Liquidity and uncertainty: digital adaptation of a complex intervention for people with severe mental illness during the COVID-19 lockdown

Abstract

Background

This paper explores the extent to which the implementation and evaluation of a collaborative care model of face-to-face service delivery for people with severe mental illness was viable during the first UK lockdown associated with COVID-19. The PARTNERS2 cluster randomised controlled trial and process evaluation were co-designed with service users and carers. The aim of this paper is to explore whether digital adaptation of the PARTNERS model for people with severe mental illness during the COVID-19 lockdown was equitable, in terms of fostering collaboration and trust in a vulnerable population.

Results

We collected qualitative data from multiple sources during lockdown and subsequently constructed case-studies of participating secondary care workers. We adopted Bauman’s notions of liquid modernity to inform our analysis, and identified that digital adaptation during lockdown was only successful where organisational policies, care partner skills and service users’ existing resources were optimal.

Conclusion

PARTNERS2 can be delivered digitally by a care partner to support people with severe mental illness to identify and work towards their goals when existing resources are optimal. However, at a time of increased need, we identified that people who are very unwell and living with limited access to resources and opportunities, remained disenfranchised at great cost.

Trial registration

ISRCTN 95702682, registered 26.10.2017

Background

People with severe mental illness (SMI) have to cope with both the impact of a chronic health condition and discriminatory societal attitudes [1]. Around two in three people with SMI live alone and in poverty, rating their quality of life as poor [2]; and most want to work, although less than one in ten do [3]. The side effects of medicines on both physical and mental health functioning can be challenging [4] and contribute to a mortality rate 20 years lower than the general population [5]. Recovery focused care has been increasingly promoted in the UK, and new models are emerging including use of peer support workers [6] and recovery colleges [7]. The community mental health framework places emphasis on integrated care particularly between primary care, secondary mental health services, social care, and the voluntary sector [8]. Models of care that straddle primary and secondary mental health provision are emerging [9]. As is the evidence base for remote mental healthcare, particularly since COVID-19 forced the acceleration of telephone and online healthcare delivery [10].

The PARTNERS2 study was co-designed with service users and carers to test the effectiveness of a collaborative care model of face-to face service delivery for people with SMI (Author ref). Briefly, the model involves a secondary care worker (a ‘care partner’) located in primary care in the UK, building rapport with a service user to develop a shared understanding and using coaching to identify goals, and linking with other practitioners and services as needed [11]. A team of service users and providers, led by a psychiatrist, provided initial training in model delivery, subsequently augmented by top-up training and reinforced with meta-supervision (e.g. supervision of both care partners and their supervisors in primary care, by a more experienced and senior practitioner) [11]. The model is underpinned by a realist programme (or mid-range) theory, about how the mechanisms of the intervention work in a given context to bring about certain outcomes [11, 12]. For example, delivery was intended to be flexible with supplementary texting and telephone encouraged when preferred by service users, and initial face-to-face contact was central. Video contact was not part of the initial formulation of the care model. The trial used a cluster randomised controlled superiority design, with randomisation of 39 GP practices across four regions in England. The primary outcome was quality of life, measured with the participant reported Manchester Short Assessment (MANSA) at baseline and follow-up [11, 13]. The parallel process evaluation involved a quantitative analysis of fidelity, using the PARTNERS2 Collaborative Care Fidelity Instrument, as well as a realist-informed qualitative process evaluation are reported elsewhere [13].

The COVID-19 pandemic began 18 months into the 29-month delivery of the intervention, as trial recruitment was ending. This provided a unique opportunity to explore the viability and impact of providing an intervention designed to foster collaboration and trust remotely with a population perceived as hard to serve. We continued both the trial and process evaluation during the COVID-19 first lockdown period (26.03.2020 – 19.05.2020, hereafter ‘lockdown’), adapting each critical component of the model (e.g. collaborative care, rapport building, coaching and supporting physical health) for digital delivery. Responding to the rapid notification of lockdown, care partners were offered additional training in the use of video based digital technologies by the meta-supervisor, in line with government guidance for using digital technologies [14, 15]. The process evaluation also required adaptation, and we tested digital data collection during this period. The trial steering committee approved continuation of the study based on evidence that data could be collected, and digital delivery possible.

Aware that the rapid digitalisation of an individualised intervention for SMI could foster inequity, we employed Bauman’s concepts of liquid modernity and liquid uncertainty to help us understand the study participants’ experiences of living through the sudden changes to our modes and practices of communicating and engaging. Bauman contended that we live in an age of liquefaction - where constant change is the only certainty, and where people are required to be increasingly flexible, expansive and resilient as our points of reference are constantly changing and being renewed [16]. For Bauman, the uncertainty and disengagement that we experience from living in liquid times can be magnified when we become ill, lose our jobs, or experience forms of social disconnect, which he argued can be magnified by digital technologies (in the place of community and more tangible forms of support) - which add to the fragility and isolation of vulnerable individuals [16].

