Exploring the efficacy of a low intensity psychological intervention for family members and carers of individuals with a first episode of psychosis in early intervention services

The study aimed to evaluate the efficacy, and explore carers' experience, of a brief carer focussed intervention in an Early Intervention in Psychosis (EIP) service using a mixed methods approach.


| INTRODUCTION
Individuals with psychosis benefit from carer support.Carers facilitate access to mental health services, recognize signs of relapse and improve engagement between the patient and services (Eassom, Giacco, Dirik, & Priebe, 2014;Jens Einar Jansen, Pedersen, Hastrup, Haahr, & Simonsen, 2018;Onwumere et al., 2018;Onwumere, Shiers, & Chew-Graham, 2016;Sin et al., 2017).Conversely, the caring role can create substantial stresses and long-term strain, impacting on the carers physical, social and emotional well-being (Charles et al., 2021).Carers may have a broad range of needs, relating to psychoeducation, emotional support, coping strategies, self-care and dealing with a crisis (Onwumere et al., 2016).This is particularly relevant in the early illness phases, when a carer may experience distress and increased worry (Onwumere et al., 2016).
High expressed emotion (EE) characterized by high levels of criticism, hostility, and emotional over-involvement (EOI), may be a family factor negatively contributing to the care of those with psychosis (Kuipers et al., 2006;Meneghelli et al., 2011) which may contribute to poor clinical outcomes and more frequent hospitalizations (Alvarez-Jimenez et al., 2012).Providing psychoeducation is effective in reducing EE in particular hostility and criticism (Öksüz, Karaca, Özaltın, & Ates ¸, 2017).A recent study looking at the effect of psychoeducation on illness perception and EE of family caregivers of patients with schizophrenia found that family members' EE significantly decreased below the baseline after psychoeducation in the intervention group but not in the control group without education (Budiono, Kantono, Kristianto, Avanti, & Herawati, 2021).Mitigating EE within an intervention may also reduce relapse rates (Bebbington & Kuipers, 1994).
Psychoeducation and carer/family support interventions in psychosis are essential when offering care that is compliant with NICE and the early intervention access standard ('National Institute for Health and Care Excellence: Guidelines ', 2014;Onwumere et al., 2016) and have been shown to reduce carers' global, perceived burden, and negative caregiving experiences (Sin et al., 2017) This study aimed to examine whether a brief (up to 8 sessions) carer intervention in EIP services in North East London NHS Foundation Trust (NELFT) would increase the overall wellbeing of carers and reduce carer burden.We hypothesised that successful completion of this intervention would improve wellbeing and reduce caregiving burden and anxiety.

| Design and participants
A mixed methods approach was used to evaluate the effectiveness as well as the experience of the intervention.The analysis of qualitative and quantitative data followed a parallel data analysis approach (Erzberger & Kelle, 2003).
The sample consisted of carers or family members of service users with a first episode of psychosis under the care of EIP services in NELFT who are routinely offered the carer focussed intervention.
The service offers the intervention to multiple members of the same family separately.If carers declined the intervention when it was first offered, they were re-offered the intervention regularly (every 3 months).All quantitative data (n = 153) were part of routine outcome measure collection and included participants' informed consent to publish.
Nine carers completed semi-structured interviews.The average number of sessions attended was five.

| Intervention
The 'low intensity' (LI) intervention is informed by a Cognitive Beha-

| Psychoeducation
Carers read through an information booklet, designed by Onwumere (1st edition, December 2001, adapted from Kuipers, Leff & Lam (1992)) in collaboration with the assistant psychologist pausing after each section to reflect on how this information might be relevant for the carer, their family and their loved one experiencing psychosis.
Topics covered in the booklet include crisis helpline information, symptoms of psychosis, useful resources, causes of psychosis, and effects of caring for someone with psychosis and treatment trajectories.

