Improving community health networks for people with severe mental illness: a case study investigation

Research questions

1. How do people with SMI use resources in their personal networks to support their health and well-being?

2. How do community-based practitioners and organisations support people with SMI to use their personal networks to support health and well-being?

3. How do primary care, community-based mental health providers and other organisations work together to develop effective personal networks for people with SMI to improve their overall health and well-being? What were the barriers and enablers to achieving this?

Research objectives

In order to answer the research questions, the following objectives were specified:

1. to map the personal networks utilised by people living with SMI to support their health and well-being using the CHN approach

2. to identify practitioners and organisations in primary care and community health services that contribute to developing effective personal networks for people living with SMI

3. to identify the enablers and barriers to organisations collaborating to provide effective support to people to develop their personal networks

4. to provide recommendations for practitioners, managers, service users, and health and well-being boards for organisational changes to establish and support, if appropriate, the CHN approach.

In our study the use of the term ‘network’ needs careful clarification. It has two meanings. Firstly it was used as a technical term used in SNA24,25 to describe the structure of ties between various nodes such as people and organisations. 26,27 Secondly it was used in a more general or lay sense to describe the connections to people, places and activities. Throughout the report care was taken to minimise the use of network as a general term, to avoid confusion with SNA, but at times it was included because this was an appropriate description of resource flows and connections within an individual’s life.

Mixed methods

In order to answer the research questions, the study used mixed methods162,163 and a three-stage synthesis process, including two context-specific case studies,164 to explore the lives of people with SMI. The study data collected from individuals with SMI aimed to be both descriptive of what was happening within personal networks using applied SNA,24,25 and be used to derive critical insights into network meaning for individuals with SMI. There were tensions in the approach adopted: we had a framework measuring incidences of connections – people, places and activities – sitting alongside reflexive analysis by lived experience researchers and the core study team. We have taken a priori research questions into the study, informed by published literature, and have carried out further consultations to understand concepts such as health and well-being or explored other published literature to make sense of emerging findings. Alongside these data sources we have integrated insights from other informants providing a practitioner perspective and organisation viewpoints. Thus the mixed-methods study has produced findings both grounded within the data and informed by deductive processing of research questions. The mixed methods were balanced, with no one approach overshadowing others either in the planning, collection of data or analysis stages. 165 Dual competency in methodological techniques166 was achieved by a team approach with all members trained in one method and some conversant in both.

Figure 2 provides an overview of the individual data collection modules and how they inter-relate.

FIGURE 2.

Interaction between study modules: planning and data collection. PPI, patient and public involvement.

A specific focus of the study was using the CHN approach to map connections of people with SMI and follow up with interviews to explore meaning within these personal networks. In network studies, the population under investigation was often highly defined, for example examining the impact of drug use in the social networks of people with bipolar disorder. 167 However, we adopted an inclusionary and broad approach, as the study was exploratory. This broad approach does present methodological and analytical challenges because we were exploring many varying factors in people’s lives, situated in people’s own explanations, which describe heterogeneity, not network categories or types. We were mindful of these challenges and throughout the report emphasise that the work was largely descriptive and exploratory. Thus, we were cautious in our conclusions but forthcoming with potential explanations and definite recommendations based on data synthesis.

Approach to data collection and synthesis

It was important to emphasise that throughout the study, we have been grounded in emerging data and reflexive in making sense of these data. 168 We have worked iteratively and sought to explore the data by returning to literature or to further data collection and engaging our patient and public involvement (PPI) group to draw tentative conclusions. Thus, data sets were collected in sequence. First, network mapping (see Chapter 3) and in-depth interviews (see Chapter 4) were collected, as details from the former were used to recruit purposively for the latter and because the interviews involved reflecting on the mapping content and process. Second, the initial organisational interviews (see Chapter 5) and the last organisational interviews (see Chapter 7) were collected in sequence because the findings of the former, and other data sets within the study, were used both to select interviewees purposively and to refine the topics for discussion. Overall within the study the data were collected and analysed largely in parallel, as recommended, to allow for learning from one data module to inform the others throughout the data collection process. 164 The combining and synthesising of data allows the exact same phenomena to be examined from different perspectives, which should result in a more comprehensive multiperspectival understanding. Mixed methods are best suited to explore complex issues that can benefit from the integration of different types of data; it was at the synthesis stage that this approach yielded benefits. 169

Informal synthesis between the data sets took place throughout the study, as data were shared by the research team and used to both inform other modules as well as identify potential areas justifying deeper exploration; this was achieved by regular team data-sharing and discussion sessions including the PPI group members.

The formal synthesis took place in three stages. The first stage of synthesis (see Chapter 7) was the two case studies which combined data from the network mapping, in-depth interviews and organisational interviews. The goal was to understand ‘meaningfulness’ in our data: we understand this in the sense articulated by others who champion interpretive case studies as seeking not generalisable causal explanations but contextual understanding of the meaningfulness in human experience. 170 The two case studies were not designed to compare and contrast; rather they sought to explore the phenomena under study in a context-specific way. This included examining both how personal networks were influenced and moderated by the local context and also identifying and examining any salient features, within the context of each case study site, and the locally defined opportunities and resources that they might provide. The second stage involved working with our PPI group (see next section) to review the findings from Chapters 3–6 and interpret them using a reflexive methodology based on personal experiences of living with SMI. The final synthesis stage involved modelling the CHN approach to understand if it can usefully contribute to recovery theory and practice. 171 A matrix was used to test our model against data summaries from Chapters 3–7. Our model contained three dimensions: external structures and systems (shaping choice and change); individual agency (how individuals were impacted by their own agency or lack thereof); and available resources (access to people, places or activities).

Patient and public involvement

A key part of the study methodology was the integration of lived experience expertise to understand local community resources in the two sites, the design and redesign of study materials and the interpretation of study findings. A team of six lived experience researchers worked on the study as a PPI group, with five completing the reflections analysis process at the end facilitated by an independent lived experience research consultant (full report in Appendix 1). A summary of PPI input for each module is provided in Figure 3. The two modules with the highest level of involvement in both planning and analysis were SMI in-depth (see Chapter 4) and practitioner interviews (see Chapter 6). The team were recruited to provide a diversity of perspective in terms of lived experience of schizophrenia, bipolar disorder or other psychosis as well as age and gender. We attempted but did not successfully recruit a black and minority ethnic (BME) perspective into the PPI group, which was a limitation of the study.

FIGURE 3.

Patient and public involvement in the CHN study.

Organisation interviews: series 1

Over the initial months of the project, face-to-face semistructured interviews were conducted with the aim of gathering contextual information on the organisation of services, and provision of support for people with SMI in both study sites.

Network mapping

The primary aim of mapping interviews was to collect data from individuals with SMI on their current personal networks. This module involved a number of stages (see Figure 4), which are outlined in more detail below.

FIGURE 4.

Flow chart of network-mapping methodology.

Data collection instruments

The mapping interviews used the following instruments (a copy of these and more detailed methodological information can be found at www.mcpin.org):

1. Demographics: a questionnaire collected key sociodemographic data including age, gender, employment, relationship, education and housing status alongside background data on mental health, such as current diagnosis, medication, length of contact with services and inpatient history (see Appendix 3).

2. SWEMWBS: the short seven-item version chosen to measure current well-being asking participants about their experience of seven different thoughts and feelings over the previous 2 weeks; shows strong validity and high correlation with the longer version. 174

3. Dartmouth CO-OP charts: chosen to measure physical, social and emotional functioning for ease of completion and comprehensive dimensions. They consist of nine measures which ask about physical fitness, social activity, daily activity, pain, social support, emotional distress, quality of life, change in health and overall health. 175

4. Resource Generator UK (RGUK): this measures access to social capital within personal social networks and has four internal scales: access to domestic resources, expert advice, personal skills and problem-solving resources. It has been validated122 (see Appendix 4).

5. Health Resource Generator (HRG): consultations and pilots indicated the RGUK did not fully measure the types of resources that were relevant to the study with its particular interest in health capital. The study team therefore devised a pilot scale in the format of the RGUK and based on resources that were identified in focus group consultations to capture access to health- and well-being-related resources (see Appendix 5).

6. Community health network-generating schedule: the team developed a version of the name-generating and name-interpreting procedure used in SNA76 to map social ties, meaningful activities and meaningful places (see Appendices 6 and 7). This tool emphasised the connections within personal networks that support health and well-being.

The aim of the network-generating schedule was firstly to accurately map networks of people, place and activity, to understand how they were composed and then to understand their impact in terms of well-being. From the beginning, the team were aware of the challenges posed by this methodology when applied to participants with mental illness. Network generation was time-consuming and cognitively demanding, with concerns over accuracy of network recall having been raised by various previous studies,76 including recalling colleagues and members of groups and friends. 176 Extending the method to collect data on connection to people, places and activities and collecting the data on this population could potentially heighten these issues. The team were also aware that talking about potentially difficult relationships could cause participant distress. To tackle this, the network interview process was streamlined to have strong face validity and was designed to be as engaging as possible, with clear protocols for minimising and responding to participant distress. Studies have shown that the choice of name generator questions have a strong impact on the network mapped, impacting heavily on size, density and diversity. 177 Consultations and piloting assessed which network questions were most appropriate for mapping networks. We picked standard name generators and added extra items on the basis of consultations. These were refined based on whether or not they generated new names in piloting (if they did not, they were removed). Activity and place were ‘domain-specific’ or contextual name generators, providing extra prompts that improve recall and accuracy.

The final process is shown in Figure 5. For all three levels of network, boundaries were drawn by limiting connections to those which were current and regular interactions, or which were meaningful to the individual. The mapped networks were at a single point in time – the time of the interview. Questions used to elicit these networks can be found in Appendices 6 and 7.

The sample of 150 was based on a sample size calculation to detect differences in mental and physical health between a higher-need group, such as those in secondary mental health services, and a lower-need group, such as those supported in primary care alone, using a social networks study of day hospital and day centre service users. 46

FIGURE 5.

Final interview process chart.

Recruitment

Recruitment process

The first wave of recruitment was exclusively through primary care. Invitations to take part in the study were sent to all GP surgeries in study sites and the Primary Care Research Network (PCRN) assisted with this process. Study researchers also presented to individual practice managers directly. When surgeries responded with interest, members of the study team met relevant contacts to explain the recruitment process, including the inclusion and exclusion criteria and a recruitment log was used in primary care to document each step of the recruitment process.

A second wave of recruitment was extended to secondary care, through early intervention services and community mental health and recovery teams, as a result of slow response rates in primary care. We targeted teams that covered the same geographical region of the South West (SW) as our GP surgeries, and those serving our London borough. The recruitment process through secondary care was that Mental Health Research Network (MHRN) staff presented the study to the team, who identified people who met our exclusion/inclusion criteria and approached them with an invitation pack. We were not able to track how many people were excluded by secondary care; we know only how many packs the teams were sent but not how many they handed out. This has impacted on our response rates, as we were unable to accurately assess how many service users were invited to take part. Our data, as presented in Figures 6 and 7, show 154 packs were given to secondary care practitioners to distribute. See Table 2 for summary of final study population by recruitment route.

TABLE 2 - Network-mapping interview recruitment by site (%)

Recruitment route	SW (n = 75)	London (n = 75)	Combined (n = 150)
Primary care	80	82.7	81.3
Secondary care	20	17.3	18.7

Data collection

The majority of interviews were conducted by the two study researchers; however, in London MHRN staff were trained and supervised to carry out 14 interviews and a second SW researcher did 24 interviews. All participants received a £20 high street shopping voucher as a gratuity for their time and contribution.

FIGURE 6.

Consort diagram for London recruitment. MH8, a register of patients with schizophrenia, bipolar affective disorder and other psychoses. a, In larger GP practices not all patients were sampled, we randomly selected 60 people to approach.

FIGURE 7.

Consort diagram for SW recruitment. MH8, a register of patients with schizophrenia, bipolar affective disorder and other psychoses. a, In larger GP practices not all patients were sampled, we randomly selected 60 people to approach.

In-depth interviews

Data analysis

The data analysis process took place in three stages; preliminary thematic analysis; within case, interactions between themes; and full thematic analysis and reflection. Data presented in Chapter 4 were based on the full thematic analysis but answer specific questions from emergent network dimensions prioritised as important from Chapter 3, making use of the within-case maps to support the process.

Stage 1: data immersion and theme production

Broad themes were inductively developed following the principles of working with ‘reflecting teams’, who discuss transcripts line by line, in a group, to consider all the possible layers of meaning. 182 Using this technique for data immersion allowed the academic researchers to work in partnership with the lived experience researchers. Jones states that it was more important that the members of the team work creatively and imaginatively in partnership than that they have any specific knowledge of research methods. 183

The academic researchers (two) and lived experience researchers (two) read through individually, and then as a group, three transcripts and discussed what they thought were the most important things within these transcripts. All members of the reflecting team made detailed notes on the transcripts, creating memos of their thoughts both while reading and as a result of the group discussions. The group discussion included considering whether or not differences in opinion were influenced by differences in experience and/or training. The aim of this part of the discussion was not to resolve who was ‘correct’ in a positivist sense, but to gain an understanding of the others’ perspectives. At the end of this session a list was produced of the emergent areas of what seemed to be of interest in the transcripts from the perspectives of research question 1 and what the research participants chose to talk about to the researchers.

The ‘reflecting team process’, as detailed above, was carried out twice more with one of the academics and one of the lived experience researchers on a further four transcripts. At the end of this process eight broad themes had been agreed on. Three of these were agreed to have emerged as important because interview participants had been prompted to discuss them in the initial network interviews (activities you participate in; strategies for keeping well; people). These broad themes could be considered to be deductively generated. Three broad themes were agreed to have emerged as important because of the temporal dimension that this more reflective interviewing style facilitated (past, future, changes in networks over time). One broad theme was agreed to have been inductively generated by the involvement of the lived experience researchers who identified hope and possibilities in what people had and could achieve, when the academic researchers were hearing predominantly negative narratives (building blocks to doing things for you). An eighth broad theme was introduced to cover information that was felt to be important, but which otherwise might be lost (other). Information about physical health and reflections on taking part in the network-mapping exercise were both covered by this theme.

These themes were applied, using the NVivo 10 data management tool, to all of the transcripts by two researchers who checked each other’s coding and discussed refinements to the theme definitions. The convergences, divergences and newly emergent stories within each theme were examined and discussed first within the analysis team and later with the wider research team.

For each participant a summary paragraph was also constructed blending key features of the network data from Chapter 3 with the in-depth interview material; this included three network words or phrases chosen by participants, demographical information, and diagnosis and key life events information.

Stage 2: within case, interactions between themes, analysis

We were interested in what was important to individuals and their experiences, rather than producing generalisable themes across interviews. Instead of undertaking a cross-case thematic analysis we were interested in what was important about these themes for the participants. Each broad theme, from stage 1, was allocated a colour, and visual within-case analysis was carried out for each of the participants to examine the interactions between themes within their accounts. 184 A landscape-oriented piece of white A4 paper was used for each participant185 (see Figure 8 for an example). On the left-hand side of the page factors that had been coded as being significant in participants’ ‘past’ were drawn, or written. On the right-hand side of the page the ‘future’ was represented. In the middle of the page a figure of a person was drawn and under this was listed information from ‘other’. ‘Activities’, ‘Strategies’, ‘People’, ‘Changing Networks’ and ‘Resilience’ were then all drawn or written around the remaining white space, with interactions between any of the parts of the themes being indicated with arrows.

FIGURE 8.

Example of a within-case analysis picture. Key: green, past; pink, strategies; yellow, future; orange, activities; purple, building blocks; light blue, change in networks; blue circles, people; red circles, other information, usually statements about health or beliefs; red jagged edges, particularly stressful things in life, including voices.

We asked the lived experience researcher who worked closely with the study team on this module to provide feedback on their role. The researcher summed up the involvement experience in three words:

• energising

• confirming (that I do have a valuable contribution to make)

• committing (I was committed to the team and the project in the same way they were committed to supporting me in my work).

I like to think that I contributed my experiences rather than my story of living with a diagnosis of SMI. My experiences, I think, enabled me to analyse the interview transcripts in a way that was reflective of the subjective experience. I think I contributed a dry sense of humour – always helpful when the going gets tough. Being on my own recovery journey looking at the past, present and future in the analysis really fitted well with my view of peoples journeys, of which social networking and relationships are key parts.

Stage 3: axial coding and thematic analysis

The first two stages of analysis facilitated the process of data immersion and allowed the research team to consider what was important across the data set and within individual accounts, with the additional benefit of the perspective of the lived experience researchers. Qualitative components of network studies can be, and are, used to provide a greater depth of understanding of how quantitatively generated categories are experienced by individuals and to explore the heterogeneity within them. 186 A consideration of the individual in-depth interviews alongside the quantitatively derived network types and the factors that had been identified as salient by the literature review allowed individual accounts to be explored for wider meaning. Uncoded and anonymised copies of the transcripts were entered into the NVivo 10 (QSR International, Warrington, UK) data management tool. Abductive reasoning was used to produce 50 deductively (from the literature review and quantitative analysis) and inductively (from the two stages of data immersion described above) derived codes. These axial codes, which included both categories and concepts, were then applied to the transcripts by one researcher and checked, for consistency of coding, by a second. 2 The axial codes were then combined into eight thematic categories by a process of discussion and reflection between the two researchers and validation with other members of the research team; there were different categories for each of the three of the quantitatively derived network types. 187 These nine codes were specifically chosen to explore defining features of these networks in order to answer research question 1. The findings of this third stage of the analysis are reported in Chapter 4 and the implications for deepening our understanding each of the network types are discussed.

Practitioner interviews

An important aspect of the study was engaging with practitioners to explore how personal networks emphasising connections to people, place and activity were viewed both strategically and practically. The original research design was for a survey to be carried out among practitioners but concerns over poor survey response rates and the case study design led us to propose a change in methodology. The telephone interviews specifically asked about personal networks and the CHN approach as well as current concepts such as recovery.

