach and new service being provided.

Staff [went] through training again on how they can sit with someone if they have the time, or how they can use it themselves…. so people engage with it. (Project ID 49)
b. Personal attributes and skills of project managers Many of the project managers leading implementation had experience of working on the frontline of the services they were describing.This meant they displayed knowledge and insight into the needs of the service as well as having a motivation and competence to be able to deliver meaningful change to the system.Only one participant noted challenges to staff 's acceptance of the service innovation due to the complexity of the digital approach needed to deliver it, which required external expertise.

Facilitators
a. Bottom-up development with engagement from existing staff, managers and service users Most projects were developed internally by members of the lead organisation, delivering them with existing staff and tailoring the national model to be adapted to the local context.This local development was felt to be a facilitator to the service succeeding as it ensured that the service was adapted to fit the context in which it was being delivered.

It's quite a geographical spread, quite a rural area, so what we've always had to do is create two places of safety (Project ID 32).
For services building new digital apps, it was noted that there was a need for engaging with external companies and experts in this technology due to the specialist nature of the project, however they all reported this as a collaborative approach led by internal plans and aims to ensure consistency.
Ten of the services noted service users' involvement in the development of the project in some way.This inclusion of service users was seen to facilitate implementation as it further ensured that the service was meaningful to those who would be receiving care and could be shaped by awareness of the issues that previous services may have had.

[using a] certain level of co-production is basically, from my point of view, communities designing solutions to meet their own needs, I think, rather than us, as organisations, creating services to meet needs for people… we've had conversations about, "How can we do this better? How can we make this work?" but we are very much at the start of that journey. (Project ID 11)
b. Using a flexible, person-centred, non-clinical environment Creating a warm and welcoming service.
Creating services that were warm and welcoming both physically and psychologically was felt by a number of managers as core to their crisis care service;

It helps people to feel valued, and I've seen other places where there is, like, plastic chairs… Do you know what? If no one has taken the time to give you somewhere warm and comfortable, how do you feel welcome? How do you feel listened to? How do you feel valued? We're just telling people that, "No matter what else is going on in your life, we're listening to you. We value you and we want to help, if we can. " (Project ID 35).
This was particularly important for projects in institutional settings such as A&E assessment rooms and s136 suites where people may be detained and not want to be there.Thus, having a milieu that respects the person's needs beyond simply reducing risk was seen as key.The key aim of several assessment or s136 suite redevelopment projects was to move the crisis space outside of medical settings so that it felt more comfortable and less overwhelming.A key advantage described reported by service managers was taking away the clinical feeling associated with crisis care spaces in hospital settings such as A&E or inpatient services.This made communitybased services more accessible and welcoming, making it more likely that people would turn to them for support before in their time of need, potentially seeking help before a crisis became so severe that admission seemed unavoidable: c.Adaptability: creating a service that responds to the reality of crises An adaptable service system that can flex and bend to the needs of individuals experiencing a crisis was seen as salient by all the crisis care services; be that as having flexibility within the service itself or by creating flexibility within the pathways across and between services.While digital places of safety were seen to provide an alternative to traditional crisis care for those in rural locations, crisis cafes and safe haven models which accept self-referral and offer out-of-hours services were considered an alternative to A&E.

"[the new service space] it's allowed us to have so much more flexibility in responsiveness to people in crisis…it has just given us options for patients to come and see us… we've got a reception area; we've got options for people to come in and wait. Whereas before they might have had to wait at home, which okay that's fine if it's safe, whereas at least waiting here it's safer… we can keep an eye on situations while we're doing other bits of work …that wouldn't have been an option at all if it was an ambulance had turned up at a house. They would invariably want to take you to A&E and not bring you anywhere else. So, yes, it's given us a greater degree of flexibility. " (Project ID 8). "We've delivered a blended model, of face-to-face and remote. Whereas a lot of mental healthcare provision has obviously retreated into a much more purely remote model. But in terms of crisis care, that isn't really an option. " (Project ID 25).
Having third sector partners was perceived as a facilitator to delivering a flexible service as these colleagues seemed less influenced by restrictions and bureaucracy that could act as a barrier for people accessing services.Participants highlighted that a lack of adaptability and flexibility increases obstacles facing service users seeking help.