| Wellbeing
This component looked at the idiosyncratic needs of the carer and the impact of their caring role on levels of anxiety and depression, selfcare and quality of life.The practitioner was able to choose from a menu of LI interventions relevant to the carers goals, including: problem solving, sleep hygiene, time management, or LI CBT for depression/anxiety.

| Carer wellbeing plans and signposting
The last 1-3 sessions of the intervention were used to summarize key learning points and highlight areas of focus for the future, including working on a relapse prevention plan.Sign-posting is relevant if carers have continuous needs and would benefit from other services, such as Improving Access to Psychological Therapies (IAPT).The Warwick-Edinburgh Mental Well-being Scale (WEMWBS) (Tennant et al., 2007), a 14-item scale of mental wellbeing routinely used in NHS services examined subjective well-being and overall psychological functioning.

| Measures
The Hospital Anxiety and Depression Scale (HADS) (Zigmond & Snaith, 1983), a self-assessment scale for symptoms of depression and anxiety with high reliability and validity was also used.

| Quantitative
The data was analysed using R. Paired t-tests were conducted using both a Bayesian and NHST approach.Where the NHST paired t-test may suggests that the data are unlikely to be observed if the symptom scores were identical at both timepoints, the Bayesian analysis adds to the interpretation by indicating which hypothesis is more probable after observing the data (where a large BF10 indicates support for the alternative hypothesis).For parameter estimation, we assigned a Cauchy prior distribution with r = 1/√2.Gaussian graphical models (GGM) were estimated to explore the relationship between constructs measured at each timepoint, and test whether the intervention not only altered attitudes, experiences and symptoms as tested in the paired t-tests, but also whether it altered the association between constructs.This approach is fitting with the Causal Attitude Network model (Dalege et al., 2016).Mapping the patterns of change may help elucidate the potential mechanisms underlying the potential efficacy of the intervention.
GGMs were estimated with least absolute shrinkage and selection operator' (Lasso) penalization and using an iterative modelling procedure: the Extended Bayesian Information Criterion (EBIC).With aim of estimating a conservative model (reducing false discovery rate), the (EBIC) hyperparameter was set at 0.5 with an edge weight threshold of 0.2.In the final model edges weights were set to zero if they did not reach this threshold, as well as in the EBIC computation of all estimated models.Missing data were handled using pairwise deletion.

| Qualitative
A semi-structured interview was used to explore experiences of caring for someone who is under the care of EIP team, their experience of being offered and having the intervention and the impact of the intervention.The interviews lasted between 45 and 60 min and were audio-recorded, transcribed and analysed utilizing the guidelines for implementing thematic analysis (Braun & Clarke, 2006).This included familiarizing oneself with the data, coding, constructing themes and reviewing themes via consultation with clinical psychologists.

| RESULTS
Three hundred and eight carers were offered the intervention between January 2017 and June 2020.Of these, 153 carers were included in the analysis having both accepted the intervention and completed pre-and post-intervention outcome measures.The mean number of sessions attended was 8 (range: 2-18).Table 1 outlines  The network graphs (Figure 1) visualize the partial correlation matrix at initial and post timepoints.While the presence of edges are conservatively estimated (threshold for inclusion of an edge was r = 0.2), the accuracy of the strength of the edges highlights wide confidence intervals (estimated using 1000 bootstraps) and as such differences between the strength of edges are not interpreted.
At the initial timepoint, the association anxiety, depression and wellbeing were strongly associated, representing psychopathology.
Only anxiety was associated with the caregiving experience through stigma/effects on the family (r = 0.24).Within the experience of caregiving factors, stigma / effects on the family was positively associated with difficult behaviours (r = 0.32), and negatively associated with positive personal experiences (r = À0.26).While positive personal experiences were negatively associated with problems with services (r = À0.31).Post intervention, the network structure was similar to the initial timepoint structure (r = 0.77).There were some observed differences.