Interview process

To help keep track of the conversation within the interview, mind maps were created to help follow responses, identify links between points and assist with decisions over follow-up probing. An example of a mind map is provided in Figure 9.

FIGURE 9.

Example of researcher’s mind map to help navigate through interview.

Organisation interviews: series 2

Synthesis case studies

Social network analysis studies are increasingly incorporating and combining quantitative and qualitative data in complementary ways. 188 We did not do this formally, but instead created a hybrid approach that used network-mapping visuals as the springboard for detailed exploration of what life looks like now and in the past for someone with SMI managing their health and well-being. Within mixed-methods studies there was a danger that combining data sources could lead to clashes between the underlying assumptions of both methodological approaches. 189 We sought to overcome this using within-case synthesis by carefully integrating qualitative and quantitative data, from a variety of perspectives in a geographically bound context, working together as a study team to check each other and draw out shared meaning. The two case studies were intended to present two contrasting cases. The case studies were quantitatively driven, in that the salient features, such as the predominant characteristics of networks, were quantitatively derived. The qualitative data were used to complement this by demonstrating what these generalisable quantitative trends actually looked like in the context of individuals’ lives. The wider context was then considered and the qualitative data from organisational lead interviews (series 2), for each site, was used to examine the local barriers and facilitators for individuals when trying to develop their personal networks through the context of wider community resources. Finally the case studies consider the implications for how practitioners and organisations can help people living with SMI by using a CHN approach.

Patient and public involvement group lived experience reflexive expert review

The case for PPI in research rests on the claim that research that has involved the public at design stage and through the research cycle is likely to be of higher ethical quality and relevance to the populations that it serves. 190 PPI has further been defined as an ongoing dialogue between informed scientific and lay deliberators, each of whom adds different kinds of value to the public practice of science. 191 More recently, these ethical claims have been challenged with the suggestion that they are premised on a paradox: that once the lay voice becomes scientifically informed the ethical claim is unfounded. 192

The study team decided to approach this dilemma with an open mind. At present, the most compelling evidence for PPI in mental health research has focused on the impact of involvement in design and recruitment to studies, suggesting a correlation between the two. 193 Much less attention has been focused on the impact of PPI on the interpretation and synthesis of data as an interactive learning journey and even less on reflexive analysis around the role of lived experience in shaping findings and conclusions. Consequently, the study team decided that the recovery focus around agency within the study necessitated, on the one hand, a commitment to active involvement in the interpretation and synthesis to remain congruent with an ethics of coproduction, and on the other, reflexive critical analysis to consider the strengths and weaknesses of so doing.

We commissioned an independent lived experience PPI consultant who had not been involved with the study to facilitate and lead an interpretative review of the analysis drawing on the expertise of the existing study PPI group who had been employed on the project in a range of research roles. The reason for choosing an independent consultant was to ensure critical independence of view, blinded from the interests and research investments of the wider research team. Members of the existing PPI group were involved with a wide range of research methods awareness and project knowledge, from very little to having a role as research assistant in the data collection. All five of the PPI group had personal experience of living with schizophrenia or bipolar disorder and wider interests in mental health service development and well-being programmes. The PPI group both collaborated and consulted with the lived experience lead throughout the interpretation process according to capacity and time availability.

Three stages followed:

• The independent PPI consultant used a clearly defined and iterative process adapted from previous life story work around personal and social recovery with people with SMI to facilitate the interpretative agency of all members of the PPI group across the range of recovery perspectives and research awareness. 194,195 This was combined with a ‘critical friendship’ approach to working with the study team in which ‘warts and all’ review was combined with working towards the greater good of research quality. This critical friendship approach to PPI has been shown in another study to improve research quality. 196

• There was no governance requirement to go to an ethics committee for PPI. However, we felt it was important to be mindful of potential current vulnerabilities within the PPI study group. An initial telephone interview covered agreement to use lived experience to interpret the analysis and checked emotional safety around so doing. Appropriate disclosure issues were explored, as anonymous perspectives were still likely to be recognisable to the study team.

• The PPI study group were initially given four questions by the PPI consultant, reflecting on their own recovery journey within their work on the project. Responses from each lived experience expert (LEE) were written up as an individual starting position to frame subsequent interpretation and develop the interview schedule. Reflexivity was built into the interview schedule by asking each LEE to consider how their answers were informed by their lived experience perspectives at each step. The group were asked: What was interesting in the data? What was surprising? What was missing that they might have expected to see? What was salient in relation to their experience? What did they think the conclusions and recommendations and next steps should be?

This proceeded as follows:

• Draft report chapters from August 2013 – network and in-depth interviews – were read. This was followed by a telephone interview with the independent PPI consultant to reflect back views from reading the material. The consultant used the interview schedule to elicit feedback from all five PPI group members.

• Second set of August 2013 draft report chapters were circulated – organisation and practitioner interviews.

• Third telephone interview with research consultant was undertaken to reflect back views.

• Research consultant drafted report based on all the feedback, supported by one member of the group who had also collected primary data to check accuracy of report.

• The draft report was circulated for comment and face-to-face meeting attended by PPI group and principal investigator to discuss emerging findings.

• Final changes were made to the draft report, which features as Appendix 1.

The original aim was to include the reflexive review as a separate chapter but on further reflection and review we considered the greatest strength of its impact on the final synthesis was at a formative rather than summative stage. The reflexive review took place at a point in time when the wider study team were still making sense of the data and was never intended be more than a positioned and situated interpretation to make the PPI collaboration within the broader work of interpretative synthesis visible; there is as yet very little formal process around recording these activities in research studies. We consider recognition of these situated limits to be a strength making the research more reflexive.

Overall synthesis

Four members of the research team, including those with quantitative and qualitative research skills, read across all the results of all of the data sets. After immersing ourselves in the data in this way we discussed what seemed to be important within it for understanding the networks of people with SMI. We inductively brought to the surface several theories about how these different elements might be functioning. These have been named ‘meta inferences’. 197 Potential interpretive inconsistencies were not avoided or ignored; following this direction ‘complementary inferences’ were used to provide insights into different aspects of the same phenomena. Therefore, when combined they provide a more complete meaning. 198 We returned to the wider literature to examine how our thoughts related to the work of others, including overarching sociological explanations, and then re-examined the data with these in mind. We first used a matrix to pull together sources of data from Chapters 3–7 in order to ‘test’ three dimensions of influence in networks: availability of resources; use of agency; and external structures and systems. These three dimensions drew on conceptions of agency23 and considerations of structure, as applied to health. 199 We explored the data for:

• signs of enactment of different types of agency within individuals with SMI and also in practitioners and organisational leaders (the agency of others)

• the availability of resources and their role in effecting change

• the impact of external structures and systems enabling, and also acting as barriers to, network development.

Second, we used the products of this analysis to answer the three study research questions. We looked at how personal connections to people, places and activities were used by people with SMI, and explored practitioner roles within them as a distinct influence. We sought to understand how organisations used network approaches that extended beyond social networks to support people with SMI. Finally we addressed how wider community collaborations were built between organisations, and the partnerships forged to provide interorganisation integrated networks of support for people with SMI. The LEE report (see Appendix 1) was an additional data source at this second stage, providing insight into what the PPI group thought was most important in the data set for answering these research questions.

Our study had been primarily empirically driven, with theoretical contextualisation, and this synthesis process aimed to produce an empirically led provisional framework for understanding people’s lives and contributing the wider recovery approach. We drew on concepts in published literature as and when they deepened our understanding of these findings.

Diverse and active networks

Diverse and active were the largest networks, containing the most people, places and activity connections overall. They describe 36.7% of the study population (55 people). Social ties were characterised by larger numbers of friends and wider connections (acquaintances and colleagues) and a high frequency of social contact. With low social network density (ties were more loosely knit) and a high mean number of different relationships, they were also the type which most closely resembled ‘bridging capital’ in structure: weaker and wider ties to different contacts and groups.

Place connections in this network type were dominated by community-based settings such as education and work, and a higher proportion of places where social interaction was mentioned, as well as place connections which had been built through informal referral, such as recommendations from family and friends. Participants with diverse and active networks spent the majority of their waking time outside the home.

Diverse and active networks also had high levels of meaningful activity including relational activities such as socialising and physical activities such as going for walks or running. On average a fifth of these activities were externally structured (either by organisations or through involvement in formal activity groups such as book or cycling clubs). There was an even split between activities done alone and those done with other people. These networks were also characterised by higher proportions of connections made within the last 5 years. The majority of place and activity connections in this type were less than 5 years old. Participants with diverse and active networks also reported higher satisfaction with their connections.

Figure 10 illustrates a diverse and active network type. Mental and physical health settings were present, but the bulk of the network was taken up by connections to community settings and there were also a wide variety of social ties. The family network was large but did not dominate, smaller nodes indicate that many family members were seen less frequently and the existence of separate components such as the leisure centre, martial arts centre and work indicate a diverse network. In Chapter 4 we draw on a qualitative interview with SUL29, where he is known by a pseudonym, ‘Kevin’.

FIGURE 10.

An example diverse and active network (SUL29). Shape of node: type of node. Circle = person, square = place, diamond = activity. Colour of node = well-being impact. Red = very negative, orange = negative, white = neutral, yellow = positive, green = very positive. Size of node: frequency. Larger nodes = more frequently interacted with social ties/activities engaged in/places connected to. Colour of text: mental health network. Blue text = mental health nodes, black = other. Note: These diagrams have been anonymised, including changing of place names where they were too specific. Where social ties and activities were not attached to places this was because they were not interacted with or done in any specific location. Similarly activities without connections to people were those that were done alone. The participant was not located in these diagrams in order to reduce complexity of diagrams but they were connected to everything within them.

Family and stable networks

Family and stable networks were the most bonded and durable network type and describe 32% (48) of our study population. ‘Bonding capital’ refers to the presence of close ties with similar people and these networks were characterised by larger numbers of family contacts including children, partners and wider family. There were relatively few non-family ties. Stability and durability was reflected in the majority of social contacts having been known for longer than 5 years. These networks also had higher numbers of informal carers and neighbours inside the social network. Network density or the proportion of network contacts who also know each other was highest in family and stable networks.

Place connections in this type were stable, the majority on average being older than 5 years. Mental health settings make up a very low proportion of overall place connections. Participants with this network type spent most of their time at home but tended to live with others; home was rated by most as a positive place. They had similar levels of full-time and part-time employment to diverse and active networks and higher than formal and sparse networks. There were high numbers of individually structured activities (those activities where activity was structured by the individual, such as reading or walking, not part of a formal group but alone or with friends and family) and a low proportion of activities which were unstructured (such as watching TV or listening to radio).

Figure 11 is an example of a family and stable network showing a strong family-focused network where home was the key hub but it was also a social place where social interaction and social activity took place. The participant was managed by primary care for their mental health and rated this setting and their GP as negatively impacting their well-being. They also had negative wider family contacts but on the whole the network was positive and was structured around activities with family.

FIGURE 11.

An example family and stable network (SUSW56). Shape of node: type of node. Circle = person, square = place, diamond = activity. Colour of node = well-being impact. Red = very negative, orange = negative, white = neutral, yellow = positive, green = very positive. Size of node: frequency. Larger nodes = more frequently interacted with social ties/activities engaged in/places connected to. Colour of text: mental health network. Blue text = mental health nodes, black = other. Note: These diagrams have been anonymised, including changing of place names where they were too specific. Where social ties and activities were not attached to places this was because they were not interacted with or done in any specific location. Similarly activities without connections to people were those that were done alone. The participant was not located in these diagrams in order to reduce complexity of diagrams but they were connected to everything within them.

Formal and sparse networks

Formal and sparse networks were the smallest, least diverse or active of all the network types and describe 31.3% (47) of our study population. They were characterised by social networks with fewer family, friend or wider contacts and a larger proportion of the network dominated by formal contacts such as mental health or other health practitioners.

Participants with these networks spent most of their waking time at home and also tended to live alone. Place connections were dominated by high proportions of mental health and physical health settings and lower proportions of community settings. A higher number of place connections had been built through formal referral (from practitioners and organisations) and a lower proportion of places involve social interaction. Participants in this cluster had the highest number of unstructured activities such as watching TV or listening to the radio.

Figure 12 shows a formal and sparse network with fewer overall connections and in which physical or mental health settings were more dominant and most of the individual’s time was spent at home where they live alone. Although two friends visited, most activities were done alone and various local shops represented the only regular and current community setting connection. What connections the participant did have were rated positive for well-being impact.

FIGURE 12.

An example formal and sparse network (SUL13). MH, mental health. Shape of node: type of node. Circle = person, square = place, diamond = activity. Colour of node = well-being impact. Red = very negative, orange = negative, white = neutral, yellow = positive, green = very positive. Size of node: frequency. Larger nodes = more frequently interacted with social ties/activities engaged in/places connected to. Colour of text: mental health network. Blue text = mental health nodes, black = other. Note: These diagrams have been anonymised, including changing of place names where they were too specific. Where social ties and activities were not attached to places this was because they were not interacted with or done in any specific location. Similarly activities without connections to people were those that were done alone. The participant was not located in these diagrams in order to reduce complexity of diagrams but they were connected to everything within them.

Sociodemographics

Over half the total study population were male and over two-thirds were white British (see Table 6). Gender did not differ significantly across network type but ethnicity did differ; fewer participants who were not white British had family and stable networks. Mean age was significantly higher in formal and sparse networks and lowest in diverse and active networks.

Mental and physical health

Table 7 summarises mental and physical health characteristics of respondents and network type. The process of grouping diagnosis is explained in Appendix 8. We also found 88% of respondents (133 people) took medication for their mental health problem, of whom 89.5% took medication orally, 6.8% both orally and via depot injection and 3.8% via depot injection only.

Sixteen respondents (10.7%) had been admitted to hospital for physical health reasons in the last 12 months. Reasons for admission included hernia operations, kidney problems, overdose, gallstone removal, cancer treatment and a shoulder operation; 86.7% recalled receiving a physical health check from their GP or practice nurse and 50.7% were taking prescribed medication for physical health reasons.

Diagnosis was grouped as explained in Chapter 2. Groups differed significantly across network type; significantly fewer participants with psychotic disorder/schizophrenia had family and stable networks while almost half of those with manic episodes/bipolar had these networks. Mental health contact type also differed significantly; more of those participants managed only in primary care had family and stable networks and fewer had formal and sparse networks. Inpatient (but not sectioning) history differed significantly and there were fewer participants who had been a psychiatric inpatient who had family and stable networks.

We note several sociodemographic differences in these mental and physical health variables; we did not know why there were these differences in the study sample and the reasons were likely to be multiple, relating to our study design and features of the delivery of psychiatric services. Gender spread was significantly different by diagnosis; 74.6% of the psychotic disorder/schizophrenia group was male compared with 47.7% of the manic episodes/bipolar and 38.5% of the other psychoses group {χ²[degrees of freedom (df) = 2, n = 150] = 13.35, p < 0.01}. The mean age of those managed in primary care was significantly higher than those managed in secondary mental health services, at 48.8 years compared with 44.8 (t = 2.13, df = 125.51, p = 0.04). Significantly fewer participants who were not white British (23.9%) than white British participants (45.2%) were managed only in primary care [χ²(df = 1, n = 150) = 6.09, p =0.01]. Significantly more participants with schizophrenia/psychotic disorder had been an inpatient: 93.2% compared with 84.6% of those with other psychoses and 69.2% of those with manic episodes/bipolar, [χ²(df = 2, n = 150) = 11.95, p < 0.01].

Time since last psychiatric inpatient admission was used as a proxy for illness stability. The median number of years since last admission was 6 and the mode was 1. About a quarter (27%) of those who had been an inpatient had been admitted in the last 2 years, 47.8% in the last 5 years and 74.8% in the last 13 years. We found no significant difference across network type but participants in touch with secondary services had been an inpatient significantly more recently than those managed in primary care, at 7.2 years compared with 11.7 years (t = 2.81, df = 76.60, p = 0.06).

Accommodation and relationship status

Table 8 summarises participant living arrangements and relationship status. We grouped housing type into three groups in which we inferred decreasing stability from ownership (most stable) to other (least stable). ‘Other’ housing types include residential homes or sheltered housing (8% of respondents), staying in a hostel (0.7%), staying with a friend or family in own room (4.7%), sofa-surfing (1.3%) or other (1.3%). We split living situation by alone or with others. ‘With others’ included living with partner/spouse (28%), parent(s) (8.7%), other family (2.7%), friends (2%), or other (8.7%). Marital status was grouped into married/in a civil partnership or cohabiting (with an intimate partner), or ‘other’. ‘Other’ statuses include separated/divorced (14.7%), single (55.3%) and widowed (0.7%).

We found that age decreased significantly with less stable housing type [F(df = 2, n = 147) = 8.81, p < 0.001] and participants who owned homes were significantly older (mean 51.1 years) while participants in the other housing type were youngest (mean 39.3 years). Diagnosis also significantly differed, with only 6.8% of the psychotic disorder/schizophrenia group owning a home compared with 38.5% of manic episodes/bipolar group and 38.5% of the other psychoses group χ²(df = 4, n = 150) = 18.96, p < 0.001. Participants who lived alone were significantly older (t = 3.63, df = 148, p < 0.01) and had been in touch with services for significantly longer (t = 2.55, df = 148, p = 0.01) than those who did not live alone, while significantly more of the psychotic disorder/schizophrenia group (64.4%) lived alone than the other psychoses (46.2%) or manic episodes/bipolar (38.5%) groups [χ²(df = 4, n = 150) = 8.51, p = 0.01]. Diagnosis also differed significantly by marital status, with more participants with manic episodes/bipolar and fewer of those with psychotic disorder/schizophrenia married/in a civil partnership or cohabiting [χ²(df = 2, n = 150) = 12.10, p <0.01]. The reported presence of an informal carer who provides support and assistance was significantly higher for participants who lived with others [χ²(df = 1, n = 150) = 9.42, p < 0.01], but we found no difference according to marital status; 65.9% of those married/in a civil partnership or cohabiting did not report having an informal carer. Many of our participants viewed partners and other relatives not as ‘carers’ but as ordinary family.