[Our lived experience colleagues who facilitate the café] tolerate individuals that […] I know clinical colleagues wouldn't in a formal setting, "That's not our criteria. Oh, no, we can't have them' … Don't have referral criteria, as in, don't make it harder for people to access lower-level services than it is. (Project ID 35)
Barriers a. Complexities of Digital Crisis Care Projects utilising digital approaches noted that there was a need to develop elements of the service externally due to the specialist nature.This created two key challenges; firstly, the external developers had limited knowledge regarding mental health and crisis care, therefore there was additional demand for the internal team to ensure a safe and sensitive tool was developed.Secondly, the need to develop externally created a conflict between what service users and staff wanted and requested, and what was logistically deliverable.
They might have what they think is a good idea, and it's quite easy to verbalise, but there are more complex needs underneath that just in terms of the way they've engaged with the interface… they might be a deceptively simple thing to do but actually it is hugely complex.(Project ID 27) As a result, these projects were resource-heavy in both time and, cost which was felt to have a considerable impact on other aspects of the implementation process.One project noted that, as a charity, this burden was felt across the team and obliged the organisation to apply for further funding to resource their internal team to deliver the service.

The process of implementation
a. Facilitators Planning and piloting.
Projects that included a pilot phase during implementation noted a number of benefits.Piloting was seen as a way to address implementation issues and plan ahead for improvements throughout the wider process.
We did a small sort of three-month pilot, to see if it worked.And then obviously we'd done the work to make the building better.And it's still running now, so that's brilliant.(Project ID 09) For one project using a new digital approach, piloting had an added benefit as it provided a means of build team skills and confidence in implementing the project, as well as increasing staff buy-in: I think the most important thing is to start with your own team and your own staff.Get them trained.Get them piloting.We got people to act as clients and then reversed it.We had a whole morning of that, chaos ensued, but we did that because we wanted to try all the different aspects of it in terms of what could go wrong.(Project ID 41) b.Service user involvement Ten of the projects discussed the important role of service user involvement at all stages of the service development and implementation.By having input from individuals who had experience of a mental health crisis, it was clearer to the service what could be done to create a safe and welcoming space for those in need during such periods of acute distress.This ensured that the service was a viable and useful alternative to pre-existing services.
Because they are a charity and we're not employing the people with lived experience, we haven't fallen into that trap that other places have about lack of objectivity and those issues.It's almost like they are a critical friend, but yet people value their input and value the co-production, and everything we did as an organisation was around the co-production really.(Project ID 25) Community-based services such as crisis cafes were seen as being more able to involve these facilitators, while more traditional facilities such as assessment suites or s136 suites were less able as they were more restricted by elements of risk management.
c. Learning from existing services Two services implementing crisis café style services found that visiting other services was of considerable benefit to their development of their service as this supported the developing team where guidance and evidence was limited.

We had [other services] come down and do some workshops with us about different models, what was good practice, what was out there. We went and visited a number of services… So based on that, what
we came up with in the pathway was that we developed this service (Project ID 25).

d. Reflecting, evaluating and executing change
Eight of the interviewees reported they were involved in a formal evaluation, and all services that received funding were required to complete monitoring forms as a form of evaluation.Other forms of feedback were used across the projects including the use of compliments and complaints, service user engagement, and informal feedback processes.
We have taken on board compliments and complaints, and we have worked alongside our liaison psychiatry colleagues, and dementia care colleagues, to make sure that we are as minded as possible of the recommendations around planning spaces for people living with mental health problems, or dementia.(Project ID 48) Many of the services noted the importance of ensuring that part of their process included capturing this feedback as part of a cyclical and iterative approach (often called 'continuous improvement') to obtaining insights through data and patient feedback and using this learning to guide and improve the service.

Barriers
a. Resources, costs and sustainability of capital funded approaches The majority of projects reported that they delivered within budget.The funding was only provided for the purpose of capital costs, thus required additional revenue funding for aspects of the service such as staffing.For most services this revenue funding was sustained and sustainable within the wider services, however two of the projects noted challenges in securing these additional resources that were required to deliver a fully functioning service.

I think some of the difficulty was that the budget was capital funds. That limits how you're able to use that money… if you set up a service then you always get that recurring cost. It would be really beneficial to have that identified ongoing resource to help with that earlier on. (Project ID 1)
Challenges related to high staff turnover, changes in staffing following COVID-19 and changes in the scope of the project designed and that which was delivered accounted for these resourcing issues and highlighted the dynamic nature of crisis services.