This paper aims to explore the extent to which digital adaptation of a complex intervention (PARTNERS2) for people with severe mental illness during the COVID-19 lockdown may have further marginalised people who are subject to multiple forms of disadvantage.

Methods

Sampling

We prioritised collecting data from care partners and participating service users, using our existing realist framework (where we had previously identified ‘messy’ areas of programme theory for further exploration in our process evaluation (Schön 1983) [17], e.g. the role of participant characteristics), rather than frontline health practitioners (e.g. general practitioners) whose workload increased due to the pandemic [9]. With further research ethics committee approval (West Midlands – Edgbaston Research Committee ref: 14/WM/0052), we amended existing study documentation for digital data collection, and tailored topic guides to capture perceptions of adaptation and impact on delivery of the model.

We sampled all six care partners, working in three different healthcare systems, who delivered PARTNERS2 during lockdown, with the aim of capturing any intervention sessions that they were able to conduct with service users (detailed below). This cohort sample were from a range of disciplinary backgrounds, working within different service delivery models [18]. They include existing care partners who had established relationships with service users via face-to-face meetings, and those new in post and who had not met service users prior to lockdown.

Data collection

All participants provided informed consent prior to each interview or conversations where research notes were made. Interviews were undertaken by CH-M, JG, DR, RD and RG-J, using realist topic guides, designed to capture aspects of the programme theory, e.g. context, use of aspects of the model, and perceived outcomes [9]. HP, CH-M facilitated the collection of records completed by care partners (e.g. details of the content and duration of intervention sessions with a service user, and details of supervision sessions). Interviews with care partners were undertaken using MS Teams or Zoom. Interviews with service users utilised their preferred media, in keeping with the methods used for intervention session delivery by their care partner, where this had occurred (MS Teams, Zoom or telephone). Audio-visual data was collected via digital software, and an encrypted audio-recorder was used to capture telephone conversations. Encrypted data were transferred securely to a General Data Protection Regulation compliant transcriber. All data were managed with NVivo for Teams [19], with all participants given pseudonyms [20].

As organisational and individual barriers prohibited care partners from conducting and recording any intervention sessions with service users during the first six weeks of lockdown (detailed below), data collected post-lockdown are included to further illuminate the experience and impact of digital service delivery [21]. These are: intervention sessions between a care partner and service user, tape assisted recall interviews [22] with both care partners and service users (where excerpts of recorded sessions between care partners and service users are replayed and participants’ perceived effectiveness of components commented upon), transcripts of formal interviews with care partners (with tape assisted recall participants and non-participants), service users and personnel involved in facilitating research (research managers), and fieldnotes from less formal conversations with care partners who were reassigned or resigned, and with service users who withdrew from the intervention. As the data collection period progressed, two care partners were reassigned to other frontline duties, while one resigned from their post.

Data analysis

Following the data analysis plan for the process evaluation, case studies were created for each care partner using longitudinal multi-modal data [23]. Data were initially analysed inductively [JF] using evaluative coding [24] to explore the viability of digitalising an intervention designed to foster collaboration and trust between a care partner and service user in an often hard to reach population. To assess the qualitative fidelity of the intervention to the collaborative care model, these evaluative codes were then tested deductively (with [CH-M and RGJ], and wider research team [LG, MC, VP, HP, SR, JG, DR, RD and RB], including members of the lived experience advisory group) to explore more fully the impact of remote delivery on the assertions and propositions that underpin the PARTNERS2 programme theory [12]. To further explore whether digitalisation was equitable, in terms of fostering collaboration and trust in a clinically vulnerable population, we then employed a more long-range or macro theory of modernity and uncertainty, to inform our sense making. Bauman saw the role of sociology as to disclose the possibilities of living differently and with meaningful choice [16]. For Bauman, the ‘liquid modern human’ is one who is subject to the perpetual forces of transition, whereby the old ways of working seem to no longer fit practice, while the new ways of living have yet to be entrenched [16]. Bauman cautioned that the widespread adoption of digital technologies has the potential to undermine the ‘veracity of realism of a mutual person-to-person trust’ such that its content and significance can be profoundly reduced [25:15]. We therefore applied Bauman’s diagnostic approach, retrospective to data collection, to further explore the extent to which the experience of using digital technology could be consider as indicative of, or a possible mediator of, the liquid times that the COVID-19 lockdown fostered (e.g. characterised by social isolation, predictive uncertainties, and reactive policy and practices) [25].

Results

We collected over ninety ‘units’ of data about digital delivery collected during lockdown and subsequently from intervention providers and service users, which varied from a sentence in a reflective practice log, where care partners were encouraged to reflect upon their experience of delivering the PARTNERS intervention, to a whole interview specifically about digital delivery (Table 1).

Table 1 Care partner intervention delivery data, collected during lockdown and subsequently

Our Bauman-informed analysis, employing the concepts of liquid modernity, liquid uncertainty and the notion of an enduring underclass enabled us to reflect upon aspects of our existing programme theory pertaining to the skills and resources required to provide and benefit from the digital provision of PARTNERS2 to consider which are amenable to revision for effective digital delivery. This additional lens enabled us to develop additional context, mechanisms and outcome configurations pertaining to digital delivery.