Table 8 shows that significantly more participants who owned their homes had family and stable networks while most participants in ‘other’ housing types had diverse and active networks. Similarly, significantly more participants who lived with others had family and stable networks and those who lived alone tended to have either diverse and active or formal and sparse networks. Significantly more participants who were married/in a civil partnership or cohabiting, and who had an informal carer, had family and stable networks. Very few participants who were married/in a civil partnership or cohabiting had formal and sparse network types.

Education and employment

Table 9 summarises education and employment status of the study participants. Only a small minority of participants did not have a formal education but fewer than 30% had full- or part-time employment. Employment percentages add to more than 100%, as some participants had more than one employment type. Qualifications included National Vocational Qualification (NVQ) level 1 or below (6.7%), General Certificate of Education (GCE)/General Certificate of Secondary Education (GCSE) O-levels or equivalent (24.7%), GCE/GCSE A-levels or equivalent (12%), higher education qualifications below degree level (12.7%) or degree level equivalent or higher (33.3%).

We see from Table 9 that network types differ significantly by both education and employment (full-time employment or long-term sickness) dimensions; both more of those with no educational qualifications and those who were on long-term sickness/disability benefit, had formal and sparse networks while fewer of those who worked full time did not.

What type of individuals were in each network type?

We used multinomial logistic regression to assess which social, health and demographic variables remained significant when controlled for other variables. For categorical response variables, the computed R² is pseudo-R², and thus the total variation was not reported but instead approximations were reported. See Appendix 12 for the full list of variables used in this analysis.

Table 10 shows that, when controlled for other variables, only four items remain significant in the model. Participants on long-term sickness or disability benefit were significantly less likely to have diverse and active network types than the other two network types, but this was variable did not significantly affect differences between formal and sparse and family and stable network types. Participants with no formal education qualifications were significantly more likely to have formal and sparse networks compared with family and stable networks than those with formal qualifications. Participants who were living with others were more likely to have family and stable networks than both diverse and active and formal and sparse network types than those living alone. Although we found diagnostic differences in Table 7, when controlled for other variables they are no longer significant.

Well-being and functioning

The Dartmouth CO-OP charts self-assessed current functioning, health status and quality of life (Table 11). All questions were framed within the last 4 weeks from point of data collection and we used a published coding frame for comparison where lower scores refer to better functioning. 200 The CO-OP scores in our study indicated poorer health and functioning on all of eight measures than for the large community (general population) sample. 200

We saw significant differences across network types in the three asterisked measures in Table 11, where in all cases, those with formal and sparse networks had the worst scores while those with diverse and active networks had highest overall health and least limited social activities; those with family and stable networks had the most social support.

Our SWEMWBS was lower than the general population. We present a multiple regression model on this scale in the last section.

Figure 13 indicates that, while SWEMWBS scores differed significantly across network type, all network types included participants with higher and lower well-being.

FIGURE 13.

Distribution of well-being scores within each network type.

Exploring diagnosis groups in the study population

Table 12 below summarises significant differences by diagnosis, rather than network type. It shows that people with psychotic disorder/schizophrenia tended to be male, with the lowest levels of full time employment, formal education, marriage/cohabiting and home ownership and the highest levels of long-term sickness/disability and living alone.

Our regression model on network type did not find diagnosis to be significant, but we can see from this table a tendency for the bipolar group to have the characteristics that were associated with family and stable or diverse and active networks and for the schizophrenia group to have the variables associated with formal and sparse networks.

Composition of social networks

We mapped all current social ties where there was regular interaction. The mean social network size was 19.9 contacts and they ranged from 5 to 64 contacts. The data were not normally distributed; we found a median size of 18 contacts and multiple modes of 13, 18 and 19 contacts.

We collected data on relationship type and summarised these into four relationship groups as shown in Table 14: family, friends, wider contacts and formal contacts. The largest groups of contacts were family and friends. Six per cent of participants had no family ties, 8% had no friends, 22% had no wider contacts and only 2.7% had no formal contacts.

Participant characteristics, place and activity variables were included in a multiple regression model to examine variance in the natural logarithm (used instead of the original scale because of non-normal distribution) for social network size. Table 15 shows that, when controlled for other variables, social network size increased for every connection to places which were community settings, for externally/formally structured activities and also slightly with increased age and higher SWEMWBS score. It was significantly higher for participants in touch with secondary mental health services than for those managed in primary care, and significantly higher for white British participants than for those of other ethnicities. Finally, social network size decreased significantly for increased waking time spent at home and was significantly smaller for participants with psychotic disorder/schizophrenia than for those with manic episodes/bipolar. Social network size was also significantly higher for participants with family and stable networks, and those with diverse and active networks, than for those with formal and sparse network types. This model explained 62.6% of the variance in the natural logarithm of social network size.

In considering explanations for the observed variations, we were aware that the larger social networks of those in secondary mental health contact may be partly explained by the participants having significantly more practitioner contacts than those in primary care, a mean of 4.4 compared with 2.6 (t = 4.67, df = 148, p < 0.01). The larger networks of white British participants consist of significantly more friends (t = 2.48, df = 124.07, p = 0.02) and family (t = 2.50, df = 148, p = 0.01) than those of other ethnicities. The smaller networks of those with psychotic disorder/schizophrenia consist of significantly fewer friends [F(df = 2, n = 147) = 5.02, p < 0.01] and family [F(df = 2, n = 147) = 11.56, p < 0.01] but more practitioners [F(df = 2, n = 147) = 5.78, p < 0.01] than the other diagnostic groups.

We also considered mental health-related contacts within networks and these are summarised in Table 16. Participants who had a significantly higher percentage of networks comprising other mental health service users were those who had been in contact with mental health services for longer (Spearman’s r = 0.25, p = 0.02) and those who were on long-term sickness or disability benefit (t = 2.15, df = 148, p =0.03).

Participants who had significantly higher proportions of ties who knew about their mental health condition were those who were supported in secondary mental health services (t = 2.13, df = 148, p = 0.04), those who had been a psychiatric inpatient (t = 2.57, df = 148, p = 0.01) and those who were on long-term sickness or disability benefit (t = 2.92, df = 148, p < 0.01).

Table 16 indicates how disclosure of mental health condition differs by social network group. Mental health condition was often kept from wider components of networks such as colleagues and acquaintances, as illustrated in Figure 14, which splits mental health disclosure by those who (in the participant’s perception) know (blue) about their mental health condition and those who do not (red).

FIGURE 14.

Perceived knowledge of mental health condition among network ties. Key: blue = knows about condition; red = does not know.

Social network ‘emotional closeness’ and well-being impact

Participants rated each social network tie by how emotionally close they felt to them through six levels using concentric circles, from inner circle (closest) to outer circle (not close at all) (see Appendix 7). Most participants (86%) had someone in their inner circle and on average 4.5 ties in the inner circle. These were primarily family (59.1%) or friends (27.6%) but practitioners (8.4%) and wider contacts (4.9%) also featured. The number of social ties within the inner circle decreased significantly with years since first mental health contact (r = –0.23 p = 0.02) and the mean number was highest for those with family and stable networks and lowest for those with formal and sparse networks [F(df = 2, n = 147)= 6.63, p < 0.01].

Participants without anyone this emotionally close had significantly fewer family contacts, with a mean of 3.1 family ties compared with 6.7 family ties (t = 3.45, df = 35.82, p < 0.01). Participants also rated each social network tie in terms of perceived impact on their own well-being, on a five-point scale from very negative (–2) to very positive (2). The majority of contacts were rated quite (37.6%) or very (26.8%) positive while only one in ten was rated as either quite (7%) or very (3.1%) negative. The proportion of negative ties did not vary significantly across network types.

Negative contacts were significantly correlated with network size (r = 0.31, p < 0.01); larger networks were more likely to contain negative contacts. The most commonly rated negative ties were colleagues, 19.2% of which were rated negative; neighbours, 18.7%; wider family, 14.3%; immediate family, 11.9%; acquaintances, 11.5%; mental health practitioners, 10.1%; non-mental health practitioners, 8.7%; partners, 6.5%; and friends 6%. Negative ties tended to be found in relationships that people had limited control over, such as colleagues, and least in those where individuals can exert more choice, such as friends and partners.

Frequency of contact with social ties (social network activity)

We measured frequency of interaction with each social tie because it was possible that participants had larger social networks of contacts but full of ties they did not regularly interact with because of illness, distance, time or other factors. ‘Interaction’ included face-to-face, phone and e-mail contact. Overall, 15.9% of contacts were interacted with on a daily basis, 32.8% at least once a week, 25.6% at least once a month and 25.3% only occasionally.

Some 23.3% of participants reported no daily contact with any individual social network tie but all participants had at least one social network tie they contacted at least once a week. We found that those participants who had no daily social network contact had significantly fewer family contacts (t = 4.79, df = 90.51, p < 0.01) and were significantly more likely to live alone [χ²(df = 1, n = 150) = 31.34, p < 0.01] and significantly less likely to be married or cohabiting [χ²(df = 1, n = 150) = 15.44, p < 0.01], and significantly more participants with schizophrenia or psychotic disorder had no daily contact with any social network tie [χ²(df = 2, n = 150) = 6.95, p = 0.03].

A ‘social network activity’ variable was created by computing mean frequency of contact with social ties and we found it decreased significantly with age (r = –0.30, p < 0.01) and was significantly lower for participants in touch with secondary mental health services (t = 2.45, df = 148, p = 0.02). It did not differ significantly across network type clusters.

Practitioner role in networks

Formal practitioner contacts comprised a mean 18.6% of total social network size. This differed significantly across network type [F(df = 2, n = 147) = 32.33, p < 0.01]; in formal and sparse networks, practitioner contacts comprised 33.0% of social networks on average compared with 19.6% of those with diverse and active networks and 12.1% of those with family and stable networks. Participants with smaller overall social networks had significantly higher proportions of practitioner ties in them (r = –0.40, p ≤ 0.01). There was also a significant difference across diagnostic groups [F(df = 2, n = 147) = 16.31, p < 0.01]. In the psychotic disorder/schizophrenia group, a mean of 28.8% of networks were composed of formal contacts, compared with 14.4% of the networks of those in the manic episodes/bipolar group and 21.82% of the social networks of those with other psychoses.

We examined three groups of practitioners: clinical mental health practitioners [psychiatrists, community psychiatric nurses (CPNs)], other mental health practitioners (third sector), and other health and social care practitioners (GPs, dentists, job advisors). While all participants in the study had a GP surgery, 70.7% had a named GP and this did not differ significantly by primary/secondary care contact or diagnosis. A total of 27.3% of participants named a dentist. A total of 22.7% of participants had at least one practitioner in their inner circle of emotional closeness: 8% had a GP, 14% had a clinical mental health professional, 3.3% had other mental health practitioners and 2.7% had other health and social care practitioners. Some individuals had more than one; for example, one participant had four clinical mental health practitioners in their inner circle. We did not find significantly more people in primary care with a GP in their inner circle.

There were no significant differences in overall network size between those who had a practitioner in their inner circle or not, but those who had a GP in their inner circle were significantly less likely to be married/in a civil partnership or cohabiting [χ²(df = 1, n = 150) = 5.41, p = 0.02]. It was possible that practitioners replaced close relationships in some networks but on the whole they seemed to complement rather than replace informal networks.

In which closeness circle, out of six, a practitioner first appeared was examined. We found that on average the first mentioned practitioner was closest in diverse and active networks and least close in family and stable networks [F(df = 2, n = 147) = 4.10, p = 0.02]. Participants who felt closer to their clinical mental health practitioners rated them as significantly better for their well-being (r = 0.34, p < 0.01).

We examined continuity of practitioner relationship. Of those participants who named GPs, 5.6% had known their current GP for less than 6 months, 48.1% for 6 months to 5 years, 20.8% for 5–10 years and 20.8% for over 10 years. Of those who named clinical mental health practitioners, 14.8% had known them for less than 6 months, 56.8% for 6 months to 5 years, 17% for 5–10 years and 11.4% for over 10 years. Participants who had known their GP for longer felt significantly closer to them (r = 0.23, p < 0.02), and rated them as having a significantly better impact on their well-being (r = 0.26, p < 0.01). Participants who had known their clinical mental health practitioners for longer felt significantly closer to them (r = 0.32, p < 0.01) and had significantly higher satisfaction with their network as a whole (r = 0.27, p = 0.01).

Social network structure: density and diversity

The structure of social networks was also examined. Participants were asked which members of their social network knew one another, to measure density; higher density refers to a larger proportion of social network members knowing each other. We considered four measures of density, all of which can be skewed by social network size, because the probability of a fully connected network decreases as network size increases. 180 Efficiency had the lowest correlation with social network size (r = 0.17, p = 0.04) and so this was selected. Higher efficiency scores equate to lower density or less connected networks. Larger numbers of wider contacts (r = 0.27, p < 0.01) and friend (r = 0.26, p < 0.01) and practitioner contacts (r = 0.24, p < 0.01) were correlated with higher efficiency but larger family networks were negatively correlated and thus more dense (r = –0.35, p < 0.01).

Density differed significantly across network types [F(df = 2, n = 147) = 17.60, p < 0.01]; family and stable networks had the lowest efficiency scores and diverse and active networks had the highest, reflecting the interconnected bonded structure of the former and the diverse settings found in the latter.

Diversity of network was considered using mean number of relationship types. On average participants had 5.5 out of a possible 9 relationship types mapped in their social networks, ranging from 3 to 9 (SD 1.5). Those with diverse and active networks had significantly more relationship types than the other networks and those with formal and sparse networks had the fewest [F(df = 2, n = 147) = 18.67, p < 0.01].

Home

Home was mapped for all participants. We considered waking time spent at home, well-being impact of home and social network ties linked to home.

FIGURE 15.

The dominance of home place in networks (n = 150).

Over half of participants spent more than 50% of their waking time at home (Figure 15). This differed significantly across our three network types [χ²(df = 4, n = 150) = 17.55, p < 0.01]. Those with family and stable networks and those with formal and sparse networks both spent most of their waking time at home, but we saw from Table 8 that the latter group tended to live alone and the former with others.

Those who spent more of their time at home had significantly smaller social networks (r = –0.46, p < 0.01) including smaller friend groups (r = –0.40, p < 0.01) and wider contact groups (r = –0.34, p < 0.01). Those who spent more time at home were also significantly older (r = 0.23, p < 0.01) and had been in touch with mental health services significantly longer (r = 0.15, p < 0.05).

A total of 86.1% of participants mentioned some form of regular social interaction at home, such as regular visits from family, friends or practitioners. Participants rated home as positive for well-being on the whole. Some 74.2% rated home as positive (32%) or very positive (42.2%), 15% rated it as neutral and 10.9% rated it as negative (7.5%) or very negative (3.4%). We found significantly more participants with formal and sparse networks rated their home as positive than those with other network types [χ²(df = 2, n = 147) = 8.68, p = 0.01], and significantly more of those renting rated their home as neutral or negative than those with other housing status [χ²(df = 2, n = 147) = 7.59, p = 0.02], while significantly more participants who were not white British than white British participants also rated their home as neutral or negatively impacting well-being [χ²(df = 1, n = 147) = 6.04, p = 0.01]. A total of 41.6% of participants had lived in their current home for more than 10 years, 41.6% for 5–10 years, 17.4% for 6 months to 5 years and 6% for less than 6 months. We did not find any network or well-being differences according to time lived in current home.

Places outside the home

A total of 1449 places outside the home that people currently and regularly attended were mapped in the study. At the analysis stage, places were categorised (Table 17, or see Appendix 10 for more details). All participants were connected to a GP surgery, as this was a recruitment criterion.

We found that, on the whole, places were rated as overwhelmingly positive for well-being and most network maps had only place connections they viewed positively, or which were viewed neutrally but served a function, for example shops or a GP surgery. Our analysis of place focused on connections across settings, associations with social networks and sociodemographics. Later we look at the relationship to resource access and well-being.

Across our network types we found significant differences in connection to total places outside the home [F(df = 2, n = 147) = 17.42, p < 0.01]; those with diverse and active networks connected to the most and those with formal and sparse networks the fewest. Mental health settings were most numerous in formal and sparse networks and least common in family and stable networks [F(df = 2, n = 147) = 2.12, p < 0.05]. Community settings were significantly less common in formal and sparse networks while similar numbers were found in diverse and active networks (mean 6.9) and family and stable networks (mean 7.0) [F(df = 2,n = 147) = 23.28, p < 0.01]. Those with formal and sparse networks connected to significantly fewer physical recreation settings and those with diverse and active networks connected to the most [F(df = 2, n = 147) = 5.35, p = 0.01].

Finally, some connection-building was evidenced by those who linked to more workplaces (r = 0.20, p = 0.01) and physical recreation settings (r = 0.20, p = 0.02) having significantly more social networks of contacts built within the last 5 years.

Duration of place contact

As with social networks, we also examined duration of contact to place connections and the balance between network-building and network stability. Overall, the majority of connections in participants place networks were over 5 years old (52.6%) while 47.4% had been built in the last 5 years. We found older participants had significantly more connections over 5 years old (r = 0.36, p < 0.01) and found evidence of network rupture caused by inpatient admission; those who had been inpatients in the last 5 years had significantly fewer place connections older than 5 years than those who had not been inpatients in the last 5 years (t = 3.00, df = 99.93, p < 0.01). Participants who reported decreasing health had significantly fewer place connections built within the last 5 years (r = 0.20, p = 0.01).

We also found participants who lived alone had significantly fewer new place connections built in the last 5 years (t = 2.09, df = 148, p = 0.04) and those who worked full time had significantly more new place connections (t = 4.06, df = 148, p < 0.01).

Referral to place: how were place connections built?