Facilitators
a. Leadership and management buy-in A key facilitator for these projects was having committed and influential leadership supporting innovation and change, thus building buy-in from senior managers and commissioners.Having senior teams who were motivated to make the service work built confidence in the project among staff and made change feel achievable.
We had good support from our commissioners, from the managers, in terms of agreeing the bid… We had support within the trust and with the mental health commissioners.(Project ID 48) I had a new service director that came in, and we sat down, and we said, "Right, how are we going to deal with this?" So, we said, "Okay, let's look at who we've got.This person has got a strength in this area… let's flatten the hierarchy, and let's unlock people's creativity and stuff." (Project ID 25).
Such engagement was beneficial particularly where commissioners were engaged at an early stage in planning.

You know, she was involved from us drawing up the plans to make the changes to the building, she was involved at that stage, really. So that worked really, really well. (Project ID 09)
This was felt not only to strengthen development and implementation processes, but also to support the sustainability of the services.There was belief and motivation to continue the service, thus further funding was secured to continue the service.

Barriers
a. Complexities of staffing services In the creation of new person-centred approaches for individuals experiencing mental health crisis, services set in non-mental health specialist areas, e.g., general hospital A&E settings, found that the complexity of providing such a service was a barrier to delivery.This barrier was related to the limited training that medical staff had in providing this style of crisis care.
The biggest thing I would say is staffing and trying to make sure that you have got appropriately trained staff to work within the environment.It is staff who are highly trained [to deliver mental health care], not just registered practitioners (Project ID 48).
Within NHS settings in particular, staff turnover was a concern for project leads as changes in staff who championed new services could have an impact on the overall drive and motivation for the project to succeed.Exemplifying this, one service was unable to take part in the study due to high staff turnover in the service that meant no one who had been involved in the funded project was still working in the organisation.
Changes in staff were felt to impact staff motivation and drive, as well as having influence on the implementation by creating setbacks at a systemic level.

Every now and again, there is a change in personnel. If it's a change in personnel in leadership, then you've got to rebuild all those relationships again. (Project ID 28)
This issue of staff turnover had an impact not only at the staffing level of service but also at an organisational level in which the need for consistent leadership was central, thus any changes in this area had deep impacts that were felt through service delivery.

Facilitators
a. Adaptability to a local context A number of project leads noted the importance of keeping in mind the wider context in which their service sits was necessary to ensure that they could meet the needs of the local population.This consideration was a key component in the decisions made by the three organisations implementing digital approaches, as previous projects had found that the wider rural setting made travelling to a physical location a barrier for some service users.

So again, the barriers for somebody in a small village 30 miles away to talk to somebody face-to-face can completely be overcome by this sort of thing [digital approaches], "You can see us face-to-face. We would love to see you and talk to you face-to-face and this is how we can do it, and it is as easy as this… we can do a test call, we can show you, walk you through it. " (Project ID 41).
By understanding the wider context and needs of the local population, they were able to develop a service that would have an impact for those in need and when current alternatives such as driving to the local A&E or taking public transport to a crisis café would be challenging.One project manager noticed the lack of uptake in the service where evidence-based approaches were used but assumptions about the local context were made, and the local context was not accounted for.

My main advice would be really think about the times of opening… I could be open all night and I wouldn't see a soul, ever. Because it's just very different. So, think about your geographic demographics in terms of things like that, don't always assume that people want to access that kind of support through nights and at weekends because actually when you look at when people attend here, it isn't after 9 at night. The reality is, unless you're in a city centre, which we are not, people don't want to come out, because public transport gets worse and it's dark. And they just don't like that level of travel at night in more rural areas. (Project ID 1)
As a result, this service adapted its model to address the local needs and developed satellite sites to accommodate the needs of service users in more rural areas.
"…We are really rural… so, we're going to do pop-ups in the local communities, to try and take the support to them, rather than them having to come to us." (Project ID 1).

b. Engaging key stakeholders and partners in development
Services noted that having external collaborations such as police and ambulance services with key stakeholders was important for troubleshooting how to create a collaborative service integrating all key stakeholders who all have their own requirements and skill sets.This was more common in NHS services assessment and s136 suite developments where such multidisciplinary teams are required.
We've built a very clear observation area that has a nice observation window and desk spaces for three people so that the AMHPS, doctors and nurses can get in there and write their reports and have their discussions… and the police came and worked with us on it.It was designed to have a dual purpose; so that it was somewhere that was safe for us to manage someone's aggressive behaviour, but it was important it was also somewhere that was welcoming.(Project ID 32) Not only was this networking and strategic partnering to create buy-in seen as important, but it was also felt to be a key facilitator from a care perspective.Where services were networked well within the wider crisis system there was opportunity to signpost and refer across systems thus creating opportunities to allow service user choice as well as genuine alternatives for people in crisis.