T A B L E 1 Sample demographics
The association between anxiety and stigma/effects on the family was no longer present, coinciding with a reduction in severity of anxiety from the initial timepoint.However, depression was now associated with stigma/effects on the family (r = 0.21).Wellbeing was surprisingly positively associated with problems with services (r = 0.22), both displaying an increase in mean scores from the initial timepoint (although not significantly for problems with services).
Associations within factors of the experience of caregiving remained the same apart from the emergence of a direct, negative association between problems with service and stigma/effects on the family (r = À0.27).both the offer and the CWI itself was mentioned as being central to facilitation of this connection (see P3, Table 3), and being offered it early was seen as helpful (see P4, Table 3).Being linked in: connection seemed beyond the sole relationship with the clinician delivering the intervention.The intervention is read as having helped carers understand roles within EIP and this connectedness with the wider team seemed to foster a sense of trust (P2, Table 3).Being remembered:

| Thematic analysis of interviews
Being contacted by the team appeared to be linked to feeling held in mind and more able to cope (P7, Table 3).All sub-themes are read as promoting a reciprocal nature to the connection to services and staff.

| Being valued
Across interviews, there were frequent references made to experiencing a sense of feeling valued post intervention.Personalization: Numerous reports were made to the personal approach separate to just reading or a carers forum.Participant 8 (Table 3) is read here as meaning that the individual space allowed for open discussions with the opportunity to personalize questions and seek clarification.Selfvalue: The way in which carers began to value their own needs was frequently mentioned, as highlighted in Participant 1 (Table 3) and for Participant 5 this seemed linked to feeling more empowered.
'…the fact there was someone who was happy to listen

| Change
A prominent theme concerned the changes they that carers specifically connected to the CWI.The sub-theme of re-framing was read as referencing a new understanding of previous and upsetting events that had occurred with the service user which helped support a shift in how they could respond (P3, Table 3).The sub-theme, language and knowledge, had frequent reference to the importance of having the 'right' language to facilitate communication.All participants discussed that sessions provided them with better terminology, with some discussing specific crisis planning sheets helping them feel more prepared (P5, Table 3).The change due to the intervention was indicated by improvement in well-being and a reduction in self-reported anxiety, consistent with previous research (Onwumere et al., 2016;Sin et al., 2017).Furthermore, difficult behaviours and stigma/effects on the family were also reduced.Post intervention we observe the relationship between depression and stigma/effects on the family in the absence of a relationship with anxiety which was present pre-intervention.This highlights that there might be a pathway between anxiety and stigma via The intervention provides a simple and easily implemented approach to address carer and family needs in community settings and contributes to practice based evidence.Engaging carers and promoting ways to formulate a joint understanding of psychosis may reduce appraisals of loneliness and de-stigmatize the experience (Allard, Lancaster, Clayton, Amos, & Birchwood, 2018) and support clinicians in their attempts to engage patients effectively (Jansen et al., 2018) and promote positive illness management (Onwumere et al., 2018).Importantly the intervention was delivered by assistant psychologists and therefore has potential to utilize a stepped care approach and is a cost-effective model which enables delivery of the intervention by other professional groups with potential implications for workforce planning.