We examined how certain places were referred to (Table 18). Referral by others was unusual and participants tended to make their own connections; however, participants with larger friend networks (r = 0.21, p = 0.01) and larger diverse networks (r = 0.22, p < 0.01) were informally referred to a higher percentage of places; similarly those with larger practitioner networks had more formal recommendations (r = 0.30, p < 0.01). Participants in secondary mental health care were signposted to significantly more places than those in primary care (t = 8.26, df = 147.78, p < 0.01). We found informal referral to differ significantly by network type [F(df = 2, n = 147) = 4.56, p = 0.01], with the highest proportion of place connections built through informal referral in diverse and active networks and the lowest in formal and sparse networks.

Social interaction in place

We measured the extent of social interaction in places connected to in three categories: places where social interaction was mentioned by the participant, places were no social interaction was mentioned but was inferred as necessary (e.g. in dentist surgeries or mental health appointments social interaction was integral to the visit) and places where social interaction was not mentioned or inferred as necessary (e.g. supermarkets, libraries, parks). Overall, social interaction was mentioned in 68.9% of places, not mentioned but inferred in 12.9% and neither mentioned nor inferred in 18% of places.

The majority of places involved social interaction mentioned by the participant. We found that the number of place connections where social interaction was mentioned was correlated with more friend contacts (r = 0.49, p < 0.01), family contacts (r = 0.42, p < 0.01), practitioner contacts (r = 0.21, p < 0.01) and wider contacts (r = 0.40, p < 0.01). The number of places where social interaction was not mentioned or inferred was correlated with smaller family networks (r = –0.27, p ≤ 0.01).

Where social interaction was not mentioned but inferred, the participant did not mention it because there was no specific social tie whom they saw regularly there; for instance, at the dentist they might have seen whoever was available rather than a specific practitioner.

Social interaction was therefore a key dimension of place. However, the dominance of shops and markets in the category of no social places mentioned or inferred may skew this finding. We found that both green and outdoor spaces and spiritual settings where no social interaction was mentioned or inferred were still rated positively for well-being, with a mean positive score and a modal very positive score for both; places themselves also had a beneficial impact.

What have we learnt about place connections?

The place connection data shows us how diverse people’s lives were; people with SMI were not living in a psychiatric ghetto. These networks of place seem to be built by the individual rather than those around them. The 150 participants used 1449 places away from home that were important enough to mention at interviews for impacting on their health and well-being, mostly positively. We did not find place, as such, an important feature of network in terms of variations across the study population and impact on well-being. However, it was clear that the type of setting connected to did matter as well as social interaction in a place.

What activities were the study population doing?

A total of 958 activities were mapped and 14 categories applied at the analysis stage (Table 19). We found a wide range of meaningful activities engaged in, from participants who spent all day watching TV to those who engaged in 16 diverse activities. Activities were extremely personalised and challenging to categorise. We found most participants (76%) engaged in some physical activity even if it was only going for walks. General recreation such as watching TV, reading, creative and cultural hobbies as well as multimedia use were also common. Everyday tasks such as domestic activities also featured in many networks.

Participants tended to rate their activities as having a positive impact on well-being, as they also did with places. We found that overall the number of meaningful activities engaged in decreased with age (r = –0.17, p = 0.03) and that those who engaged in more activities also had more friends (r = 0.41, p < 0.01), family (r = 0.18, p = 0.04) and wider contacts (r = 0.19, p = 0.02).

Across network types, participants with formal and sparse networks engaged in significantly fewer physical activities and those with diverse and active networks engaged in the most (F = 3.59, df = 147, p = 0.03).

Level of structure in activities

In order to understand activities further, we labelled them according to structure level. Unstructured activities were those that passed time without an obvious purpose, such as watching TV or listening to the radio. Individually structured activities were those where the structure was set by the participant and had an obvious purpose such as exercise, study or socialising and externally or formally structured activities were those where the structure was set by others, such as formal education, work, volunteering, exercise with a personal trainer and so on. We found 13% of activities had no structure, 71.2% were individually structured and 15.8% were externally or formally structured.

We found that larger social network size was significantly correlated with individually (r = 0.27, p < 0.01) and externally or formally (r = 0.57, p < 0.01) structured activities.

Social and non-social activity

The majority of activities were those that the participant tended to do alone (59.1%) while 40.9% were activities that the participant usually did with others.

A total of 46.7% of activities were done alone and yet were structured. We found that those who engaged in more of these ‘alone but structured’ activities had larger friend groups and connected to more places (r = 0.25, p = 0.02) including libraries, cultural recreation, spiritual settings and physical recreation settings but did not have significantly different well-being or functioning. Unstructured activities done alone were associated with smaller family networks (r = –0.18, p = 0.02) and lower perceived social support (r = 0.22, p < 0.01) and physical fitness (r = –0.18, p = 0.02). Those who engaged in more unstructured activities alone were significantly older (r = 0.16, p = 0.05) and had been in touch with mental health services for longer (r = 0.19, p = 0.02).

Activity duration

The majority (69.1%) of activities mentioned were those that participants had done for longer than 5 years, with fewer than a third (30.8%) of activities having been started within the last 5 years. We found that older participants had a significantly lower proportion of activities started in the last 5 years (r = –0.32, p < 0.01) and the percentage of newer activities to be significantly correlated to newer social networks (r = 0.26, p < 0.01) and newer place connections (r = 0.43, p < 0.01) in terms of the proportion of each that were less than 5 years old. Inpatient admission also seemed to impact on activity duration, with participants who had been an inpatient in the last 5 years having a significantly lower percentage of activities and hobbies that were older than 5 years (t = 2.36, df = 113, p = 0.02).

Summary: what we have learnt about meaningful activities

Activities mentioned as important tended to be those which were structured and which were done with other people. Those done alone or unstructured seem to be least beneficial for networks overall and we saw significant differences across network types.

Access to social capital

A key interest of the study was in how connections to people, place and activities provide access to resources. The RGUK measures access to social capital resources through personal contacts. Table 22 compares our study findings with other projects that had used the RGUK. The total possible score for the RGUK was 27, with four sub-domains: seven for domestic resources, nine for expert advice, six for personal skills and five for problem solving. It seems to suggest that access to social capital was lower for our respondents than for the general population, but it was higher than other studies involving people with SMI or depression. This could potentially be explained by our sampling locations, as all the other studies except one203 collected data in London, where people with SMI appear to have less access to social capital (as in our study), or other differences in the study populations. However, Table 21 shows the score was lower for those with formal and sparse networks than in any of the studies cited, at a mean of 9.5. Moreover, RGUK score differed significantly across diagnostic group in our study (F = 9.90, df = 158, p < 0.001); the psychotic disorder/schizophrenia group had a mean score of only 11.6, compared with 16.6 for those in the bipolar disorder group and 14.0 for those with other psychoses.

We examined variance in RGUK using a multiple regression model including all network and participant characteristic variables as well as network type. The variables used are presented in Appendix 12. Table 23 summarises the significant variables and shows that, when controlled for other items, RGUK scores were higher for participants who had someone in their inner circle compared with those who did not, for participants who were working full time compared with those not, for participants who had higher percentages of social ties rated as negatively impacting their well-being, for increased connection to places where social interaction was mentioned and for increases in social ties who knew about the participant’s mental health condition. It was lower for participants with higher efficiency scores (less dense networks had lower social capital), for participants who did not have any formal education qualifications, for participants in London compared with the SW site, for participants who spent more time at home and for lower physical fitness scores. Of the total variation in RGUK score, 60.7% was explained by these variables. Aspects of social ties, place connections and activities all remained significant in the final model. When examining RGUK scores across network types (see Figure 16) we found variations: people with formal and sparse networks tended to have lower scores.

Figure 16 indicates how RGUK score varies within each network type and highlights the segregation of scores between network types. Few participants with diverse and active or family and stable networks had the lowest RGUK scores, while participants with formal and sparse networks tended on the whole to have low scores on this scale and very few were around or higher than the mean score of 14.19.

FIGURE 16.

Distribution of RGUK score by network type.

Sources of social capital

Figure 17 shows that social capital was accessed primarily through immediate family and friends, with limited resource access from practitioners. Table 24 indicates the groups for which practitioner sources of social capital were significantly higher. The percentage of resources provided by practitioners decreased significantly with time since last inpatient admission (r = –0.19, p = 0.04) and was higher for those with smaller friend (r = –0.22, p < 0.01) and family (r = –0.21, p = 0.01) networks and larger practitioner networks (r = 0.42, p < 0.01).

FIGURE 17.

Sources of social capital: % of total social capital resources accessed by relationship type.

Access to health resources

To supplement information from the RGUK looking at social capital, we created a bespoke schedule to assess access to health resources among the study population. Table 25 summarises significant independent variables explaining variance in HRG group and shows that, with increasing age and increased percentage of place connections made in the last 5 years, participants were less likely to be in the high HRG score group. Participants who were volunteering, those on long-term sickness or disability benefit and those living alone were also less likely to be in the high HRG score group, while participants with more family contacts and who engaged in more social activities were more likely to be in the high HRG score group. Sixty-one per cent of the total variance in HRG score group was explained by these variables.

Sources of health resources

As Figure 18 shows, health resources for those who had them were overwhelmingly provided by family and friends, although practitioner provision was higher than for RGUK. Acquaintances provided very few of these resources. Table 26 summarises for whom practitioner provision of HRG was significantly higher and it is notable that many of those in the HRG high group had formal and sparse networks.

FIGURE 18.

Sources of health resources: percentage of total health resources accessed by relationship type.

Networks and well-being

Only 34.1% of variance in SWEMWBS score was explained by the variables in Table 27, which indicated that SWEMWBS scores were higher for participants who had higher access to social support, were married/in a civil partnership or cohabiting and engaged in volunteer work and for males than females. It was lower for participants who had children, for increased percentage of social ties rated negative for well-being, for participants who had never been in a psychiatric inpatient ward, for those with an informal carer and for increased number of place connections in mental health settings.

What were the differences between primary and secondary care contact?

Because our study included both participants managed only in primary care (38.7%) and those managed in secondary care (61.3%), we considered significant differences between these populations.

Table 28 shows that, when other variables (sociodemographic, health and well-being, employment, housing and network) were held constant, participants who had a higher percentage of social ties known for less than 5 years, who had higher mean numbers of relationship types in the social network, who had a practitioner in their inner circle of emotional closeness, who spent more of their waking time at home, who engaged in more unstructured activities and who had a higher percentage of place connections which were mental health settings were more likely to be in secondary mental health service contact rather than managed in primary care. Meanwhile, participants with no formal education qualifications, who reported more limited social activities because of physical or emotional health, who had a higher mean frequency of contact with their social ties and who engaged in more activities alone were more likely to be managed in primary care.

Data limitations

We acknowledge that our network-mapping data were limited by potential sampling bias due to low response rates, 15.01% in London and 23.65% in the SW, and when reflecting on findings this must be taken into account.

What characterises the personal networks of people with severe mental illness?

We found great heterogeneity in the connections to people, place and activities. We see diversity in some of our participants’ lives that rebukes assumptions that the majority of this population are inactive and living in a mental health ghetto of day centres and other institutional spaces. 205 Our clustering analysis has grouped the individuals into three broad network types with distinct sets of structures and network characteristics. While the clustering is an artificial construction it is useful for showing the diversity of individuals’ people–place–activity networks. Other studies have used similar methods to identify network types, albeit on social ties only. 206–208 For example, a study on the mental health of older adults also identified that less resourceful network types, lacking friends or family, were related to poorer mental health. 179 Another study developed network types and found a more diverse and loose-knit network was more active and with better access to resources than smaller, more excluded or homogeneous networks. 46

Our network types should be interpreted in the context of two study locations; in another SMI population, network types might cluster in a different way. Moreover, while we had a ‘diverse and active’ group, the mean social network size (19.9 social contacts) was less than in studies of the ‘general population’. It was difficult to compare with other studies as methods of network size estimation varied, and often studies were interested in only one type of social tie. However, population estimates suggest, for example, 73 active ties based on listing known contacts using open questions209 and 123 active contacts based on the exchange of Christmas cards. 210 Studies that directly compare mental illness and general population networks also show smaller networks for people with mental illness. 45,109 Therefore, compared with the general population, the connections within this group were likely to be less ‘diverse and active’.

The inclusion of place and activity variables adds more detailed understanding to these network types. We can see the differing balance of formal and informal settings, and of community and health settings, in individuals’ lives. Home has emerged as an important place in networks. The extent to which people spent their time at home was significantly negatively associated with social network size and access to social capital, suggesting value in community engagement. Both participants with formal and sparse networks and those with family and stable networks spent most of their time at home, but those with the latter networks tended to live with others and also had higher access to social capital. Having more place connections outside the home, particularly those in the community, was also associated with larger social networks, and those places which were social were associated with higher social capital. Working full time was positively associated with social capital scores. This was consistent with other work, showing the importance of employment for resource generation – financial resources but also social relationships. Increased engagement in social activities was associated with having higher access to health and well-being resources.

Our participants accessed a wide variety of settings in the community. Over a third of the study population used libraries, which was more than the general population, estimated in 2009 as 12.8% of the adult population visiting once a month or more. 211 Place connections which were safe and free to access, such as parks, provided important community spaces to get out of the home, meet other people and build identities that were not defined by mental health. 212,213 Our data suggested that the community was an important resource, and fuller networks were those that connected to and made active use of community resources to enhance well-being, while people who were the most isolated were those that spent a lot of time at home doing unstructured activities, living alone and with few visitors. This may seem unsurprising, but connectedness as a dimension of recovery8 is not given the same priority in mental health services as risk assessment or medication management for SMI; social aspects of living with SMI are given minimal attention in NICE treatment guidelines. 38,39,214 Deinstitutionalisation and community connection also depends on the existence and availability of community resources and spaces outside the home to connect to, and these places being truly in the community rather than institutionalised settings,205 as well as reducing barriers to access such as stigma. 215

The importance of meaningful activity for our respondents supported findings in other studies. 127 Types of activities which were meaningful might vary from person to person216 but feeling productive, having routine and being socially engaged were key aspects of the beneficial impact of meaningful activities. 131

Our three network types might provide a useful lens through which to distil myriad network variables into types that mental health practitioners can work with more easily. It provides detail on the extent and type of connectedness that a client has, which provides a context for thinking about interventions, as well as defining some of the resources available. Our resource and well-being data indicated that both diverse and active networks and family and stable networks had healthy components to them. Although we see a tendency for people with schizophrenia/psychosis to live alone and have fewer family ties or resources, our modelling of who had each network type showed that diagnosis was not a significant explanatory variable: many participants with psychotic disorder/schizophrenia develop diverse and active networks (42% in this study). Long-term sickness and disability as an occupational status seemed to be the key variable distinguishing diverse and active from formal and sparse networks; participants who reported being long-term sick tended to have formal and sparse network types. Long-term sickness as a risk factor for poorer social connection, negative social identity and lower levels of physical and other activity was well supported in other work. 100 In these data, living alone was a key distinguishing factor determining membership of formal and sparse networks compared with family and stable networks. Living alone was likely to be a risk factor for isolation and loneliness – which themselves have well-established negative health implications200 – especially where connection to community settings outside the home was lacking. These findings are cross-sectional and exploratory; moreover, causation may be two-way: diagnosis may affect if someone is living alone or long-term sick. However, our findings indicated that it may be useful to view a person through qualities of their network in addition to the label of a diagnosis, in order to understand a person and the social support that they require. This was the approach also being advocated in the treatment of long-term conditions, recognising the central role of social support in recovery. 102,103

Employment was an indicator of access to resources, as shown in the RGUK regression model. Employment can be beneficial because it facilitates access to social interaction and provides structure, a sense of identity, self-esteem and improved finances,132 while gaining employment can improve mental well-being and social inclusion,133 but people with mental illness face discriminatory barriers in access to employment. 217 In addition, not all employment is equal, and low-paid work which lacks meaning for the individual may be no better for health and well-being than long-term sickness. 218

The networks we mapped contained contacts, places and activities that were mostly viewed as positive or neutral for well-being. The study team reflected on this and it was likely that, as our focus was health and networks to support health and well-being, adverse or difficult relationships were less readily brought to the fore through the interview questions. Despite efforts to encourage people to relate to positive and negative contacts, places and activities, we acknowledge that we might not have been fully successful in getting people to relay negative aspects of networks. Our finding may reflect the stability of our study population at time of interview, and represent individual control and agency over personal networks where individuals had pruned their networks to remove negative ties. Those with smaller networks may value any tie even where conflict or negativity exists. 96 Moreover, the existence of some negative ties is not in itself a problem if people can manage them.

Another observation was over disclosure of mental health within network settings; a recent review highlighted complexity in disclosure decisions. 219 Fewer wider contacts, including colleagues, had been disclosed to and those with formal and sparse networks had the highest level of disclosure. It may be that the ability to control this information was an important aspect of network management. Participants with formal and sparse networks had a larger proportion of practitioner contacts who knew about their mental health condition, while those who had diverse and active networks, and had greater choice over whether or not to disclose, seemed to tend not to do so.

While it was difficult to compare methods, the mean size of the social networks we mapped was larger than in many studies, including the mean size of 15.2 contacts found in the recent systematic review of individuals with psychosis. 220 This could be for a number of reasons. Our method was different: it was expansive in terms of network boundaries and included places and activities as extra prompts for social contact so we may have mapped a larger network through methods used. However, a study on people who had been in a mental institution found a larger social network size of 23. 82 Our data also suggest it was the type of relationship (close relationships) rather than the number which was most important for resource exchange and this was supported by other work. 221

Finally, the wide range of places and activities important to people in our study population underlines the personal nature of recovery from mental illness and progression towards greater citizenship;69,222 what was important varied by person and one size does not fit all. While the specific activity may differ, the literature suggests having meaningful activities in one’s life can increase life satisfaction and improve subjective quality of life in this population. 127

What was the mix of formal and informal health-care provision within people’s networks?