Barriers a. Policy changes
Overall, all service managers who reported detail on the adoption and penetration of their crisis care project expressed positive views regarding how the new service or changes to an existing service had integrated to their existing services and crisis offer.While the majority noted that the set-up and delivery of their project had good fidelity to their original plan and that the changes were implemented as originally prescribed, one service found that, due to policy and guideline changes in relation to Psychiatric Decision Units, the original project was no longer feasible.Therefore, the outcome was delivered differently than intended, with an urgent care hub created instead of a psychiatric decision unit: Information came from the CQC, RCPsych, and from the National NHS England team around a shift away from recommending psychiatric decision units, the type that we had put into the original bid.So, we went about reviewing a slightly more therapeutic space that was attached to our Health Based Place of Safety to respond to that national directive.(ID 09, NHS Service)

b. Delays due to Covid-19
As the funding was provided in 2019/20 there were considerable barriers facing all projects as a result of the COVID-19 pandemic.Some services noted delays in progress due to internal changes in staffing and ward spaces, while others noted external delays from companies doing construction work.
There was a construction delay [due to covid-19] … some people have moved on during that time.(Project ID 4) In addition, service managers noted that demand for crisis care had increased for some of them during the COVID-19 period, reflecting the impact of the pandemic on service users' needs and service delivery.For one A&E assessment suite, the increased demand for inpatient admissions and staffing issues in the hospital meant that service users were waiting longer in the assessment suite.
Another service providing an out-of-hours crisis café noted that during the pandemic the café was unusually busy.This led to the facilitators being overwhelmed and required problem solving regarding how to turn people away when demand was high, particularly whilst trying to sustain social distancing.While acknowledged as a challenge, they felt that this increasing demand was also a positive reflection of the service in that it was a genuine alternative to people feeling increasingly worse or having to attend A&E.
We only had enough money to start with to run it for three days a week, and demand is growing… Do we stop people coming?We can't have more than six people visiting at a time so we've had to say, "If you are a regular, and actually, you're here for a game of cards, and someone new presents, you will be asked to move on, and you understand that?"It's not easy, but we are trying to manage it best we can.(Project ID 35)

Perceived impact
Overall, the project leads reported positive outcomes from the implementation of their project.Participants reported a range of perceived positive impacts not only on their own service but also across the crisis care systems.Table 2 below outlines the perceived outcomes according to Proctor et al. 's taxonomy of implementation outcomes [15].Outcomes presented in Table 2 reflect perceptions of service leads rather than validated evaluation data.