Liquid modernity

Bauman identified liquid modernity as the current epoch, characterised by constant mobility and change in identities and relationships, and in which individuals risk becoming disembedded and excluded [16, 26]. In our analysis, we identified examples where digital service delivery amplified the extent to which some care partners and service users could appear ‘embedded’ in or ‘dis-embedded’ from effective service provision. For example, Grace was able to draw upon and adapt her prior experience of the PARTNERS model, whereas Sarah struggled to provide effective digital delivery, due to having limited experience of PARTNERS2 and working within a service delivery model where meaningful integration of the PARTNERS service with the wider mental health system had not occurred.

Grace had three years’ experience of delivering the model since the formative evaluation (Author reference). Prior to lockdown, the meta-supervisor reflected that Grace, having recently been working in a crisis service, required support to take more time to understand and not rush into plans with service users (practiced through role plays within the training) to operationalise the PARTNERS2 model. For Grace, this training allowed her to align her core values with the programme theory:

“I think with that first training … it was the hardest thing to do, not to give any advice!... Because we’re so used to being, you know, the rescuer... And we’re not putting the responsibility for stuff back on to people, so they can draw on their own strength, you know?...I started to recognise some of that. [Grace_INT2]

During lockdown, Grace suggested that her experience of digital model delivery was enabled by a secure knowledge of the model and pre-existing relations with some of the service users on her caseload. This allowed Grace to focus the intervention on service users’ immediate goals:

“Four [service users] I had previously met, and I can pick up and continue the relationship that we had face to face.”[Grace_COVID1]

“You take things at people’s pace, don’t you, you know? And a lot of people wanted to sort of talk about COVID-19, and who they missed in their lives, you know, that they can’t have contact with and that type of thing. The ones that I had previous to the end of March when lockdown happened, those I am much further along in the process with.” [Grace_COVID2]

However, she also noted how learning from the care partner training had to be adapted to enable ‘remote’ rapport building and the establishment of trust with new service users, especially those she had not met before lockdown:

“I think it’s probably taken longer. You’re not able to pick up those cues that are non-verbal… [New service users] are beginning to recognise my voice, store my number, which is always a good sign, you know, that they’re engaged in the process… So, over the last few weeks I have worked on building that relationship over the phone… so that gives you hope, if you like, that you can carry on and try to live up to what you’ve said. If I hadn’t spent that time, I would have had to deliver the intervention in an almost kind of rote kind of way.” [Grace_COVID1]

During lockdown, Grace established contact with Keith and provided five intervention sessions, lasting between five and 45 minutes via phone and video (MS Teams) with additional input via email. Keith had a long-standing goal to re-establish a previously unsuccessful small business, and in addition to PARTNERS2, was receiving cognitive behavioural therapy. Having identified and explored this goal, Grace encouraged a stepped approach to achieving it, which she summarised with Keith:

Grace: It’s about breaking things down, isn’t it… bit by bit.

Keith: Yeah, how am I going to get there and, do you know what I mean, achieve to do that… I totally understand that, yeah, I get that.

Grace: Yeah. What information do I need, you know, for this business, what’s going to help me, do I need – I don’t know, I’m thinking off the top of my head here, do you need some accountancy, you know -

Keith: OK.

Grace: So, it’s breaking stuff down, that’s your ultimate goal, isn’t it, it’s to go back into business, but because things sort of fell apart the last time… this is going to sound really strange, but in a good position, really, to learn from your mistakes. [Keith_Grace_SES_B]

Post-intervention, Keith reported that Grace had successfully facilitated a collaborative relationship, by maintaining contact between the digital intervention sessions, and that the development of trust had been fundamental to their work together:

“[Grace] has always been bang on with me… if she’s on annual leave or anything, she’d let me know… if I’ve had an appointment and I’ve not been able to do it, I’ve also just emailed her and she’s emailed me back or texted my phone, so, she’s good… If I can trust somebody… I won’t say I keep them at arm’s length but I still try and find the negatives… Yeah, she’s never given me no doubts and I trust her.” [Keith_Grace_TAR_B]

Grace’s own reflection on her practice suggests a more nuanced understanding of the model, using lockdown as an opportunity to focus on smaller goals:

“What I might have done is said “What can we do in the meantime?” In the here and now... ‘Cos when you think too far into the future with the restrictions that we’ve got in place, I think that we might increase people’s anxiety. So, I think keeping people in the here and now and what we can do here and now… It’s yes, we’ve got restrictions, but we’ve all got to live by these restrictions, haven’t we… But it doesn’t mean that we can’t go back to them at some point when things improve.” [Grace_COVID2]

Delivering the PARTNERS2 model digitally was more difficult where care partners were less experienced, and existent models of service delivery inhibited the integration of the PARTNERS service. For example, Sarah was new to the PARTNERS service at lockdown. With less experiential knowledge of PARTNERS2, Sarah drew upon her previous experience of working in a Community Mental Health Team (CMHT):