We considered access to social capital and to health resources to answer this question. We found these resources to be provided primarily by close informal ties such as family and friends and less so from wider and practitioner contacts. Importantly, those who lacked family and friends to provide these resources had lower resources; they did not seem to replace these resources with practitioner contacts. More health resources (HRG) were provided by practitioners than social capital resources, but their support was still a minority of overall HRG.

Elements of place, social relationships and activity all seemed to impact access to social capital. Close social ties within the inner circle seem to be particularly important. Social network contacts who know about the participant’s mental health impact positively; this was controlled for type of relationship and suggests that, regardless of relationship type, disclosure may foster the trust and reciprocity required for social capital to grow. Employment also improved access to social capital, as was found in a study of South East Asian women with SMI. 122 The finding that increased percentage of network viewed as negative contributes positively to social capital was of interest. Contacts can be disliked but still useful; we found colleagues were viewed primarily as negative, so this fits with the impact of employment. Social capital was a resource accessed through social networks so it was unsurprising that place connections which were social had an impact. Participants in the SW had higher social capital and this will be explored in the case studies (see Chapter 7).

Did these networks differ in composition, range, size and density across subgroups?

Controlled for participant characteristics, only four factors were significantly associated with network type: housing status, long-term sickness and disability, living alone or not, and formal qualifications. This was important because it suggests that, on the whole, anyone can develop a more resourceful network type regardless of their diagnosis, age, ethnicity or gender, even if influential factors might be hard to change. We find support for this idea in that formally structured activities, time spent at home and community setting connections explain variance in social network size. Longitudinal work would be required to unpick how networks develop over time. However, our work highlights the heterogeneity and complexity of networks and suggests one size does not fit all; is it also likely that mean scores will apply to a particular individual. Therefore, it may be helpful not to attempt to isolate such factors but instead to use the network approach to map networks on an individual basis, supporting a person-centred approach to interventions.

How did network strength and breadth relate to well-being?

Well-being was our only outcome measure which was not inherently social. We found activities such as engaging in voluntary work, higher social support, being married/in a civil partnership or cohabiting and being male compared with female to be significantly associated with higher well-being scores, and increased connection to mental health settings, having children, a higher proportion of negative ties, lack of psychiatric inpatient history and presence of an informal carer to be significantly associated with lower well-being. It was difficult to explain these results using our network-mapping data set without longitudinal data. We could only make suggestions such as research showing hope was a strong predictor of subjective well-being,223 which possibly explains the importance of volunteering and cohabiting relationships. We found well-being and all functioning measures including perceived physical fitness, quality of life and overall health to be lower than the general population across all three network types. Those with formal and sparse networks did the fewest physical activities and connected to the fewest physical recreation settings. This suggests that more attention was required on physical fitness and well-being for people with SMI and in particular those who were least connected through people, place or activity resources. While many networks had important connections missing from them, participants valued what connections they did have; most connections to people, places and activities were rated as positive for well-being.

Our well-being model was weak. SWEMWBS measures well-being over the previous 2 weeks,160 though subjective well-being is influenced by both short-term and long-term factors. Many of our network factors and participant characteristics were more longstanding, thus day-to-day life events could affect the score more strongly than network features. Finally we note that there was consistent perceived value in network dimensions across our data, with the words used to describe networks mostly positive descriptions. Personal networks to support well-being contain support, friendship and emotional closeness. This suggests that personal network assessment is a good way of examining strengths and assets on which to build individual plans. For the minority who saw their networks as negative and for those with negative connections, there are additional pointers for change.

What was the role of practitioners in networks?

A key answer to this question lay in evidence that improving networks can improve mental and physical health, and well-being, and that network interventions are possible. More evidence is needed, building on existing work describing effective interventions in various health settings and in mental health224,225 and the barriers practitioners face in addressing social factors in the lives of their clients. 226 The network mapping offers limited insights but the approach has potential. The process can offer a route into understanding the reasons why some networks are sparse, and into working together with individuals to identify potential routes into building connections and resources. The practitioner may have a role in strengthening both bridging and bonding capital. 117

The study looked in particular at the mix of formal and informal ties in networks. The majority of networks were not dominated by mental health: places, people or activities. Only one in four contacts was a practitioner or other mental health service user. Only two out of three participants identified a mental health place setting on their network map. In terms of social capital and health capital, most was gained from non-formal contacts such as immediate family or friends.

Mental health crises will disrupt or rupture networks, affecting activity patterns and connection to place, whether through social withdrawal, stigma227 or loss of employment and relationships. Participants who had been an inpatient in the previous 5 years had significantly lower percentages of long-term activities and place connections. In our study population, those with family and stable networks were less likely to have ever been an inpatient. We note that having family was important in personal networks and thus, where these relationships were absent, network enhancement strategies may be required to fill the gap. We observe that practitioner relationships can sometimes fill a friendship void but they did not provide access to resources in the same way as family or friend contacts. They were therefore not the solution except as a change agent, to empower, build skills and share ideas, but not to become the key network resource. A key strategy might be turning wider contacts into friendships where family were missing; practitioners might assist this process.

We also found that practitioners feature in all three network types in the inner circle of emotional closeness. To some degree, practitioners had an important role regardless of what else was in the individual’s network; they appeared in the inner circles of social networks regardless of overall social network size. We did not find significantly more people in primary care with a GP in their inner circle. GPs were in inner circles regardless of mental health contact.

While the amount of direct referral from practitioners to new place connections was low, this did not discount practitioners or informal network members encouraging participants to get involved in new things. We saw that more community place connections was associated with higher social support; it may be a two-way link where supportive network members encourage the individual to take steps themselves to make new connections. Informal contacts were more important than formal ones for referral but by far the most important factor was the individual themselves. The practitioner may be vital in helping build the confidence and resources of the individual to go and make these connections themselves. 125

How did people with severe mental illness use resources in their personal networks to support their health and well-being?

In conclusion, this chapter describes the heterogeneous social ties, places and activities that individuals with SMI draw on to access resources for health and well-being. There was evidence in the literature that connections to the community, to social ties and engaging in meaningful activities can improve access to resources, subjective well-being and quality of life. 11,111 Our study supports this evidence and gives some insight into how people, place and activity interlink. In our study the most important aspects seem to be community places which were social (and which supported engagement in activity); meaningful activities which were social and structured; and the informal network – not just the family, but friends and wider contacts too. Our data suggested most of these connections were built by the individual with limited referral by practitioners, but practitioners can have a role in encouraging or building the confidence of clients to make new connections. 228,229 Getting out of the house regularly was associated with increased access to social capital and larger social networks, yet the majority of our population spent more than half their waking time at home and had lower well-being and social capital than the general population. Therefore it seems that more needs to be done to support this group to develop and make use of personal network resources more fully. Network perspectives offer an insight into the types of connections people had and the extent of isolation, and crucially they provide detail on specific opportunities and gaps in the networks of connections that individuals have which can inform service interventions towards improving social inclusion and social capital and increasing activity levels for this group.

The next chapter will explore the meaning of these connections to answer the question of how people with SMI use personal network resources to support health and well-being more deeply.

Network meaning and social capital

Formal and sparse networks were characterised in the quantitative analysis by a higher proportion of contacts with health and social care practitioners, potentially providing fewer opportunities to develop an identity beyond mental illness or generate social capital. However, despite more limited resource opportunities, many people in this group described ways in which they derived sense of meaning and had created identities beyond their diagnosis. The non-formal aspect of their lives was valued; although people had fewer connections, those they had were important and useful. We found that several participants had purposeful strategies to ground their lives in daily routines. We focus in this section on examples where strategies are based on non-SMI connections to explore and understand non-formal aspects of a personal network labelled ‘formal and sparse’.

Geoff (SUSW34) described his approach as doing ordinary things as much as possible. He talked about how everyday interactions with ‘ordinary working people’ helped him to learn about relationships and how to communicate. He cited ‘banter’ with (non-SMI) friends as important in terms of helping to manage his voices and dispel paranoia. This helped him in terms of emotional well-being: interacting with ordinary working people encouraged Geoff to develop a sense of routine which was manageable and had meaning for him. He discussed how his flexible job with a local greengrocer ‘creates structure for me, um, well again it’s the responsibility thing’.

The example of Geoff also illustrates the barriers to routine imposed by mental illness. We learnt his medication made him sleepy – ‘I just feel, I’m very very tired’ – restricting his routines but overall he recognised that listening, noticing things and ‘maybe strike up a conversation and um, that sort of thing, you know it’s quite important’. His physical health was valued and he was an active walker. Geoff acknowledged that relationships were important – ‘it’s about being needed by other people in the same way we need other people as well’ – but can feel sad that he is not in a long-term relationship. Geoff explained that, although he accepted limitations caused by SMI, his well-being was nurtured through reciprocal friendships and work. Geoff values the building of social capital that these reciprocal non-SMI relationships provide.

Adam (SUL08) was similar to Geoff with respect to age, living alone and being diagnosed with psychosis. He explained that he was purposefully spending less time with people with SMI. Although he said that he was comfortable in the company of other service users, he stressed that he was socialising with a wider range of groups: friends from the pub, a social club with his neighbours and a church community. In those social circles the issue of mental illness would rarely come up and he added that he did not want other people to think that he was ‘crazy’. At the time of the interview Adam, who had been unemployed for a long time, revealed his personal struggles with not working, because, as he phrased it, ‘work is healthy and so the more I didn’t work the more my health d . . . declined’. He mentioned routines including going to the gym and a book club, as well as applying for jobs, but believed that his age counted against him. Ordinary life goals were a motivation for Adam: he emphasised his commitment to work and his ongoing efforts to find a job, in part so he could have ‘a companion to live with because it’s not fun to live alone’. Adam described that a consequence of having more non-SMI friends was that he needed less specialist mental health support; in particular he was dismissive of psychiatrists, who he felt could do little to help him any more. However, he was reliant on his depot injection but did not like attending ‘that mental place’ as there could be ‘violent customers’; he valued instead doing ‘normal’ things: ‘I try to do things normally, in the normal way and talk to them normally.’

Geoff and Adam exemplify a recurring theme in formal and sparse networks where there was a motivation to create routines and opportunities involving people to fill gaps, such as an absence of family and employment. We found other examples in sparse and formal networks that contained less ‘social’ connectedness but where identities and meaning had been forged that were linked to skills and interests. One such example was expressed by Ali (SUL75), who described himself being ‘outside of society’. He presented himself as belonging on the margins of society, and throughout the interview he stressed that he identified with socially excluded groups and had engaged in artistic activities such as drawing sketches of the homeless. Neil (SUSW26) likewise presented himself as having an artistic identity, and reflected that, although he has not been able to work for many years, he made a useful contribution to society, for example, through writing. Neil had built cultural capital and he valued this resource. He described himself as possessing ‘a literary grounding of a thinking man’. He also said that his painting prowess had developed to the point where he felt part of an artistic community. By successfully making the activity of painting meaningful to him he was now also developing connections with people, as evidenced by going away on painting holidays, and that moving away from the chaotic SMI community was empowering.

I paint. . . . I’ve got to the stage where I think I can honestly say I am an artist. Paintings are getting good enough, now. And um I just went on holiday with a group of friends and um it was fabulous. I shed a lot of my illness there.

Formal and sparse networks also included examples of people who were embedded within the mental health system and reliant on practitioner relationships for emotional, practical and specific mental health support. These people had very limited non-formal resources to draw on. The balance of formal to informal varied in the formal and sparse network grouping, impacting on access to resources and thus social capital, and this balance changed over time.

One way this balance could change was through reduction in the proportion of formal network contacts following discharge from community mental health services. This happened to Neil, who said he was initially apprehensive when the security of such support was removed, ‘there was a time when I was not getting support. I was concerned. I thought, if only they were still here . . .’ As time developed, however, he came to feel more self-reliant and independent by moving away from the mental health system, even though this was not his personal choice at the time:

I’ve learnt to survive and I am now strong emotionally, my stress thresholds are improving daily, um, and I’m in reality and, um, whatever that reality is. I’m in this reality and, um, life is good. And, um, I, I just don’t need to be cared for any more, I just don’t need it.

But now I say what do I want them for? Namby pamby, looking after me and telling me I’m a good boy? I don’t need that. I, I believe in myself now and I’m well enough and confident enough to go out there and mix with society and not be a threat to some innocent member of the public. It doesn’t exist. It’s gone. I, I don’t need any care now, it’s not necessary. It’s all down to me.

Neil’s example demonstrates that, although people with SMI can become dependent on mental health support and the ‘going it alone’ may cause anxiety, ultimately being away from the system, at certain times, can be helpful, creating space for confidence and self-reliance to develop. It may have been that for Neil moving away from service dependency was a factor in enabling his sense of cultural capital to flourish.

Valuing mental health supports?

The research explored the interplay in personal networks of informal support from family and friends, as well as wider ties, with formal supports from services. In this section we draw on examples to examine how people feel about mental health services being a presence in their lives. Within formal and sparse networks we found greater dominance of formal connections but these were not always well received: some individuals struggled with the loss of autonomy felt from experiencing mental illness and some people resisted any reliance on mental health services. Ali (SUL75) articulated this dilemma:

Well, the change would’ve come from when I was sectioned, 2002, and I think there’s been a complete shift towards . . . also a sense of, ah, of resentment of the Mental Health Services, but also a dependence on them. And maybe that the resentment and dependence feeding off each other, because often, when you’re dependent on someone you resent them for it, you know. You want to be your own . . . you’re not autonomous any more.

There was active engagement from services in Ali’s life. In contrast Bill (SUSW42) had limited engagement with services, which suited him, despite describing loneliness and isolation alongside paranoia. He was a man in his fifties who talked about how he was prone to feelings of guilt and isolation to such an extent that he had stopped activities he used to enjoy but again, as with other participants, was concerned to find employment and move off benefits: ‘I should be at work’ he explained. There seemed to be a disconnect between himself and health-care practitioners. Practitioners viewed him as mentally ill, whereas Bill viewed his problems in life as related to loneliness and being an outsider; he did not recognise a mental illness identity. Bill saw a GP and that was sufficient support:

I mean, I’m not going to try to get out of this job illness by claiming I’m more ill than I am, so, you know, no, I don’t think there’s much more they can do to help.

In contrast to Ali and Bill, others wanted high levels of support and welcomed having relationships with practitioners, speaking positively of the value they placed on these contacts. Pauline (SUL61) explained that her social worker and GP were highly respected relationships in her life, whose positive qualities compensated for other negative contacts, especially family:

You can’t have a family member because the family members have got no intelligence compared to the social worker, you know, and she’s trained for helping people. She’s the one that can help people. The other people can drag you down to unbelievable misery.

I’ve more trust and friendship, not friendship exactly, but respect I think the word is from these two [her GP and social worker].

So by going, dressing up smart and going to see the social worker, and coming to tell the GP about my troubles makes me feel that I’ve got the proper support in life just because they’re exceptional people as well.

It was evident that Pauline would turn to her GP and social worker for a wide range of support, whether emotional support or illness related, and these connections encouraged engagement in new activities. Both had roles as important contacts in her network as well as facilitators of network development. She explained that, although she was afflicted by voices which triggered much social anxiety – ‘they’ll start talking about fire, I’m going to be set on fire, I’m going to be run over, I’m going to get beaten up, I’m going to be scarred’ – she was still able to pursue lessons in sign language. As well as being a means of learning a new skill, it enabled her to address her extreme anxiety by being committed to supporting others and generated new social connections:

Doing the sign language is important because I’m meeting new people and that’s the thing. It’s like needing a blood transfusion, you know, for your life and, um, it’s important for me to meet new people. I’m going to be helping deaf people who are isolated, like myself, and I really understand how bad it is, you know, when you’re totally on your own and you can’t talk to people and you can’t make friends.

Pauline described the value she placed on her social worker encouraging her to engage in new activities; the latest area they were addressing together was her physical fitness. Developing new relationships was particularly valued by Pauline and she looked to her GP and social worker to help her make decisions over where, and how, to seek new connections.

Some people with formal and sparse networks turned to health-care practitioners for social contact to make up for a deficit elsewhere in their life; Gregory (SUSW10) explained that he went to see his GP sometimes, ‘just because I need to speak to someone.’ Even though he rejected his diagnosis of schizophrenia, he appreciated that his GP had broached the subject in an ‘open and tactful’ manner.

Notable elements of formal and sparse networks, alongside challenges of managing SMI, were inner strength and resilience, identified as ‘building blocks’ of recovery by one lived experience researcher. These might be very small signs of agency within the network or more substantive examples where individuals were driving actions in their networks, developing a greater sense of autonomy and independence. Practitioners were acknowledged as one influence in recognising the potential for change. Barry (SUL78) had taken the mantra of his social worker to heart: ‘don’t let life beat you’.

We summarise in Table 30 what we have learnt about formal and sparse networks in terms of five key themes highlighted in the first section of this paragraph. These were areas within our literature review (see Chapter 2) and were used to structure researcher reflection notes within the data synthesis process in order to answer the study research questions.

Family routines and support as a positive feature in personal networks

For some of the people interviewed, family life was an intrinsic part of their recovery journey, a valuable source of support for their own illness, delivered by children, partners and other relatives. It was this support that helped to give their lives a sense of structure and meaning, sharing activities and providing responsibility and purposeful roles. Families were described as providing far more than emotional support. We focus in this section on the role of families in recovery.

Jackie (SUSW09) reflected on how she had become more outgoing and that her network had grown as a result of her routine of looking after her young son. She had recently developed some close friendships and had also engaged in more activities. Routines revolving around her son stood out as positive features, providing variety and escaping spending too much time at home, which reinforced a cycle of guilt and sense of isolation. However, it was also evident that this came at a personal cost and that her determination to support her son could be challenging: ‘I’ve always said I’d never let it affect him, never show it to him, which is . . . that can be hard.’

Sometimes when you’re having a really bad day, it’s quite hard to pretend you’re not. Some days if I’m really tired I feel really like I couldn’t do something and I know that he needs to get out, that can be hard, but I do it.