Main findings
In this study, we aimed to explore what helps and hinders the implementation of innovative mental health crisis care projects in England.When taking stock of the current picture of crisis care models and evidence, Johnson and colleagues [19] highlight the need for a greater understanding of how best to implement change and innovation in crisis care across a wide range of contexts.This study provides insight into the facilitators and barriers of such expansion and new models of crisis care.The use of the CFIR constructs provides a nuanced insight into the integral role of developing services with leadership, collaborative working, and staff resources at the core of the service.
Furthermore, this study notes that the need for localised, flexible, and person-centred approaches to work as effective crisis care models reflect that overall accessibility, support, and the extent to which an integrated and flexible crisis response from helpful and empathic staff is highly valued by service users [20].The results from this study indicate the importance of lived experience input within the development of projects as well as in their delivery with services noting that co-production and codelivery of innovative services such as crisis cafes was central to their acceptability and sustainability among service users.Utilising such bottom-up approaches in this field is important as this ensures that the service users' needs remain integral throughout implementation and beyond into the ongoing delivery of care [21].
While the findings in this paper suggest the strength of utilising a bottom-up approach to developing crisis care services to allow for personalisation and flexibility to account for service context, this creates a tension with the need to utilise evidence-based models.Previous literature has noted that having high fidelity to a model can be associated with better outcomes [22].As reflected within this paper, the complexity of crisis means that evaluation is hard to do in such contexts, however such tensions between the local evolution of services and need for evidence-based models with well-developed methods for improving and monitoring fidelity may be helped by utilising Quality Improvement (QI) methods [23].Namely, the use of learning from existing services, piloting, and reflecting to ensure iterative change to improve services were noted as important elements that facilitated the process of implementing changes to existing services or for the creation of new services.These findings outline that it is helpful where possible to inform service development and justify continued funding through utilising this QI approaches within service improvement and innovation in mental health care.
While most of the implementation domains highlighted key facilitators to support such service change or implementation, there were clear barriers that impacted on the projects, especially the complexity of delivering crisis care in context of resource constraints which exacerbate the workforce challenges facing our healthcare systems.Most of the perceived barriers identified in this study were within the 'Outer Setting' domain due to these external economic, political, and social agents, which are likely to be common in any organisation as key barriers within implementation of services.Considering these challenges, the adaptation of responsive and flexible -Allowing care to be more integrated across different services "This allowed us-and continues to allow us, to provide that support in a way that gave more equality of access, and again, safety of discussions, safety of conversations without people needing to physically come down to the building." (Project ID 41) Fidelity -Overall, the majority of services reported that the set up and delivery of the project had good fidelity with their original plan -One service was an exception, reporting that due to policy and guideline changes in relation to Psychiatric Decision Units meant the original project was no longer feasible "All aims achieved and fidelity to the funded proposal apart from receiving accreditation." (Project ID 09) "Information came from the CQC, RCPsych, and from the National NHS England team around a shift away from recommending psychiatric decision units, the type that we had put into the original bid." (Project ID 09) Implementation Cost -The majority of services reported the implementation being within budget -For the two sites which came in over budget, the extra cost was met by the NHS Trust -A further two sites reported being within budget but had some limitations e.g.recurring costs not being covered "…we couldn't, as a charity, have invested £40,000 in that building, to have done the work to enable us to do this.So, to be fair, yes, we probably wouldn't be where we are now, without that funding…" (Project ID 42) Penetration -Several sites reported high penetration, with new initiatives integrating well into the existing care pathway -In some cases, this led to decreased pressures and demand for other services -Some services mentioned wanting to increase the roll out further, for example bringing it to remote areas via 'pop-up' centres "Good usage, integrates across the pathways to relieve pressures across the directorate" (Project ID 25)

Sustainability
-All sites reported that the service would be sustained in the future -In some cases, adjustments to the existing model were being planned for the future service -All but one site reported that further funding had been secured to allow for the continuation of the service "Increased demand allowed investment to fund further nights and sustain for 3 years as they saw the value" (Project ID 35) bottom-up approaches informed by service user needs and the acknowledging the nature of crisis was necessary for these projects to achieve their reported successes, an approach recognised as important in the delivery of complex healthcare services [24,25].
Within the present study, services experiencing such resourcing issues and the lack of continuity of staff created barriers to implementation as this created a challenge in keeping momentum for innovation and changes.This likely reflects the wider context, as mental health services are experiencing workforce shortage with staff turnover being a growing issue [26].Alongside the challenges noted in the study in relation to barriers created by the COVID-19 pandemic, this adds further emphasis on a need to consider implementation through a wholesystems lens those accounts for national level factors [27,28].This is further emphasised when noting the importance of leadership as a core facilitator to completed implementation of such projects as it is necessary to consider the challenge of relying heavily on such factors for the perceived success of a services.Previous literature has noted that the failure to sustain alternative service models can be because of a reliance on charismatic leaders and local champions, without whom they may not thrive [29].

Strengths and limitations
While the results of this study outline the completed implementation of a range of crisis care models, there are limitations to these findings.Firstly, these results are based on interviews with service managers who provide a single stakeholder perspective, and have been invested in the service, thus are more likely to provide a positive perspective regarding its implementation.Triangulation with quantitative outcome data would be necessary to provide a fuller picture of the impact that such implementation of service has for the wider crisis care pathway and service users receiving this care.Furthermore, as part of a wider evaluation of the services there may be response bias as participants are keen to provide views that outline their perceived success of the project.The study had a low response rate due to a number of projects not partaking for a range of reasons.These services may have had important insights into wider challenges and barriers that the included projects did not experience, thus there is a need to encapsulate these voices in further work into crisis care implementation.Lastly, it is important to highlight that these service implementations occurred within the specific context of the UK and mostly within NHS services which limits the generalisability of such findings to other countries in which healthcare is delivered differently.
A strength is the collaborative, iterative approach to our analysis.Initial deductive coding was conducted by seven of the study researchers.The second stage of inductive coding was conducted by study researchers working together with four Lived Experience Researcher colleagues.The developing coding framework was refined further with input from the wider study team, including colleagues with a variety of mental health clinical experience.This collaboration came with challenges to ensure consistent coding, but the resulting analysis was informed by a range of views and stakeholder perspectives.