| LIMITATIONS
A key limitation of this study is the absence of a measure evaluating the effects of psychoeducation.Our results, in particular the sub themes of reframing and language and knowledge support the hypothesis that the effectiveness of psychoeducation is contingent on carers' knowledge about psychosis, their cognitive appraisal about the caring situation, and subsequently, their perceived burden and (self-efficacy in) coping with caring (Albert, 1978;Birchwood, Smith, & Cochrane, 1992;Lazarus & Folkman, 1984).
Similarly, EE as an outcome measurement would have enabled further understanding of the relationship between expressed emotion, psychoeducation, wellbeing and overall experience of caregiving (Jansen, Gleeson, & Cotton, 2015).Reduction of EE is an implicit effect of the intervention as highlighted by the sub theme of behavioural changes related to style of communication and expression of concern.The focus on improving communication and knowledge about psychosis might lead to an empathic stance and reduced EE and might impact on relapse.
The evaluation of routine service provision limits the generalizability to other carer populations and settings, although it strengthens its ecological validity.In the absence of follow-up data, it is difficult to comment on the durability and overall impact of the gains on carer wellbeing or patient outcomes.Future evaluations could replicate the intervention using a controlled design to test its efficacy and better understand mechanisms of change and longitudinal impact of the intervention.
To conclude, a brief low intensity intervention focusing on psychoeducation and carer wellbeing that can be delivered by assistant psychologists has the potential to improve levels of anxiety, wellbeing and reduce negative aspects of caregiving.
. Psychoeducation is commonly delivered via individual or group programmes or via internet or mobile app and can involve clinicians taking on the role of the information provider.Considering the increase in demand for treatments, it is necessary to explore alternative methods for delivering psychological interventions.Task-shifting to lay health workers is one such alternative.It involves shifting tasks, such as the delivery of psychological intervention from highly qualified health workers to workers with fewer qualifications and minimum training, to increase the coverage of health care and to use resources more efficiently (McInnis & Merajver, 2011).Several trials have successfully used task-shifting to deliver psychological interventions for a variety of mental illnesses.A Cochrane review investigating the effect of non-specialist health workers on people with mental health conditions found that compared with usual care, interventions delivered by non-specialist health workers may increase recovery of common mental health disorders (van Ginneken et al., 2013).Little is known about the efficacy of psychoeducational interventions delivered by 'non-clinically trained' practitioners such as assistant psychologists; however, this has implications for workforce planning, the delivery of such interventions and their cost effectiveness.
vioural framework and designed to be completed in eight 1 h, one to one sessions, delivered weekly, in-person at the service.Assistant psychologists attend a training programme facilitated by clinical psychologists.They received weekly supervision and monthly peer-supervision, facilitated by a clinical psychologist.The intervention was divided into three key parts: Participants completed outcome measures in the first and final session of the intervention.Measures took approximately 10-20 min to complete.Choice of measures were predetermined by the service needs and mechanistic targets for intervention, focusing on caregiver burden and overall wellbeing and anxiety.The Brief Experience of Caregiving Inventory (BECI) (O'Driscoll, Farrelly, Burgess, Szmukler, & Shaikh, 2018) self-report measure was used to examine the experiences of individuals caring for someone with a serious mental health disorder.BECI is comprised of 19 items and four factors (stigma/effects on the family, positive personal experiences, problems with services, difficult behaviours) and displays good reliability and construct validity (O'Driscoll et al., 2018).
the demographics for the sample, which were collated from clinical records.Paired samples t-tests were performed to compare initial-post scores across measures (outlined in Table 2).Based on significance testing and Bayes factor, there was very strong support for change due to the intervention (significant initial-post differences) for BECIdifficult behaviours (d = 0.45); BECI-stigma/effects of on the family (d = 0.34), WEMWBS (d = À0.89) and HADS-anxiety (d = 0.73).
Three higher-order themes were identified including 'connection with services', 'being valued' and 'change', with a further eight sub themes.Connection with services and being valued indicate successful engagement with the team and development of epistemic trust.
to me.I felt I am leading on it like, in terms of the way I am talking.I was given the forum to say what I wanted rather than just being talked at.It felt helpful.I felt empowered'.(P5).
Finally, behavioural changes, concerns a change in interactions with loved ones and time spent taking care of themselves.This included references to communication and not arguing about differences in beliefs, as illustrated by Participant 9.4 | DISCUSSIONComplementary outcomes from the qualitative and quantitative analysis suggested that a brief intervention, delivered by assistant psychologists has a positive impact on the carers overall wellbeing and functioning, and reduced overall distress associated with the experience of caregiving.CWI was viewed as helping carers feel more connected to the service, more valued and as contributing to beneficial changes in perceptions of their caregiving role and responses.The intervention was seen as providing an understanding of psychosis, including the causes, symptoms and recovery and associated language to facilitate communication with services and their family member.By creating a narrative of what had happened during early stages, helped them to be open to progress.Carers expressed the desire for individualized interventions that focused on their specific emotional distress in addition to psychoeducation.This may have been facilitated by the intervention offering expansions on psychoeducation, which enabled thought to be given to carer wellbeing.
depression.Residual symptoms of depression were associated with stigma/effects on the family.While stigma/effects on the family decreased, people who remain depressed post intervention experienced high levels of stigma.A reduction in stigma may be due to carers developing a better understanding through the psycho education component (Portia Monnapula-Mazabane and Gbotemi, 2022).Surprisingly, we observed a relationship with an increase in well-being being associated with greater problems with services.This might be hypothesised as carers developing more agency round seeking support and interacting with services, whereby previously uncertainty related to anxiety and shame related to stigma may have reduced selfefficacy.A complementary theoretical proposition can be deduced from the qualitative findings: carers feeling an increase in self-value and connection with services which might feed into overall selfefficacy.This is consistent with a recent review which found that the quality of the therapeutic relationship with staff and the service was the single most important factor in determining whether the experience of accessing EIS was a positive or negative one(Loughlin, Bucci, Brooks, & Berry, 2020).Overall, the results highlight that the intervention enables successful engagement with services and development of epistemic trust, that gives rise to the ability for carers to engage with the psycho education in a meaningful way and address their own well-being needs.As such the intervention might have an indirect effect on anxiety mechanisms by targeting knowledge and understanding which helps people to deal with uncertainty and therefore you see a knock-on effect on carer well-being.Given the sample size, these observations and hypotheses are presented tentatively and require replication in a larger dataset as well as additional research to highlight underlying mechanisms of change and the relationship between psycho education and wellbeing and other psychological factors.
Descriptives from initial and post measurement along with statistics from the paired samples t-tests Abbreviations: BECI, The Brief Experience of Caregiving Inventory; HADS, The Hospital Anxiety and Depression Scale; WEMWBS, The Warwick-Edinburgh Mental Well-being Scale.