Nick (SUSW51) likewise talked about how his child care routine was a positive influence. In a number of interviews there were accounts of how the pattern of family life in itself provided sense of meaning and purpose. Samantha (SUSW60) reflected that in the past she had thought that she would not be able to settle down and live an ordinary life because of her earlier experiences of bipolar disorder, but that getting married and buying a house with her husband was a significant event providing security and stability, a solid base to connect to the ‘world’ from:

We just felt this is a nice place to live, and it’s got a good feeling about it. And we’ve been there ever since and don’t want to move.

The enjoyment of her life at home was a key part of her recovery story in contrast with her experience at work, which she described as ‘extreme pressure all the time’. Samantha also explained how having a supportive and intuitive husband was helpful in terms of managing bipolar disorder on an ongoing basis:

My husband is the most important one: we’re very close. He knows all about my situation, and he’s very supportive. He seems to have got this sort of intuition as to, sort of, when if I’m sort of not quite right, if I sort of leave him to sort of cheer me up, or if I need him just to leave me alone for a bit.

There were other examples in interviews of family being central to identity and stability. Gavin (SUL34) described his uncle and father as being ‘like a rock foundation to try and help me boost myself up’. Seeing his children once a month was also identified as the highlight of his life:

I know I’m not sort of like, um, like 100%, you know, with it but at least, you know, like, you know, I can show them the love and caring and respect, do you know what I mean? And they, they, um, they understand, do you know what I mean?

Gavin identified his family as providing essential support in terms of helping him to stay well. He explained, for example that they regularly telephone him to check how he is doing:

. . . which gives me that, um, that inspiration of happiness, because like they say, it’s a serious thing. All you need is a phone call can cheer you up, you know.

The support Gavin received from his family seemed particularly important in contrast to the rest of his network. Many aspects of his life had diminished over the years, including engagement with key activities such as work. The family was a core part of Gavin’s social relationships and without them he would have been left socially isolated. Dora’s (SUSW61) interview also demonstrated the enduring value of family support, especially through times of upheaval and change. She had recently separated from her husband and had moved to a new home with her two daughters: ‘we’ve been through a lot together’, and she reflected that they had grown stronger as a result. Dora also said that it was important for her to be accepted in the local community, which was likened to ‘a goldfish bowl’ at times, for the sake of her daughters. The need to be accepted by the community felt pressuring and challenging, partly because of the stigma of mental illness, but also for a central part of her well-being. She also reflected on the impact of her illness on her eldest daughter: ‘sometimes I feel sad that they do [understand about my mental health] because it changes you as a person . . . it affected her school work. I’m sure, you know, I’m sure she would have had an easier time if this wasn’t there’. Her extended family, despite her separation from her husband, also remained important in her routine; she supported her in-laws, doing jobs for them. This evidenced again the importance of reciprocity in relationships and the different forms it can take.

The influence of family in personal networks

Our interviews revealed some challenges for recovery with respect to family support. We found that high levels of family involvement can come to dominate to the point where these relationships act as a barrier to new opportunities to build connections. A similar risk might be found with over-reliance on practitioners in formal and sparse networks, though family connections tend to last longer; the key was what roles were occupied by family and how far the individual with SMI was in control. Where family dominated to the extent that people could not imagine an alternative future and as a result lived with short-term perspectives on managing SMI, there can be a detrimental impact on personal recovery. We focus on examples that reflected on the influence of family in the network in this section.

Eleanor (SUSW38), who was in her early forties at the time of the interview, discussed having a close attachment with her mother to the extent that when she was a teenager she dreaded becoming independent and had been living at home ever since:

I’ve always had a really, really strong relation, I was always a mummy’s girl . . . And I have always had a very, very strong dependency on her. I’ve never been able to be independent in the way that other people grow up and become independent and look forward to their independence. When I was in my teens I absolutely dreaded it. I couldn’t bear to think about working or responsibility. I’ve never worked. I’ve been to interviews, but I’ve never managed to work.

Eleanor revealed that over the years she struggled to engage in activities under her own initiative but she was aiming to take more responsibility over her own recovery – ‘I think I do need to learn coping strategies’ – even though this was deemed difficult and she felt weighed down by her depression, ‘I realise now that it is down to me and I think that when I’m ready it will happen.’ We highlight that it can be challenging to break dependent relationships, and acknowledge that ‘being ready’ is an important part of recovery from SMI.

Joe (SUSW77), a man in his early thirties, discussed how he was very reliant on others to help create a sense of routine and engage in activities. He said that he felt close to his mother, who provided him with ‘organisational’, as well as emotional, support, while he revealed that ‘to do stuff’ with friends was a key strategy to help cope with his voices: ‘concentrate on that rather than concentrate on what people [voices] might be saying.’ Joe talked about potentially moving away from the small town where he lived to a nearby city because he regarded it as more ‘cosmopolitan’ and ‘job opportunities and education opportunities are better there.’ However, there were few concrete plans shared to understand how this ambition would be fulfilled. Such a move would impact on his family as well as himself and therefore barriers to change could be both internal and external to Joe.

At the time of the interview Ed (SUSW16) was recovering from a recent hospitalisation. He was experiencing an associated loss of control and found it very challenging to conceive of what his future might hold: ‘I just cannot answer that.’ His life was focused on the short term, taking one day at a time. He reflected that he could go from feeling well to unwell in a short period of time, and that he was trying to maintain a balance between being active and resting but found it difficult, as his health and well-being were delicately poised. He said that he was reliant on other people, particularly his wife, to gauge how well he was doing, but some of the support she was providing made him feel as if he was being left in the dark:

I’m fairly sure she’s doing things to support me with other people but isn’t telling me what she is doing, and I’m not necessarily aware of it so she is protecting me maybe from certain questions or dealing with certain things.

Ed’s story provided a clear sense of the impact of managing SMI for the individual and relationships within the family. Struggling to cope day to day, knowing people were supporting you, was positive but could have disempowering consequences.

It is important to state that people with SMI can restrict the opportunities of those around them at times. Bradley (SUSW68) explained that he would often devolve tasks and responsibilities in order to alleviate his own stress: creating a routine which was simplified and more manageable for him involved placing extra burden on his partner:

That’s right. I’m not going to do it if I don’t have to, sort of like thing, you know. I’ll just bung it to [name of woman: partner], some of my voices said. I’m going to laugh it off, see, but it’s quite cruel. But it’s quite funny, well, you know what I mean.

In other interviews people with family and stable networks did not always identify their family as constituting a helpful part of their life and recovery. Families were sometimes identified as a passive or unwelcome presence. Bradley (SUSW68) revealed that he had endured a difficult relationship with his mother and step-father, who had lived close by throughout his life: ‘they were cruel to me in the past, and cruel to me now.’ However, he had developed a close friendship with a man whom he described as being ‘like a father figure’, who helped him to engage in a regular exercise regime and also helped him to cope with his voices. His experience was mirrored by Jackie (SUSW09), who explained that she was estranged from parts of her family. Over recent years she has had a very negative relationship with her sister. Contrasting with this, Jackie explained that she had formed a relationship with a close friend, whom she likened to a surrogate mother and who provided her with social and emotional support lacking from her own family: ‘she accepts me for who I am and she’s just pretty laid back and always good with advice.’

Carl (SUSW41), who was in his twenties, adopted a disparaging attitude towards others and did not identify any relationships as important or significant: ‘I’ve got a very strange and maybe slightly dismissive attitude towards maybe everyone I know.’ He mentioned that he met with his mother only out of a sense of obligation, and that he did not really trust anyone, including his wife, whom he had recently married.

I’m quite happy being my own person all the time and if that means that I end up with a closer-knit circle of friends, so be it. And if it means that I get stuck at home a little bit, then so be that too.

Daniela (SUL41) described members of her family as having a less than positive influence on her well-being: ‘they’re just disruptive people to be with, I think. And when you’re feeling very bad, that’s really just more than you cope with, in my experience.’ Although she reflected that her family constituted part of her network she ‘can’t stop seeing,’ she explained how she compensated by developing her life elsewhere by deliberately adding people to her network who increased her engagement in social and cultural activities:

I wanted to find people that . . . I mean, that . . . there are certain things that I like doing in my life, like I like going to exhibitions . . . something like that. But it’s much nicer to go with somebody than go alone, and none of the people that were currently in my network were either . . . they were either not interested or not available.

The importance of family-like relationships was illustrated where challenging birth family connections can be replaced by others that can provide similar types of emotional and practical sport. We summarise in Table 31 what we have learnt about family and stable networks following five selected areas, supported by the literature review in Chapter 1.

The potential within diverse and active networks to support well-being

The interviews explored with participants the well-being impact of connections in their networks to people, places and activities. A notable aspect of diverse and active networks was efforts by individuals to manage multiple connections and maintain routines as part of managing SMI. We focus in this section on the well-being impact of networks by exploring the strategies used to support personal recovery.

John (SUL39), who was 29, explained that he had an active approach to recovery, ‘with schizophrenia, the more time you have on your hands sitting down doing nothing, the more you think about things.’ Personal resolve to create a routine to aid recovery came across strongly in the interviews. John reflected on the past and his determination to avoid relapsing back, describing it as ‘like an inner determination, isn’t it? That you don’t want to get into that situation again.’ He also underlined the importance of engaging in a range of diverse activities since becoming ill; he reflected that he had more free time, and hence opportunity, to engage in new activities such as learning to swim and speaking Arabic. He was also motivated to do so, finding out about opportunities, seeking them out. John also indicated that he had strategies for managing different parts of his network, for example he would no longer visit friends’ houses if he thought street drugs would be consumed:

I think cannabis was one of the main reasons why I ended up going into hospital and developing my medical condition, so, basically, um, I would say, um, sometimes, um, if we’re going to meet up, rather than going to his house where there’s going to be cannabis being smoked, um, we just go out for something to eat instead.

This touched on an important aspect of diverse and active networks, because they reflected a wider range of social groups and interests than other network types, and they often required more active management. One such strategy to support well-being identified in interviews involved not disclosing mental health problems to specific parts of a network, guarding against the stigma of mental illness labels. Kevin (SUL29), for instance, revealed that he purposively decided not to disclose to his martial art group his experience of bipolar disorder: ‘they might think, you know, if he’s unstable, he shouldn’t be learning a martial art, do you know what I mean?’

Um . . . (pause) what doesn’t work. My biggest thing, possibly, would be this group here [pointing at the map] . . . of me not telling them. So that-that’s . . . I wouldn’t tell them the reason why I’m not coming, or . . . um, I won’t . . . I tell them very little about my past.

Kevin’s management strategy for disclosure was to be very careful before talking about his bipolar disorder with contacts because he believed that there was a lack of widespread understanding and that he might be judged differently. He attributed multiple break-ups with past girlfriends to the fact they did not view him ‘in the same light’ when they found out about his diagnosis of bipolar disorder:

No, she-she knew I had bipolar disorder I have had a girlfriend in the past, back in-in-in my late twenties, who I told and she never looked at me again in the same light.

In his interview it was evident that Kevin found the label of being diagnosed with bipolar disorder challenging to accept and was resigned to having to manage the symptoms of mental health in the future; however, he viewed being discharged from specialist mental health services as a future goal that would represent a ‘kind of victory’. He also expressed the importance of self-determination and his own personal ‘empowerment’, and added, ‘I wouldn’t lead a passive life.’ He regarded his own recovery as a ‘slow build’ that he had developed purposively through routines to provide a sense of structure outside his home:

It is easy to isolate. See, this is why I do my martial arts. This is why I do, you know . . . go to mass on a Sunday. It’s a . . . it’s . . . go to the supermarket every night, rather than buy your shopping for a week. It’s to get you out of it [the house].

The participants’ accounts demonstrated that diverse and active networks could encompass a routine of activities in the community, activities which within themselves have meaning and enhance well-being, but also act as a safety net against social withdrawal and help to promote a sustainable recovery. This was particularly important in networks where a person was not in employment, building alternative routines and structures. Such routines were often self-initiated; Ellen (SUL28) explained that she was successful in identifying and accessing community resources to support her recovery. Her library, for example, served as a hub in terms of accessing new social groups and activities, such as a book club.

The potential durability of diverse and active networks is demonstrated by the wide variety of interests they included so that if something stopped another connection would often fill the void. Roger (SUL10) described how playing football had been a major part of his recovery, valuing the ‘camaraderie’ and that ‘being part of a team is quite, um, does make you feel more part of something and secure’. Because of a physical injury, at the time of the interview he was temporarily unable to participate in football, but found a substitute activity, of sorts, in the form of swimming with his girlfriend: ‘that’s almost as good. It’s a little less social but it’s also kind of a release.’ Roger’s narrative highlighted some of the ‘building blocks’ of diverse and active networks, and by association recovery: developing structure and routine through being active (supported by others), rather than being defined by illness, which in turn can lead to a growing sense of connectedness.

. . . after a while my mum got a bit more fed up with me being as depressed and then kind of encouraged me into going to uni and my girlfriend would be sort of more willing to push me into doing things when I was depressed rather than sort of backing off and then letting the illness be, more define me, more just pushing me into doing things and, you know, encourage me that I can do them and if I do these things then maybe I will get better quicker, you know, I think making you feel more capable rather than making you feel incapable, which you already feel.

Teaching helped him to feel part of a community:

I mean just walking to and from school, you know, the pupils wave to you, you know, yeah, you kind of feel like you’re contributing to that community, so that’s good.

The ‘good’ that Roger described translated into feelings of positive well-being for him. We note that diverse and active networks tended to contain an array of ‘building blocks’ so that if one connection closed others could be pursued. Also, people managed connections that were no longer supporting their well-being by removing them where possible.

An important aspect of accounts from people with diverse and active networks was how they had responded to problems in their past, overcoming obstacles. Memories of being treated in the mental health system were cited. Andrew (SUSW46), for example, became unemployed after a severe ‘manic’ episode, and was told by mental health professionals ‘you’ve been very ill, don’t expect to walk back into your career.’

So there’s a lot of negative suggestion I received. And so to come back from that takes a huge amount of courage and a huge amount of strength.

He defied this gloomy prognosis and at the time of the interview felt ‘strong’ and was working full time providing support for people with severe mental health problems. Practitioners acted as barriers in some networks by holding attitudes that demonstrated lack of hope and limited life expectations for their clients. Likewise Nigel (SUSW36) drew on his past experience of mental illness to help others through one-to-one mentoring:

And within that project I’m doing the one-to-one work with him, helping him because I’ve been through similar sorts of things in some ways in mental distress and now recovered bipolar even though I’m still vulnerable and need support in the system. I’m doing things to help people if I can and then to helping, and he’s helping me at the same time – it’s not just a one-way relationship, it’s a symbiotic relationship.

The well-being gains of helping others, establishing reciprocal relationships, were noted among the diverse and active group. These creative decisions demonstrated how individuals recognised the importance of connectedness as a specific strategy for managing well-being. Personal resolve to create a routine to aid recovery also came across strongly in interviews with people with diverse and active networks. We found an emphasis on generating new connections and activities, often forward looking. Part of this could entail taking personal ownership of recovery, rejecting medical explanations and treatments such as medication, as illustrated by Michelle (SUSW53): ‘I stick with my labels and don’t use the medical labels and that helps me’. The network-mapping approach supports this personalised focus. Michelle, for example, said that the initial network-mapping interview highlighted the fact that, although she had a diverse array of interests, there was a ‘big gap’ in her life and she addressed this by forming a close attachment with a dance community.

That was really useful and I’m wondering whether I’ve filled that because I’ve become aware that that was empty? So that was really useful for anyone to do I think.

There was an interesting contrast therefore between Michelle’s experience of the medicalised/diagnostic model of treatment, which she felt detracted from her sense of individuality, and the network-mapping approach, which embraced it.

Were bigger networks better for managing well-being for those with severe mental illness?

There are a lot of normative assumptions attached to studies of social networks because of the emphasis on measurement of social ties and studies that suggest larger numbers of social contacts are positive for mental health. Work assessing social capital also had this bias because resource generation was aided by social contacts and thus larger social networks provide access to more opportunities for resource generation. Our network mapping showed how connections to place and activities were also positive for generating social capital because, through connections to place, social contacts may emerge. Qualitative data also helped unpick the problematic ‘bigger was better’ narrative by exploring the complexities of optimum personal network connectedness. This is the focus of the next section, using examples that emphasise the heterogeneity of connectedness experiences.

We found not everyone felt satisfied with having a diverse and active network. Not all social ties were qualitatively equal; some were felt to be more positive, or negative, while others were considered inert. The same was true for activity and place connections. Margaret (SUL05) expressed an active desire to reduce the amount of social contact in her life; at the time of the interview she revealed she had ‘ostracised my family at the moment’. She reflected that people in her life could be emotionally draining, even ‘overwhelming’: ‘I tend to internalise other people’s problems. I’ve always done this and I find it difficult to remain emotionally detached.’ Given this, she mentioned that she would like to have fewer friends in her life in the future, stating ‘I need time to myself’. She also expressed mixed feelings about living in London: on the one hand she found it stressful and challenging to navigate, but she also acknowledged, ‘without it I’d just shrivel up . . .’ The accessible culture (museums, galleries, music) constituted an important part of her life and she considered art to have a therapeutic value, especially the ‘masters’.

Zara (SUL48) reflected on the paranoia of worrying about what the different social groups in her life, which she referred to as ‘boxes’, would discuss when she was not around. As a result she tried to keep them apart, a strategy used across the network types:

I don’t know, I’ve always been like that . . . I do, I do have paranoia about people talking about me. And especially because they, you know, if I, if I have been unwell recent, recently, or around people, and I wouldn’t want people talking about me behind my back, and if I keep people, kind of, separate then they can’t do that.

Diverse and active networks do not always hold together and when this happens people might require additional support from others, including practitioners, to buffer against negative impact on health and well-being. Suzi (SUSW48) reflected with some sadness on how her social group had become reduced by factors beyond their control:

What’s really sad and has really got me down, has got me this year, is my university friends, one by one, dropped off. There are different reasons for each of them, I suppose, but that’s how it’s happened. And I’ve tried to, sort of, keep things up a bit on Facebook [Facebook Inc., Menlo Park, CA, USA]. But no, people, some people just . . . I mean, a lot of people just, kind of, live where they are now, don’t they? And they’re not necessarily bothered about you even though you were a good mate at the time.