Implications for policy, practice and research
The findings from this research have some potential to guide local implementation of innovative crisis services as they point to some potentially actionable strategies which service planners and developers can adopt.These include adapting service provision to the local context, involving service users, engaging with a range of key stakeholders and leaders, as well as learning from others in the wider field of crisis care or from others adopting similar approaches.
Furthermore, these findings can provide insight for commissioners into the potential space for capital funding projects like this crisis care scheme as a bottomup approach to developing initiatives that can help to address local need.These findings outline the potential to produce benefits to local crisis care systems in a meaningful way with relatively modest funding; from the total investment of £15 million for these BPOS projects, the majority of services involved within this evaluation kickstarted and have sustained with investment of less that £500,000 [13].Considering the potential impacts that such services may have on service users, local communities and services, such modest investments could support the development of considerable innovation and create alternative pathways, which is highly salient against a backdrop of workforce challenges and economic cuts.By learning from the perceived successes of these services, commissioners and service manager can develop insights into approaches that have, or have not worked, and adapt this to their local context and needs.As a complex field of healthcare, it is vital that service managers across the field can learn from others in the successes and challenges faced, as well as noting the need for flexibility and adaptability in developing and implementing crisis teams and services.
Given the importance of co-production highlighted within our results, future research should consider widening the participation in such evaluations of crisis care to account for service user views and experiences of using these types of services, as well as ensuring that service users are actively involved in the development of such services.Further work is needed to actively utilise these service user voices, as well as quantitative, longitudinal, and health economic approaches to build the evidence base through measuring real-life outcomes and the impact that these improvements and innovations have on the wider crisis care systems to help inform future initiatives and investment.
In addition, this study noted the impact of staff turnover on the implementation of these services.Given the growing workforce and resourcing challenges facing mental health services following the COVID-19 pandemic, it is likely that longer term services will experience further staff turnover which will impact motivation and leadership of these innovative models, as well as growing barriers from the 'outer setting' .As a result, future research should consider longitudinal approaches to examine the longer-term impacts of implementation strategies.

Conclusion
Against the backdrop of increasing workforce challenges and increasing demand on mental health inpatient beds, as well as the poor experiences reported by service users, there is a recognised need to improve, update, and innovate current crisis care offers.Results from this study suggest that a range of models can help address the heterogenous needs of the local population.Such services can be implemented where they utilise a whole-systems approach, involving service users and relevant professional stakeholders beyond mental health services in planning and developing the service.Flexible, integrated, and collaborative approaches that are informed from the bottom-up are desirable, supported by consistent, active support from senior managers and commissioners.

Lived experience commentary BPOS crisis care (written by Beverly Chipp)
What happens to a person when they are in crisis is the most significant part of their mental health journey, particularly if we aim to avert hospitalisation.Establishing an evidence base for models is essential to support bold investment and yet there cannot be a single best model as people are different, and circumstances and drivers of crises can vary.Nevertheless, we must know what works, and offer choices.This paper helps to unpack what facilitates and hinders, so is a valuable basis for iterative improvements of each model.
Those with the powers may be thinking hard about the technicalities of implementing innovative services but putting thinking aside for a moment and focusing on feeling; Imagine yourself in crisis.In that jangled state, what do you need?Intuitively you may long for a homely space, quiet environment, and people you know and feel safe with.These natural, human instincts we can all understand, yet too frequently new facilities still look bleakly clinical and staff turnover prevents trusted relationships developing.Pressured frontline workers would also benefit from better interior design and should have a say, potentially improving retention.Unsurprisingly the paper finds that staff reported involving service users in design and implementation of services contributes to their reported success, and yet the "bottom-up" approach recommended by Baker in 2001 still seems rare.Peer workers can help to deliver essential components of any good service, namely kindness, patience, and common human understanding.
It is reported that there are knowledge gaps regarding the clinical efficacy of many crisis models, but strong indications could be gleaned from clients' qualitative feedback, because our subjective lives are ultimately the embodiment of clinical effectiveness.
The bottom line is knowing where to go when crisis occurs.If effective crisis services equate to avoiding hospital admission yet over 25% of people don't know where they could go [4] then investment in good crisis alternatives and their promotion could reduce inpatient bed pressures by a quarter.

Table 1
Overview of Themes by CFIR Domain

Table 2
Perceived Impact categorised by Proctor et al. 's (2011) taxonomy of Implementation Outcomes Able to offer out of hours support: one service reported being able to provide online out of hours support to help reach people who could not access other support, or who lived far away.