“I just used my gut instinct and leant on my skills that I’ve picked up working in the mental health frontline services for the last 17 years” [Sarah_Reflective_Practice_Log_March_2020]

Sarah worked across a large rural and socially economic deprived area and perceived that working across three CMHTs meant that few service providers really understood the aims and objectives of the PARTNERS model as few service users had been allocated to her caseload from each team. She suggested that remote working impeded communication with the health professionals with whom she had been engaging, but with whom rapport had not been fully established prior to lockdown:

“I think PARTNERS would be a key part to merging that primary and secondary care, and actually taking the responsibility off secondary care as well… Actually, meeting with the CMHT and meeting with those key people that have been caring or care coordinating those individuals to say this is what it is, they’ve signed up [to]… the care partner, does this, this and this. That’s not been clear in [location]… I feel like I’m a spare part, sometimes, I’m not in the greased cog, so, sometimes I’m not communicated with as regards to, like, care plans or if there’s been an incident …” [Sarah_COVID]

These findings suggest that, if care partners are already familiar with and confident at delivering the PARTNERS model, or able to draw upon experiential knowledge aligned with the philosophies underpinning the programme theory, then the PARTNERS model can effectively be delivered digitally.

Liquid uncertainty

Bauman’s work extends the metaphor of an ever fluid society to the idea that people behave like particles, and have fewer strong bonds with other people; but that they are ‘assumed to be capable of designing our own lives and mustering everything needed to pursue and see through our life objectives’ [25: 101]. Under these conditions, Bauman cautioned that the widespread adoption of digital technologies has the potential to undermine the ‘veracity of realism of a mutual person-to-person trust’ such that its content and significance can be profoundly reduced [25: 15]. Our examples suggest that digital competency does not compensate for a lack of experiential knowledge of PARTNERS. Rather digital delivery can further compound poor understanding of the PARTNERS model, and undermine the potential for a care partner to deliver effective collaborative and personalised care.

Four of the six care partners attended online training in the use of digital video technologies immediately prior to lockdown. This facilitated discussions about the form (e.g. media) and content (e.g. collaborative care) of digital adaptations. Care partners told us that their organisations used a range of bespoke digital platforms that were difficult for them to learn to use, or which impeded communication with their primary care colleagues (e.g. to add meeting notes to service users’ records). These platforms were unfamiliar to the service users that they worked with, and who – if they had access to remote technologies at all - were more familiar with WhatsApp and FaceTime. Over the course of lockdown, some Trusts commenced use of mainstream platforms (e.g. Zoom or MS Teams).

Although Grace was successful at delivering the PARTNERS2 service digitally, this was despite a long-standing anxiety around using mobile phones and video platforms. However, she was able to develop digital competence and role-model this vulnerability with service users as lockdown progressed:

“Whilst setting up video appts I have learned more about service user strengths and where they may need more assistance to get to this point. I have also personally been able to become a little more proficient in using the technology and although I am still anxious using video, that I am more comfortable with its use.” [Grace_Reflective_Practice_Log_April_2020]

Two months later, Grace discussed a more developed digital proficiency:

“People don’t realise that when I send an invite that that link actually leads them to opening a Teams account, yes, so they’ve got to register, first of all, and then they can join the meeting with myself… Just sort of trying to show people how to actually access the system… I’ve stayed on the line with them, sometimes on the phone, whilst they’re trying to navigate their way through the system.” [Grace_COVID2]

Whereas Grace shared her digital vulnerabilities as a means to build rapport and develop a shared understanding, Hannah emphasised a pro-digital aptitude that pre-dated her role as a care partner:

“When working in IAPT [Improving access to psychological therapies] I did a lot of telephone work, so some CBT [Cognitive behavioural therapy] interventions were telephone, lots of assessments were telephone. In [other] work I see people across the country on Zoom... [I prefer] video… Because we can see each other, we can have a deeper connection and there’s possibilities to use software such as screen sharing, that’s easier. Yeah, picking up on non-verbal cues, I think, is really important… Knowing not to talk over each other and seeing who, you know, when somebody’s speaking, I think that’s so much easier.” [Hannah_COVID1]

Like Grace, Hannah was also able to reflect on the impact of COVID on the nature of goals, and the need for flexibility in delivering the model:

“I think the type of goal will change because of the social circumstance, I don’t think it’s because of the method that we’re using, I think it’s because of the situation that’s causing us to use these methods! … But we can be flexible and it will be fine, it’s just trying to be really adaptive and think about smaller steps, remote steps, and I guess a heavier usage on the voluntary sector than possibly before, voluntary and charitable sector.” [Hannah_COVID1]

At lockdown, Hannah had one years’ experience of being a care partner. She had met Alice prior to lockdown, and subsequently they met six times during lockdown via phone, text and Zoom (with sessions lasting five to 60 minutes). Alice’s specific goal during this period was to become pregnant and manage her antipsychotic medication appropriately, with a longer-term goal to secure safer housing. However, it was only when Hannah and Alice were discussing a housing application, by video, that Alice realised that the PARTNERS2 intervention is time-bound:

Hannah: It’s just us thinking about when was the last letter written [to local Council], is there anything as a care partner that I could do to help take that pressure off [GP] maybe, if I write the letter?