Diverse and active networks can also have negative elements, or aspects that need attention. We found when people with SMI were seeking out opportunities in the community this could mask issues closer to home, potentially destabilising their mental health. For example, Liam (SUL32) described how his untidy home influenced how he felt about himself, and was wary about inviting people into that private space; this was made more frustrating by health practitioners’ lack of understanding of how important this issue was for him and his inability to address the problem on his own:

I’m also escaping because my flat is in chaos. There’s a slight sense of hoarding, slight lack of elimination, slight problem about putting the place in order so it looks good. So I get up in the morning and . . . and I’m more likely to get on the phone to my girlfriend and try to arrange to meet her than I am to deal with sorting out the stuff inside my flat. So there’s . . . some of this engagement of healthy stuff is also an escape from dealing with my own stuff and getting my own stuff together. . . . If you were to walk in my flat you’d say oh, there’s something going on with this person. So, um . . . this, um . . . that’s kind of . . . remains a challenge, and it’s indicating to me that something’s not quite right with me and I need to try and work on it but I kind of avoid working on it. And if I . . . and I also feel if I put it right I’m going to . . . it’s . . . I’m going to end up making it in a mess. When I try to discuss this with health workers they kind of dismiss it. It’s like, this guy’s only going on about an untidy place, but it’s very significant for me, and I can’t get it together.

Although in this example there were frustrations about the ‘dismissive’ attitude of practitioners, it underlined how important it was for mental health workers to actively monitor connections to people, places and activities as well as their impact on well-being.

The role of practitioners in diverse and active networks

In this final section we consider the role of practitioners in diverse and active networks; although they constituted a smaller proportion of network ties, they were still present. People with diverse and active networks often emphasised the importance of practitioners for encouraging interpersonal skills and support and valuing the social aspects of people’s lives. Zara (SUL48) reflected that her social worker had played a constructive role in her recovery, describing her as ‘really helpful’; she also added that the relationship was particularly important when she had been less well and she had learnt to be more honest when she was struggling. Her experiences illustrated that, for some people with SMI, having effective ongoing social support can fulfil a vital role, and that it can be easier to talk through issues with professionals rather than family or friends:

keeping me rational and letting me talk through stuff and not [sigh] because, I guess, like, I don’t want to talk to my friends about my stuff, so, and I don’t want to talk to my partner about a lot of stuff . . .

Practitioners helped individuals develop their own coping skills. Zara highlighted that the importance of mental health workers being able to listen was more important than the actual technique or therapy used. Contrastingly, practitioners with a perceived deficit of empathy were criticised, found to be disempowering or disrespectful. Duncan (SUL77) reflected that he had negative experiences in the past of mental health workers who had been overpowering, or spent sessions obsessively making notes: ‘they don’t say a word, nothing, and that’s it, and they just write whatever they write and you don’t see it.’ Practitioners whose encouragement was viewed as unhelpful were also criticised. Kevin (SUL29) reflected on a time in the past when he was struggling and found some of the encouragement from practitioners to engage in social activities to be misplaced. He said of one mental health worker, ‘she insisted I went out and to sort of social interactions to improve my social life. It’s a bit like telling a man with a broken leg to do . . . to do a marathon’. An important emphasis in people’s own accounts was how change had to be led by them; SMI does not take away individual interests and preferences but motivation can dwindle and stigma can act as a significant barrier. Wider societal barriers, such as difficulties gaining employment or public attitudes towards mental illness, did inhibit individual agency.

In contrast to Zara, discussed above, who said she would rather talk to professionals than friends and family about her problems, for Roger (SUL10) the opposite was true: his family had become more influential and services less so.

I think in some ways it’s easier to ignore the advice of professionals actually. Although I respect their advice it kind of . . . when it’s people that are more close to you and personal then you feel more, like you’re doing something wrong by going against that advice, whereas professional advice is more like they could give that to anyone.

For John (SUL39), too, services had become less involved in his life. In the past he had a proactive CPN who supported him in recovery from hospitalisation, providing help with medication, accommodation and looking ahead to a career. Currently he said he was receiving more support and encouragement from his family, and the role of mental health services had diminished. Some people were actively shaping their support requirements for themselves. Andrew (SUSW46) reported negative experiences of mental health services but was currently in a positive psychotherapy relationship, which he arranged himself, and felt was more attuned to his needs. Michelle (SUSW53) felt well supported by a shiatsu teacher who helped her ‘see through all of my surface crap.’ This contrasted with past experiences on psychiatric wards, and patronising psychiatrists: ‘there’s no way I can reach them or they can reach me.’ The theme of individual agency being in conflict with health-care system constraints was also emphasised by Maggie (SUL05), who, reflecting on her past experience, said ‘I’ve never felt that I’ve lost control personally. It’s other people’s control that’s been imposed upon me’.

Change also featured in diverse and active networks. However, it is important to stress that recovery was not a linear one-way transition towards empowerment and agency; people with diverse and active networks could struggle with managing their networks, or aspects of them. It was thus important that services were adaptable to increasing or reducing support as required. Donna (SUL23), for example, felt stuck in the mental health system and wanted to break away, but also recognised that she was not well and needed to be supported. This sense of dependency was also paired with fear of losing independence: ‘I try to avoid the walk-in clinic, it’s fear and a fright of them trying to re-admit me back into hospital.’ Donna reflected that she would like to move on but added ‘I’m not ready yet’. It was also evident that the potential threat of discharge was one which caused her great angst, being told ‘there’s nothing more [services] could do, that would be dreadful’. Service involvement can be regarded as all or nothing by people with SMI, with discharge viewed as the loss of all support. Donna said that the ideal situation for her would be one where she could move forward and lead her own life, but with the reassurance she could call on support again if she needed it: ‘But they could say to me “yes, we’ve done so much for you, there’s nothing more you could do” that would be dreadful.’

We summarise in Table 32 what we have learnt about diverse and active networks in terms of five dimensions we have brought together in this chapter to reflect on.

Clarity and appreciation

Taking part in the exercise was described by some participants as interesting because it presented information in a different way. Kevin (SUL29) found it ‘clarifies’ a ‘complex situation’. Bradley (SUSW68) explained: ‘Yes, it was good. It was quite surprising, yes, how close I am to some of them, how not, and who knew, you know’. The value of taking part in the network-mapping exercise was explicitly referred to by some participants, including those for whom it just made them appreciate what they had. Roger (SUL10) said:

Yeah, well it’s made me realise how lucky I am in a few ways I guess, to have so many people I can draw upon. And yeah, I mean it’s made me kind of realise how important the social side of work is as well, but yeah, which I didn’t foresee happening before I got this job at all.

Bryony (SUL65) found that the exercise had made her value the people in her personal network even more. ‘I think it makes you think, like, you probably should value, value the people a bit more.’ The exercise also allowed people to express satisfaction with their personal health network overall, not just components within it. This aspect was of particular importance to Eleanor (SUSW38), who thought others might criticise the size of her personal network at 14 contacts. What was important for her was not the overall numbers, but that her personal network functioned well for her.

It reminds me of just how safe I am and that compared to millions of people I’ve got, although there aren’t like, hundreds of people all over here, like some, sort of socialite or something, I’m happy that I’ve got the family that I’ve got. I mean, you could know dozens and dozens and dozens of people but only really have one friend out of the whole lot. So I don’t think it’s a bad thing having a small group.

Sam (SUSW60) also emphasised that, although she thought others may be critical of the number of people within her personal network (the network size was 18), this was not a problem to her because this size of her personal network worked well for her. She described her network as ‘close’, ‘happy’ and ‘big enough for me’.

As I say, I’m not sort of unhappy with my map. There aren’t that many people on there. I’m sure you’ve spoken to people whose maps’ been completely filled in . . . but I don’t . . . you know that’s not something that bothers me. I don’t want to be sort of surrounded by hundreds of millions of friends all the time. I quite like my own sort of space, which is another reason I go on holiday on my own. I like to be on my own.

Value added

The utility of the network-mapping exercise went further than clarity and appreciation for some participants. The ways in which the exercise added value included raising awareness of resources; identification of gaps and vulnerabilities; and stimulus to plan proactively for the future. Several participants reported that it was worthwhile, in and of itself, just to think about what resources were available to them and how they used them to support their well-being.

Jackie (SUSW09) reflected ‘It makes you realise who’s in your life and where they come really’. Joe (SUSW77) found the clarity of people’s roles in his life beneficial: ‘Clears up what my life is and my friends and family are like. My life could be a lot worse’. Valerie (SUSW15) was relieved that, although there were lots of medical people on her personal network map, because of her physical and mental health problems, there were also ‘other connections’ as well. Undertaking the exercise had reminded her of activities that she used to enjoy, but no longer participated in, such as going to the theatre, and as a result she was now challenging herself to ‘break the barrier’ of not driving at night, so that she could do this again. Raj (SUL07) had also been motivated to make better use of the resources within his personal network including intentions to socialise with his friends more and to do a vocational course.

Mapping personal networks had made some people aware of resources within them that they could make better use of. For others, the exercise made them aware of gaps and vulnerabilities within their personal networks. Ed (SUSW15) was facing numerous uncertainties about the future and had been in contact with SMI services for only a year.

I think it clarified for me where, that actually what I thought was a network of supportive friends, really came down to two or three people.

Adam (SUL08) in undertaking the exercise had become aware that he needed more friends. He talked to his friends about the exercise and they responded by saying to let them know if he ever needed help or support. For Daniela (SUL41) the network-mapping exercise had made her more aware of distancing herself from those who had a negative impact on her well-being and an awareness that she could ‘possibly’ focus on the gaps. ‘I do everything to sort of mitigate those . . . those contacts by just emotionally removing myself from a situation’. Pauline (SUL61) also found that being aware that things were not going well with several people at the time, and seeing this mapped out, helped her to be aware that she needed more people in her life. She said that the exercise had helped make her plans to help others in the near future ‘clearer’. Agnes (SUSW40) also experienced the process as a stimulus to review her life and proactively plan for the future.

It reminded me of things, having another look at it. Um, instead of just blundering my way through life, I’m sitting down and reminding myself of how everything is.

Support

Taking part in the personal network-mapping exercise had resulted in some participants benefiting from changes they had made as a result, including socialising with their friends more and making proactive plans for the future. Donna (SUL23) had also made changes as a result of the exercise; these changes had not, however, resulted in the benefits that she desired.

It made me realise that I need to go out a bit more, maybe to talk, um, but I have done that. There’s not much improvement, um.

It was clear that a number of participants required support to both unlock the potential well-being opportunities within their personal networks and to expand them. Liam (SUL32) found the exercise ‘excellent’ but would be likely to require this type of support, as he noted in the interview: ‘But what support is going to be available to these people to help them engage with these networks that have been helpful to them?’

Bill (SUSW42) presumed that the purpose of the exercise was to show that he lacked ‘real’ friends in his life, describing his network as ‘sparse’, ‘sad’ and ‘lonesome’. He said that he could not change his life and gave a list of reasons for this, which included a lack of finance, motivation and musical skills. This individual would need support from someone else to see how he could begin to break down these obstacles and work towards what he would like to see in his personal health network.

As well as support, other suggestions were made, in the interviews, to increase the utility of the personal network-mapping exercise for those undertaking it. The first suggestion was to improve the materials used and have more time. Liam suggested that it had been very useful to have two meetings to think about and discuss the exercise. He found that he had been quite emotional in putting the details down in the first meeting. He found that the second interview, as a ‘follow up’, was more ‘powerful’, because he could be more detached and rational.

The potential value of practitioners, or other support workers, undertaking the personal network-mapping exercise with people could be clearly seen in the response from Gavin (SUL34) when he was asked if it had made him want to make changes in his life.

The impact on mapping out a network is a very serious thing. This network is not just something on paper. It’s something of life, and what it is is it shows you where you’ve been and where you can get to. Where you’ve been is everything that’s happened but where you want to get to you still have to try, and it’s on the network and it’s still there. And it’s still there with people, like I said, like, you know, social worker, people that can help you.

In summary we found broadly positive support for personal network mapping because:

• It was person-centred, focusing on individuals and what was important to them in their lives and how this developed over time.

• The process itself was a useful way to open up conversations about people’s interests, relationships and place connections. The visual dimension facilitated another way for people with SMI to chart and explore social recovery, emphasising the value of resources within themselves and their networks.

• It provided a structured way to explore two abstract concepts – agency and connectedness – looking at the past, present and future.

How useful was the concept of connectedness for managing well-being for people with severe mental illness?

We learnt across networks that building connections to people, places and activities fashioned individual identities. Research has shown the importance of valuing identity formation, and understanding the complex social and political processes that exclude or include people with SMI, and the follow-on impact on their identities. 232 A recurrent theme in people’s accounts was reference to work, either as a presence with positive and negative impacts on well-being or as an absence within the network. Work connected people to important resources, such as ‘ordinary working people’ and ‘healthy activities’. It provided a benchmark identity that people were judged against, both by themselves and by others. Where work was absent, connections to interests provided an alternative role to that of being ill. Across all network types, people sought to distance themselves from the status of being mentally ill, and building connections to different groups of people, creating routines around the ‘ordinary’, helped to achieve this goal. Stigma was a barrier, and we heard how people managed non-disclosure in order to protect themselves from negative associations attached to the mental health system. However, this was not always possible and stigma was a major issue for practitioners, as it is one reason people are so reluctant to seek ongoing help for fear of being tainted with labels such as ‘mental’ or ‘crazy’. 156,227

We heard how connectedness provided people with purpose in life and routine. We also heard how people felt connected in different ways; limiting network contacts was an important well-being strategy, linked to creating manageable and coherent life-worlds,233 but so too was seeking new interests and friendships or acquaintances. Through relationships to people, places and activities, individuals with SMI develop appropriate timetables and structures that support well-being. Being able to contribute to society in a meaningful way, be that through sharing time, skills or friendship, was an important part of recovery. Friendships that involved reciprocity were highly valued by people in each network type. Quality of relationships, rather than the number of ties, emerged prominently, and people across network types explained how they are active in managing their network to rid it of troublesome contacts, where these are under their control. Where there was lack of network structure and routine, practitioners, or others, can help individuals with SMI look for ‘building blocks’ on which to plan person-centred recovery goals when they are ready to do so.

Practitioners also had a role in people’s actual routines: they were part of the connections map for people with SMI. All participants in the study had a health-care contact for managing SMI, but some was light touch through the GP and others had intensive support packages involving multiple members of a Community Mental Health Team (CMHT). We identified varying levels of engagement with practitioners. We found resistance on ideological grounds, such as believing problems were not illness based but had social origins, because of the problems caused by stigma and poor-quality staffing where practitioners lacked empathy and hope and carried a deficit view of mental illness. However, the value of connections with practitioners, and their knowledge, skills and expertise, was also identified, such as when practitioners noticed the potential for change in their client’s life, empowering new skill-building and giving confidence, or where help was provided to extend social support networks and develop new social capital, particularly through bridging contacts. We noted that an important aspect of the dynamics of personal networks was the ability of individuals themselves, and those around them, including family and friends, to identify opportunities for new connection-making and enacting internal or external resources to take advantage.

Reflecting on connectedness, one dimension identified by participants was the process of ‘moving on’ from mental illness and how developing ordinary routines supported their well-being journey(s). The fluctuating nature of mental illness meant that change was a strong feature of people’s accounts as routines were adapted around how well people felt and past illness experiences were an important reference point for future decision-making. Recovery journeys are non-linear and complex. We were particularly interested in looking for signs of active decision-making and an individual’s control over their personal network. Building blocks for change were found in all interviews, though the degree of described execution, turning ideas and plans into actions, varied. Unsurprisingly, people with formal and sparse networks and limited social support reflected on the challenges of turning their ideas and goals into plans. Those with diverse and active networks had more experiences to draw on and strategies for dealing with setbacks.

How are effective personal networks built?

In addressing the study research questions, we looked to the qualitative analysis to steer us on the role of practitioners in personal networks and how network resources are used by people with SMI to support health and well-being. Participants told us about features of effective personal networks. These included:

• Resilience and self-determination within individuals to actively manage SMI and social recovery.

• Management of stigma and discrimination, both anticipated and experienced, resulting from a diagnosis of SMI; having active strategies to build an identity that was not defined by mental health as well as combat stereotypes of being mentally ill.

• Social support from friends and family, with reciprocal exchanges of support valued.

• Identification of opportunities and having resources within networks to make use of these opportunities to build new social capital. Practitioners play a useful role, not in providing social capital, but in supporting new connections to activities and places that can lead to new social contacts.

• Informal connections that were nothing to do with psychiatry. Ordinary relationships, activities and place settings are valued because they provide a purpose and identity that is not SMI; these may be mainstream activities or those on the ‘margins of society’, based on individual preferences and interests.

• Strategies and plans to address an individual’s goals which are led by the person living with SMI. Others within the network will make suggestions but recovery is owned and driven by the individual. Setbacks are supported.

• Meaning in life. Personal networks that contain routines and structures tend to provided individuals with a sense of purpose providing meaning and identity.

• Specific pursuits and activities that generate well-being, balancing physical and mental health needs. Medication management motivated some people to do physical exercise to manage side effects; others identified cultural interests to stay healthy and ‘shed the illness’.

• Practitioners. On the whole, formal supports were viewed positively, and across all network types health care that was provided by skilled and knowledgeable empathetic professionals was highly valued and a necessary part of managing well-being.

We have emphasised individual agency23 within the analysis: people with SMI driving changes in their lives and making decisions across each network type. This can be emotionally draining, requiring support from their social support network and clinical expertise. We found the network categorisation process was helpful in guiding individual reflection on connectedness and well-being. What emerges as important was not the relative benefits of the three network types (formal and sparse, family and stable, diverse and active) but the conversation about why the network looks that way, what were the key influences in it and where someone wanted to go next. Do people want to develop more varied connections or not? Is stability the goal, requiring reduction of some current network influences?