Table 3
illustrates the themes with example quotes.I sort of thought I'm not sure what else I could learn, but I was wrong…and I could have missed out, asking twice was right to connect me…I think maybe there was something about the timing of when it was offered…like it could have been offered sooner and then I wouldn't have hesitated' (P3) 'The earlier the better' (P4) 'Yes.I don't argue anymore, I can just think I will walk away, before I would have argued my point, argued that what he was saying was wrong.That's helped our relationship, I think'.(P9) 3.1.1|ConnectiontoservicesThreesub-themeswereidentifiedin relation to connection to services; timing, being linked in and being remembered.The timing of F I G U R E 1 Initial (left) and post (right) partial correlation networks.Anx, anxiety; DB, difficult behaviours; Dep, depression; PPE, positive personal experiences; PWS, problems with services; SEF, stigma/effects on the family; WB, wellbeing.T A B L E 3 Identified themes linked in '…the team seems very linked, you know?Like I trusted that key things would be shared, like what I said would get passed on after the session if I asked for it to be'.(P2) Being remembered'I think you just feel like everyone is busy and they won't need me getting in contact…It was so nice to be contacted, to know they thought about me (after his relapse).It definitely helped me manage…' (P7)Being valued Personalization '…I read a lot on the Internet…but she (assistant psychologist) just put it in a different perspective, our perspective.She explained certain things that didn't I understand.If I had a question she would help'.(P8)Self-value'Youknowthey're(the sessions) something that made me then spend a little time on myself instead of worrying about everyone all the time…I didn't do that before' (P1)ChangeRe-framing '…she would be abusive to me.She was very hurtful and it was all part of it.I did not know that before and so it was upsetting…then I knew it was part of her illness so then I just used to…give her the comfort she needs' (P3)'…before we didn't know who to call if a setback or a psychotic episode…I mean we didn't even know what to call it, how to say it.Once you have that language to say on the phone they then know what you are talking about, they know what questions to ask you.It was not knowing how to describe it, and the sessions helped that.' (P5)