Alice: I think it will be, not this November, next November, when I’m going to need to pull you in and grab your help….

Hannah: So, have I said this to you, it’s for a year, it’s a year long?

Alice: Oh, OK, so you won’t be with me by the time that I get on to [location] Council?

Hannah: No.

Alice: No, right…

Hannah: Did I not explain that before, I thought I had?

Alice: No, I didn’t know that, I didn’t know…

Hannah: So, the project runs for a year, and then it comes to a close, and we make a plan for the ending, so it doesn’t just feel like it’s all been taken away and cut off, we’ll kind of think about long-term planning and things like that…

Alice: Six months has gone so far. [Alice_Hannah_SES_B]

That the ending of care partner provision was a surprise to Alice is at odds with how intervention ‘endings’ should be planned for and managed according to the programme theory, and this lack of transparency undermined the relationship between Hannah and Alice:

“I was a bit taken aback, ‘cos I didn’t know it was just for a year. And I have been a bit gutted… because I had got used to her and I had put trust into her to speak about things that I wouldn’t normally speak to anybody about and then after a year I don’t really see her again… that’s going to end soon and then what? I’m not going to get no support, I had no support since I was diagnosed with bipolar, only from my family. So, having [Hannah] was a big thing for me… Not just being put on medication and having a medication review every three months, that’s not support, but [Hannah] was… But she’s not going to because I’m not going to be able to join [local] council for roughly about two years… It just feels like I keep getting appointed these people that help me so much and then they just get taken away from me... And it’s a bit gutting. [Alice_Hannah_TAR_B]

“I felt so guilty, I thought I must have mentioned it, but I hadn’t explicitly said the time duration and things like that… And if I was to do it again, I would make sure that that did not happen!... [SU] obviously thought it was years of work, it feels like, ‘cos this application she’s talking about is in a couple of years’ time.” [Hannah_Alice_TAR_B]

Post-lockdown Alice informed Hannah that she had become pregnant and discontinued her antipsychotic medication because she feared that it might harm her baby, however stopping her medication led her to report that she felt unwell. Subsequently Alice cancelled planned intervention sessions with Hannah and withdrew from the study.

That digital technologies could further inhibit intervention delivery is illustrated by Nora’s use of both phone and video. Pre-COVID, Nora had required additional training to deliver the model as intended. She worked with Tina during lockdown– a service user who was dealing with a range of health challenges, while awaiting the release of her partner from prison. She had met with Nora twice prior to lockdown, and contact during lockdown was via three telephone-based intervention sessions, which lasted between five and 26 minutes. During the intervention sessions Nora sometimes worked with Tina’s suggested goals, but at other times was more prescriptive:

Nora: Just thinking about you and your relationship with [partner] at the moment, what sort of things are going to be in place to help you feel safe?

Tina: …Um, I don’t know at the minute. I’m trying to get partner to try and work out some tools that we can use, and I’m sort of thinking about it myself, so that we can come together and have a – almost like a care plan between us. Plus we speak to a family therapist every fortnight…

Nora: Over the last few months we’ve been able to establish a shared understanding about the problems and difficulties, you know, you’re trying to manage every day…However, the sort of general wellbeing practitioner that I am sometimes thinks that, you know, maybe if there was some more education or support available around your diabetes and your budgeting…

Tina: Yeah, well, to be honest I spend so much on what I eat. Ok it’s only been the last week that I’ve been to [local shop]…

Nora: We spoke a little about you wanting to think about giving up smoking because you know that doesn’t help your physical health although at this time that’s not the right time for you to be thinking about that.

Tina: right, well, actually I’ve thought about it. I don’t want to give up smoking. [Tina_Nora_SES]

In the subsequent tape assisted recall, Nora suggested that remote goal setting and structuring sessions with Tina via telephone was difficult, as it was harder to assess her remotely, despite having had input about both during training and subsequent supervision:

“It’s very difficult staying on track with this lady, she has a diagnosis of emotional unstable personality… I wanted to try and just get her back to talking about her physical health issues so that I could hear that she’s actually aware of what she’s doing… when you ask her about her future or you ask her about making goals, they’re predominantly made around things that can’t be solved automatically… Through the shared understanding, you know, I really do understand how difficult things can be…[Meta-supervisor] and I talked about having an agenda….So, you know, when I’ve ended, we’ll say, OK, so when I’m reflecting on what we’ve talked about I say “So, today we’ve talked about X,Y and Z, how about next time our agenda be about, let’s see what’s happened with X,Y and Z?” . [Nora_Tina_TAR]

This need for structure was similarly recognised by Tina, who also suggested that Nora’s intervention delivery by phone could make the intervention sessions feel rushed, and diminish their therapeutic value:

“Sometimes I think that we could do with a little bit more structure… But on the phone we can’t really do that… it’s quite difficult to actually come away from the phone and think “What was that about?”, because it’s gone… she might ask a question but she can rush past it to something else, when actually I needed to finish what I was saying, so I feel quite rushed about that…Over the phone it becomes more of a chat than a sort of therapeutic thing, sometimes.” [Tina_Nora_TAR]

James had also met Nora in person prior to lockdown, and was living with his parents while looking for work. He told us how deeply uncomfortable he was about participating in intervention sessions via video from his bedroom, and he was insightful about what had been missing from the PARTNERS sessions for him:

“I don’t think there was enough times that [Nora] and I connected… the first thing I said would be the thing that would become the focus and it wasn’t always the most relevant or important thing… at least two or three times she said she would send me emails and I never received anything… I just have a feeling that it just felt a little bit shallow, maybe…. And I think [Nora] could have been that person, but there just wasn’t enough fuel in the engine to get it going, you know.” [James_INT_PM]

This ‘lack of connection’ was particularly manifest in relation to goal setting. James was a science graduate who, pre-lockdown, commenced work in a distribution centre. For James this was a means to an end, rather than a long-term goal:

“I appreciate that a major part of [Nora]’s job is trying to get people to focus and maybe find a new way of thinking and have positive steps ahead of them that they can try and achieve, so she was obviously referencing back to when I first started talking to her on that day of wanting to get a car, that’s my primary reason for having a job, so yeah, that felt relevant. I think I remember thinking that maybe there wasn’t enough” [James_Nora_TAR_PM]

By the time of the follow-up interview, James had left his warehouse job and was being supported by an employment service to find a more fulfilling alternative.

Our analysis identified that, if care partners are confident at using digital technologies, or comfortable with being vulnerable so that they can model their learning with service users, then digital delivery can effectively underpin delivery of the PARTNERS service.

An enduring underclass

In Bauman’s account of modern life, liquefaction – the ever-increasing pace at which change occurs – leads to people ‘falling outside of society’ [25: 3], with those already experiencing material hardship and mental health challenges particularly at risk of being left behind. Bauman emphasises that these people have not done anything wrong but are ‘collateral damage’ to the socio-economic challenges that humans have failed to resolve [27]. Above, we identified some of the conditions under which PARTNERS2 can be delivered digitally by a care partner to support people with severe mental illness to identify and work towards their goals when existing resources are optimal. However, in keeping with Bauman’s notion that the excluded become ever more so, we also identified a ‘digital underclass’ who by virtue of their (often multi-layered) socio-economic circumstances struggled to participate in digital intervention sessions [26].

In addition to prior experience of PARTNERS and digital competence, our analysis identified organisational and structural factors which could inhibit service user engagement. Sarah told us that during lockdown the simultaneous operation of both PARTNERS2 and the existent model of service delivery was prohibitive:

“I’ve come to realise, as I’ve been working on PARTNERS… that actually the practicalities of taking on that caseload of people that are on that surgery aren’t always as clear cut. So, I’ve got people on PARTNERS that are still being seen because they’re on low-intensity caseload, which means they have a serious mental health issue but they’re not really getting a lot of intervention. But they are still seen by a CPN [community psychiatric nurse] because they’re on, like, a mini treatment order and for some people they still get to see a support worker from the mental health team. So, they’re not totally discharged as such, or given to me to care coordinate wholly, like the manual suggests…” [Sarah_COVID]

Sarah also told us explicitly about the poor socio-economic circumstances in which many of the service users live. Here, living remotely meant that they had no access to any kind of technologies:

[Rachel] She’s living in a lorry… It’s harder to do it remotely because say I went to visit [Zoe]… she lives in a caravan… I think mental health, working with these people, these diagnoses, it’s really important to have that personal approach, it’s about them, yeah…. [Sarah_COVID]

Other care partners similarly explained the impact of wider structural factors. Older people, often with more enduring mental health problems, were less likely to be comfortable with digital technologies or have access to smart phones or the skill set required to set up new technologies at distance:

“My patients preferred to be on the phone. And that’s probably an age thing… they’re not so tech savvy…[One] would have been open-minded enough to try it, but it would have taken a lot of the session time just to explain how to go through downloading apps…” [Emma_COVID1]

I tried to do one with somebody [Geraldine], a face-to-face video… it’s her son who’s the service user, but they always come together… we spent quite some time trying to find the camera and switch her camera on, not having seen her computer at all… In the end we gave up, and we went to phone. And she called her neighbour to try and help her to set up Teams, only to discover that her computer does not have a camera at all. And, I mean, I’ve never heard of a computer without a camera! [Grace_COVID2]

Indeed, Hannah noted that the technologies with which service users engage could be both signifier and moderator of ‘sociability’:

“Those that have the access to online, like remote working… Skype and Zoom, they also might have more options socially, like they can use this to be more social. Whereas the people that are just on telephone they might not even have a smart phone [Jeff], some of mine don’t, and they have no other means of social contact, so I feel like for some reason, they’re coming into, you know, the people with less technology are also then having less social interactions with people, is what I’ve noticed… for some people we’re not going to be able to do more than talk because of what they have access to…”[Hannah_COVID1]

For Hannah, this disadvantage not only impacted on their access and use of digital technologies but that this mediated the extent to which they were able to engage with the model more generally:

“The people on video are the ones where we’ve been more clear about what the goals are, the people on telephone it’s more difficult and it’s taking longer… But three of them out of the four that I’m thinking are very high functioning full-time workers with high levels of education. So, it might be more the way that they think, you know, they might be thinking that way more usually than people that aren’t engaging in work or are more unwell right now, possibly.” [Hannah_COVID2]

However, Nora explained that for some service users, it was the nature of their symptoms which made using technologies particularly challenging:

“There’s a man that I spoke to today for the first time and because of what he goes through in his experiences, using anything digital is really difficult for him, and even though he’s using the phone it really triggers his voices and, you know, he really believes that he could be being listened to or observed in a way that, if you’re alone together I would say that he doesn’t as much… So, I think for him it’s probably taking a lot to actually attend just on the phone, because of how much his psychosis relates to being spied upon in a digital world.” [Nora_COVID2]

These findings suggest that, if people who are already unwell and living with limited access to resources and opportunities are provided with digital solutions that are not fit for purpose, then digital delivery cannot effectively underpin delivery of the PARTNERS service. To do so would require the provision of appropriate digital platforms in primary care (or indeed reversion to face to face service delivery), as well as adequate staff to support their use, but more significantly greater societal impetus to address deep engrained structural inequalities.

Discussion

Supplementing our existent realist programme theory with a long-range theory specifically focused on the more existential challenges of living through a time of crisis, enabled us to reflect on the flexibility of the PARTNERS model and the potential of digital adaptation. In effect, this enabled us to test Bauman’s hypothesis that, in a ‘software world’ that encourages individualisation, as uncertainty increases what some of us might actually require is ‘more society’ or structured network of support [16].

We identified that incongruence between NHS protocols and research practices inhibited the rapid digitalisation of the PARTNERS intervention, and adaptation with video, as opposed to only phone/voice connectivity was only optimal where care partner skills, organisational policies, and service users’ resources were aligned. We needed to revise the PARTNERS programme theory for remote delivery (Figure 1), and emphasise the need for practical solutions to use easily accessible digital communication, as well as support some staff and service users to take on and routinise new ways of connecting. Knowing even then it may not be enough to support true individualisation of support through deep interpersonal connection.

Fig. 1

Liquidity and Uncertainty

For effective digital delivery, care partners needed to understand and engage with the intervention and have the skills to deliver it. Specifically, the ability to establish a shared understanding digitally was central to effective care partner- service user communication during lockdown [28, 29]. Under these conditions, care partners were able to employ coaching to enable the service user to identify personally meaningful goals and mobilise relevant resources to manage their own health and wellbeing [11]. However, for coaching to be effective, both the care partner and the service user had to actively contribute and invest time in exploring what is possible [30]. Goal setting required the development of a trusting relationship and the development of rapport [31] but not all care partners developed the knowledge required to deliver these aspects of the intervention [11]. This was particularly difficult for care partners who were new to the PARTNERS service at the time of lockdown and who lacked prior knowledge of delivering the model. Both the care partner training and intervention manual emphasises that care partner needs to be flexible and personalise the intervention to service user needs [11], and we identified that an individualised approach was sustained when regular supervision supported an enduring change in practice [32]. Where supervision was not provided, care partners could deliver the model by rote or draw upon other modes of professional practice, which may not have aligned with the philosophy of PARTNERS.

For digital delivery to be effective, a care partner also needs to be confident at using digital technologies; or be comfortable with their vulnerability to using technologies which they could harness in practice with a service user. This was difficult because of the wider context (e.g. a global pandemic, in which care partners and service users were experiencing their own ontological insecurity, and where agile technologies were often lacking) [33]. Tele-psychiatry is an established method for providing support to people experiencing mental illness, via video technologies or telephone, at distance. It has been used to provide collaborative care and support a therapeutic alliance [34]; although there are concerns regarding selection bias in studies, such that service users amenable to tele-psychiatry are over-represented [35]. We identified that the use of video to deliver PARTNERS was appropriate for service users with psychosis, and some expressed a preference for this form of remote delivery [36]. Counter to studies that report that video can provide more equitable service provision to people in remote locations [34, 35], we found that people in remote areas had poor access to broadband and were least able to access computers and smart phones. Some service users and care partners reported that the use of video could diminish the quality of interpersonal interactions [37, 38]. Telephone delivery could be used to establish ‘psychological closeness’, in the absence of physical proximity [39], however PARTNERS was not designed as a digital intervention and was therefore reliant upon post-hoc technological adaptation [34]. This was in a context of the lack of consistent guidance or prioritisation of need across host organisations [40] and inequality of access to digital technologies (for both service users and care partners), due to the speed of the lockdown, but also enduring socio-economic disadvantage [41].

Bauman’s conceptualisation of liquidity and uncertainty informed our understanding of using remote delivery methods during a time of increased ontological insecurity. Bauman argued that the constant state of change associated with postmodernity would increasingly disenfranchise people that struggled to ‘keep up’, such as those experiencing mental illness or social deprivation, and that this would be compounded by the withdrawal of the welfare state and other institutions that provide a safety net [42]. While we identified the configurations in which the remote delivery of the PARTNERS model work, we acknowledge that some potential service users were ‘left behind’ in the rush to digitalise the intervention for lockdown.

Bauman argued that any act of exclusion of people who are viewed by the state as unproductive would render them even more isolated [26], such that they would be required to find and navigate ‘individual solutions to socially produced troubles’, made even more challenging for those outside of cyberspace [43]. Portentously, Bauman considered that pandemics had the potential to magnify our sense of liquid uncertainty and create further ‘collateral damage’, but that our route to recovery would be via open discussion with a wide range of people in order to widen the possibilities available, rather than develop a ‘one size fits all’ solution [27]. We recommend that a digital PARTNERS2 service would benefit from more contextualised training and supervision of care partners. More generally we propose that strident efforts are needed by researchers, service providers and service users to work together to develop and provide agile interventions that can be individualised not just therapeutically but in the mode of delivery to service users’ needs, otherwise the most severely mentally ill and deprived service users will remain disenfranchised at great cost.

We caution against seeing the digitalisation of mental health services as an efficient form of ‘levelling up’ [44]. Rather we see it as a potential magnifier of limited effectiveness, with a significant proportion of those who are already disenfranchised at risk of sliding further from view [41]. This is in keeping with concerns about the impact of digitalised health care provision and the lack of impact assessments for patients who are most socially deprived [45]; and assertions that the lockdown was a composite of existing mechanisms (e.g. ‘social segregation’), and contexts (e.g. ‘relative isolation’) which produced significant disparities in healthcare - which we now risk sustaining [25]. People with SMI require a wider set of opportunities and choices for engagement and participation, and mental health services need to address the social determinants of inequalities, if care is to be personalised and collaborative (Mental Health Taskforce 2016; NHSE 2016, NHS 2019, NHSE 2019) [46,47,48]. We are now implementing our learning from PARTNERS2 within the policy programme for community mental health transformation (NCCMH 2019) [8]. This involves training and supporting mental health workers to deliver the PARTNERS2 model, within a new place-based, multidisciplinary service that is attuned to the multiple layers of disadvantage that many people with severe mental illness experience. To underpin this new learning we are working with system change leads, using meta-supervision and peer support to foster an enduring, yet agile, community of practice. It is in this local and supportive context that Bauman believed that the most vulnerable have the potential be re-embedded in society (Bauman 2000) [16].

Strengths and limitations

A strength of this research is that we captured evidence of the plausibility of mechanisms, in order to offer theoretical generalisations about how an intervention might operate under different conditions to those in which it was designed [49]. We planned to deliver a ‘rapid’ process evaluation specifically about remote delivery, but in keeping with best practice guidance, we have balanced the collection of trustworthy data with speed [50]. We built multi-modal case studies around individual practitioners, and mapped their pre- and post-lockdown/digital practices across their work with different service users [51]. Data pertaining to all of the included care partners, and from service users who had a range of experiences, allowed us to explore the ‘predictive uncertainties’ of continuing to provide a trial of an intervention in an unprecedented context and with a composite of mechanisms (e.g. social isolation and enforced closure of voluntary sector organisations) [25]. This enabled us to identify and address the most salient uncertainties regarding whether remote delivery was viable and the extent to which delivery of the model was equitable [52].

A limitation of this research is that we were unable to collect recordings of service user interactions during lockdown, due in part to existing NHS protocols that did not allow for digital service delivery or remote data capture [14, 15]. Our subsequent implementation study is seeking to address some of these concerns, by embedding researchers in host organisations and employing ethnographic techniques.

A further limitation is the lack of service users from diverse ethnic backgrounds included in this dataset. Lockdown coincided with increased COVID-19 mortality rates for different black, Asian and minority ethnic communities, as well as the Black Lives Matter protests, both of which have contributed to intensified fears and anxieties [53, 54]. Further research must explore the needs of service users from diverse ethnic and cultural backgrounds, plus any preferences for digital or face-to-face service provision.

Conclusions

PARTNERS2 can be delivered digitally by a care partner to support some or perhaps most people with severe mental illness to identify and work towards their goals. To be effective, the care partner needs to be familiar with and confident at using the PARTNERS2 model, or be able to draw upon experiential knowledge of practice aligned with the philosophies underpinning the programme theory. The care partner also needs to be confident at using digital technologies, or comfortable with being vulnerable so that they can model their learning with the service user. Service users are receptive to PARTNERS2 being delivered remotely when they have had the opportunity to establish rapport with a care partner at their own pace, when they are comfortable with and skilled at using digital technologies, or are able to be receptive to learning how to use them. However, at a time of increased need, we identified that people who are very unwell and living with severely limited access to resources and opportunities, remained disenfranchised at great cost. Its is highly like these individuals will need proactive face to face support.