We also identified potential for network mapping as a therapeutic intervention tool:

• The network-orientated approach was about unlocking the potential of a network, rather than seeking out an unrealistic ideal. Personal recovery takes many forms.

• Network mapping was not abstract but it was applied and focused on what was important for well-being in people’s lives.

• The tool itself was a useful way to open up conversations about meaning and identity. The visual dimension of mapping helped people chart and reflect on their connections.

• It was also a way of exploring agency and connectedness past, present and future; how people have got where they are and where they are headed.

Recovery and connectedness in secondary services

The way in which an organisation approached working with people with SMI depended on its role and responsibilities in provision of specific supports for this group. We found, unsurprisingly, that secondary care providers described their role in a very different way from other organisations; there was a strong acknowledgement that the recovery approach, promoting equality and empowerment, had influenced the nature of relationships between organisations and people living with SMI in the last few years:

Over the last couple of years the recovery model also been embedded in assessments routinely carried out by mental health teams on the London site, our main agenda right now is to a recovery model of illness rather than a treatment model of mental illness.

SL114, local authority adult social care

Where recovery was emphasised, the importance of developing positive, goal-based relationships was stressed, within a person-centred biopsychosocial framework, rather than through the filters of a diagnostic, medicalised model of treatment. There was an emphasis on aiming to ‘see it through the eyes of the individual more effectively’ (SSW202, NHS). An important aspect of this had been the desire to break down traditional client and organisational relationships with ‘professionals on tap not on top’ (SD200, NHS). However, while aspects of this relationship dynamic may have evolved, there was also an acknowledgement that:

It is difficult, it is a fine balancing act in the power relationship, and service user is in a sense a little bit disadvantaged in that. But we are trying to engage with service users as much as we can; we want their opinion, we want the carer’s opinion.

SL117, NHS

Recovery-focused care was not easy to deliver. Recovery, when cited, was viewed as an overarching philosophy in secondary care services shaping organisation priorities, which caused some tensions between staff and service users. Interviews in both sites identified ongoing challenges in changing deep rooted organisational cultures and practices; there was still ‘a cohort of practitioners who promote dependent rather than independent relationships’ (SL111, NHS). There was no mention of forging interdependency within relationships between service users and the organisations there to support them.

Leadership teams in secondary care, on the whole, demonstrated awareness of the core elements of delivering recovery-focused practices, acknowledging also that more engagement was required, as people with SMI have often been left with the impression ‘they were not really listened to’ (SL100, NHS). The data were missing clear examples of how this could be achieved.

But I don’t think hand on heart that we could say that the level of engagement is where we would want it to be and that service users and people that use our services and their families and carers feel that they have as much influence around the direction of travel as they would like to see.

SW202, NHS

Organisations were aware staff needed training and support to change established practices. Also absent in our data was any discussion of how these tensions could be managed collaboratively. Such tensions were identified by leaders as individual organisational challenges, conveying a sense of silo working. There was an identified lack of interorganisational collaboration to address staff or service user resistance to change.

Another notable ‘silence’ in the data was statutory and third-sector relationships with primary care with regard to policy orientation and recovery. Our stakeholder interview with a London GP lead talked about the interface between secondary care services and primary care being patient centred in approach, but the major policy agenda of recovery was not touched on.

I guess [our role] it’s to provide services for patients that are appropriate to their needs and those that take into account both what they would like but also what is seen as being useful or right for them according to their diagnosis, so something that is seen as being evidence based, but there needs to be within that a consideration of what the patient would like as well.

SL101, NHS

In terms of social inclusion for people with SMI, primary care identified another practice void:

I know that’s been happening but I struggle to see on the ground what has actually happened. I don’t see an awful lot of change. Again it’s an area that always comes up, to do with how we do that. I haven’t seen a lot [here].

SL101, NHS

These are two illustrations of possible disconnect between mental health policy and practice on the ground from the perspective of primary care. These indications of disconnection were reinforced by the lack of use, by the interviewee, of terms associated with recovery, such as empowerment, self-determination, self-management and meaning in life.

Secondary mental health care providers, both NHS and third sector, were becoming more aware of the positive aspects of social networks for people with SMI, and there was an increasing acknowledgment that services should play a part in helping individuals to develop them, ‘to link people into networks of support and to work with them’ (SL114, local authority). A socially orientated outlook on the part of service providers was consistent with the recovery model ethos. However, there was a perception expressed in some interviews that this represented an area where people with SMI would potentially require enhanced support from organisations because of personal difficulties in sustaining personal relationships: ‘what I do see has been important for a lot of people with mental health problems is maybe they’ve lost some of that, haven’t got that wider network . . .’ (SSW200, NHS). One challenge for leaders of services, and thus the services themselves, was to see beyond generalised stereotypes to actively deliver person-centred services, working with people to overcome barriers and produce positive outcomes, hence reflecting the heterogeneity of the populations they were seeking to support. It was a challenge they were looking to meet to both aid the individual but also relieve pressure on services:

. . . the Assessment Service will look at social networks of people and what support we can tap into rather than us providing the support, so we do that all the time; the same with the Crisis Team: they will map the social network of somebody and use all the possible support that there is out there.

SL117, NHS

Secondary care organisations felt they had a role in helping people with SMI reflect on their own strengths as a means of building up their network of connections, either by embracing new forms of activity or by re-engaging with previous hobbies.

We encourage people to look at voluntary work quite a lot, because of the various benefits it can give you in terms of potentially doing something they enjoy that might give them skills towards future employment.

SSW212, third sector

So say you discover in assessment this person used to sing in a choir but hasn’t done so for 10 years – could they be encouraged to do it again or offered something? Encouraging people to go back to things they might have given up.

SL109, local authority

Social support and stigma: surviving in the context of change

Longstanding conceptual drivers frame the secondary care organisation vision for working with social networks: social support, social inclusion and reducing stigma are key themes currently. A strong linkage for organisations involves getting people ready for work: building confidence, building skills and using resources within networks to support employment-related activities. Encouraging people experiencing SMI to do something productive and take an active role in their own recovery has become an increased expectation on the part of service providers, and requires partnership working between stakeholder organisations. The personal benefits for people with SMI being engaged in productive activity, such as employment, were emphasised in interviews:

Well, I think that we as a service would feel strongly that what we do is about promoting well-being and a person’s sense of purpose and ability to contribute to be an active part of society and to be of value.

SL115, third sector

For some people it might be just a short-term goal in managing their time through something meaningful so it might be I’m linking them in for some walks for health or a local college course to develop their IT [information technology] skills or even something like a psycho educational course like a mind over mood course, to support their health and well-being and their wellness. For others it might be that they’re being made redundant and they’re looking at retraining, so there would be a piece of work done around CV [curriculum vitae], confidence-building, looking at perhaps doing a strengths analysis, have a look at their developmental needs and working on an occupational plan.

SSW209, third sector

In the London site, employment advisers had recently become embedded in some mental health teams. This had proved challenging, as there was initial resistance from clinical members of staff who ‘didn’t seem to understand why there were employment advisors within a mental health team’ (SL115, third sector). The social aspects of supporting SMI can get overlooked in the health-care system even when these dimensions are mentioned in NICE evidence-based advice on the care of adults with psychosis and schizophrenia. 39

Interviews with community leaders indicated that, outside health and social care, encouragement to engage in ‘ordinary’ activities was commonplace and the focus of their work. Engaging people with mainstream activities and providers has been established mental health policy for the past 10 years, built on principles of social inclusion. 234 The interviews captured the views of a number of mainstream community resource leaders, but they were mostly hesitant and unsure about working with people with SMI: ‘we are not at a stage where serious mental health is high on our agenda’ (SL103, other). However, an education programme highlighted elements of good practice that providers of mainstream adult education services can employ in order to ease the transition for people with SMI, to provide ongoing ‘confidence-building and reassurance’ (SL112, other).

One of the main challenges for community-based organisations was assessment, in that assessing the capabilities of people with SMI could often be difficult. One participant remarked that people with mental illness can overpresent themselves in a positive way, and ‘tell a good story’ which may be at odds with what is actually suitable or sustainable for them at the time (SSW207, other). This emphasises the distance services and systems have to close in order to build trust, and better support people with mental health problems; telling a ‘good story’ is rarely just about illness and symptoms, but is often about trust and engagement in a system of care. In addition, people with SMI may lack ‘the emotional strength or confidence’ to deal with knockbacks, ‘so that could be a negative if things don’t work out the way they should’ (SL104, third sector). A key challenge for services was to turn negatives into positives and approach ‘things that don’t work out’ as opportunities to try new strategies. There appeared to be a lack of confidence within organisations in supporting people with SMI in community resources outside health and social care. This was a challenge for statutory services, who required the community to pick up support ‘for free’ that was previously provided by mental health services. It also represents a challenge for individuals with SMI, as some interpret community organisational resistance as part of the wider stigma problems that they face in living with mental illness.

Mental health organisations spoke about a new deal with service users and families but reflected that poor service join-up could leave individuals feeling unsupported and isolated. There was an acknowledgement that recovery work should promote positive relationships between service users and organisations, but it was also underlined that this should not entail a ‘lifelong association’ (SSW208, third sector). Involvement should be limited in nature: ‘The role of services is to kind of come alongside and to try and interfere with people’s lives as little as possible whilst enabling them to get on with life and living’ (SSW210, third sector). One lead reflected: ‘we can help you help yourself and I think that has to be a message, maybe a hard one, to give’ (SL111, NHS).

Organisations acknowledged that the challenges people with SMI face include public stigma and variable access to social support. Stigma impacts through structural discrimination, for example budget cuts, as well as public attitudes or self-stigma beliefs and behaviour. There were several points raised over the role of an organisation with regard to building connections. First, concerns were shared across both sites that mental health provision has been particularly vulnerable to health spending savings. The capacity of practitioners to undertake network development activities was under threat.

Well I do know that in this borough a lot of the support services for different groups have had to close because of cuts. In terms of providing to different groups, this service offers I think 16 different languages, so we need to be mindful of what is out there.

SL107, other

For my other role I guess it’s a lot to do with service changes and trying to improve services to best care for people, and often it’s having to do it with money in mind. At the moment we are trying to make a lot of cuts so it’s trying to be efficient with what you’re providing and a lot of it comes down to services being reduced but how can we make it better – so it’s difficult.

SL101, NHS

Second, the sector was also worried about other government policies impacting negatively on those with SMI, such as changes to welfare benefits including housing and the work programme. The concern was about both pressure on services and impact on individual networks.

I am quite worried that [the work programme] will push people into seeking more intensive mental health support.

SSW211, third sector

However, these pressures were also regarded as potential opportunities; libraries diversifying and providing job seekers with support was one example that we were given of this. Libraries and other council services have increasingly been mainstreaming the support that was offered to all groups of disabled people, including those with SMI.

It’s not just access to books, we are the friendly face of the council and we get people who come in who are unsure about bidding for social housing for example. They used to produce a magazine to view the properties but now it’s only online so obviously if people are coming in and having difficulties using that system we will help them with that. If people are using our terminals for council business we don’t charge them and we understand we know they might need more than half an hour. Our [name] library is partnered with Work Zone, an employment agency that do training especially in retail – they’re near the [name] Centre – and help people get jobs there and that’s open to everybody.

SL108, other

Third, the networks of people with SMI were seen as vulnerable to change in line with the introduction of assertive discharge from CMHTs, resulting in individuals increasingly being moved from the care of a consultant psychiatrist to the GP.

I think the concern I always had with that is that whenever people aren’t under the CMHT they also seem to lose track of other services . . . So there’s a little bit of a gap then, so we agree that this person doesn’t need to be under the CMHT but they probably could do with some of the extra support that they were getting.

SL101, NHS

Although organisations were clear about the value of working with users of services in a network-orientated way, in a period of service reorganisation and public finance reviews, it was unclear who would, or should, take ultimate responsibility for this care strategy. One possibility was for families to become more involved, which would require organisational change in the way families were valued and supported by mental health services. The most common theme we identified across interviews was caution because of their potential negative influence on well-being and networks.

If someone [family member] is supportive and comes along to appointments or encourages them to attend appointments, that’s huge, if someone doesn’t want to let their family know that’s a big issue in itself.

SL101, NHS

The general principle is, you know, that people coexist best within families and as long as we recognise where harm is and we manage that I can’t see, I don’t think that is an issue.

SSW202, NHS

The organisation leads’ perception of family illustrated key dilemmas services and their staff face in weighing up their active role in connection-building or in encouraging that role in others.

An uncertain climate

The interviews, conducted in summer 2011, highlighted a number of key changes with regard to the relationship between commissioners and providers of services. The context was seen as extremely challenging across all sectors: the expectation from commissioners, and managers of contracts, was increasingly that organisations delivering mental health services will have to provide more for less, and ‘become more enterprising’ (SL102, third sector).

The biggest challenge is resources because they will be shrinking of course, and the whole big society approach and what it means, it is many tentacled, the fragmentation of what that means is just enormous. However, it seems to me it’s a real challenge to expect the third sector to be delivering services for less.

SL105, third sector

There had been a move away from long-term contracts, towards flexibility in approach. The backdrop to interviews was a time of great change, when services and organisations were in the process of being merged or decommissioned, with the function of public health shifting to the local authority, while commissioning was being restructured from primary care trusts (PCTs) to CCGs. The shift to personalised budgets was seen to have potentially huge implications for providers (especially in the third sector); ultimately when personalisation was fully realised it could mean that ‘your service will live or die by how popular it actually is’ (SL102, third sector). It would result in a demand-driven system based on choice and preferences of people with SMI rather than a supply-driven system based on what services believed was necessary, or wanted to provide. Mental health payment by results (PBR) pilots were also being established, though not in the study sites, so organisations were preparing themselves for further reforms impacting on the future finances of services.

In this demanding change context, strategic multiagency collaboration was seen as crucial to implementing improvements in mental health care. Working strategically with other organisations was considered to be essential across interviews; one participant stressed that collaborating in partnership with others helped to engender better outcomes for people with SMI: ‘we have a central role but not necessarily the most significant role for people in terms of helping them recover and live as good a life as they can’ (SSW200, NHS). The aim, from a planning and contracting perspective, was to provide individuals with a ‘seamless service’ and draw on expertise across sectors: primary, secondary and third sector (SSW200, NHS). Looking into the future, a key issue surrounding treatment and support for people with SMI was: who will take on the leadership role and co-ordinate such an interconnected care solution?

The third sector

A key stakeholder in the mental health landscape was the third-sector specialist, spanning mental health provision and generic services such as housing. There was an acknowledgement in interviews that the third sector must become more enterprising and innovative. Key factors engendering effective collaborative relationships, involving third-sector providers, were examined. One aspect that came through strongly was the need for close relationships between ‘the key individual personalities in the organisations that are involved’ (SSW08, third sector). It was also emphasised that partnerships should be established on the basis of organisations being aware of their strengths and weaknesses:

Where each organisation is aware of their strengths and weaknesses. So we work with [name of organisation] who are a large welfare to work provider, they are generic, non-mental health specialists, private sector, they rely on us because we know how to work with people with mental health problems, and have provided them with specialist support to build confidence and work on their goals, and so on.

SL102, third sector

The climate of increased competition and lack of awareness of other potential collaborative partners were identified as the main barriers preventing co-operative working being more fully realised in the third sector:

There’s an element of competition inevitably as well, business to be found.

SSW08, third sector

Well I think commissioning in the third sector has caused all this competition sometimes, so people are almost set to compete for outputs sometimes as opposed to working together, so that’s one barrier.

SL102, third sector

Enablers and barriers to developing organisational partnerships

A number of enablers to effective partnerships were identified, underlining the crucial importance of effective networking between organisations. An emphasis was placed on shared goals and visions across providers: ‘recovery should be at the heart of that, but that should be mirrored in the relationships that we have with our partner organisations as well’ (SSW202, NHS). Strategic partnership working was argued to be the way forward, underpinned by joint agreements if possible. Coming together collectively to achieve better outcomes for people experiencing SMI was expressed as the motivating goal, rather than individual practitioners, or organisational entities, seeking to monopolise credit:

I want us to work together more with people in groups so that we have a collective ability to heal people, rather than me saying I did it because I was a psychiatrist, so it was my psychological intervention. No it is the collective, so they might do the psychological, they might do the medication or we come in with the social, so that is why the psychosocial thing becomes really important.

SSW205, NHS

The personal aspects which engender effective organisational partnerships were also emphasised across interviews. The personal touch of organisations engaging together ‘face to face’ was viewed as a crucial element of relationship-building, fostering trust and mutual respect:

I think face-to-face relationships to be honest. Yeah, you can . . . you know there is increasingly a move in the local authority to brokerage, to call in off contracts off frameworks and having intermediaries manage that, but to be honest with you I think unless you sit down with agencies and have face-to-face conversations, networking, all of those personal ways in which, that is you know basic to any organisation and it is the nature of the working relationship that is fundamental really.

SSW201, NHS

The constantly shifting context was described as a huge impediment to partnership working. Local reorganisation, especially in London, was argued to render it more challenging for organisations to work together. Although working together and pooling resources were seen as desirable, there had, of late, been a climate of intensifying competition between third-sector providers, which made this more difficult to achieve. Lack of communication was also identified as a major barrier across interviews.

As to other barriers to joint working, I don’t know, but it’s not ideal at the moment, and people are forever doing, not mapping exercises like you’re doing, but mapping what services are available in the borough, and mapping someone’s journey through it all, and how we should all have each other’s phone numbers and talk to each other more, and we don’t, we don’t do it enough.

SL102, third sector

The difficulty of finding out what other services were actually out there, and being able to map available resources, was identified as an impediment to joined-up working. Overall, interviewees indicated that the ideal scenario for effective organisational partnerships involved co-operation, strategic working together, and shared visions and goals, along with effective communication. In stark contrast to this the reality was described as one of intensifying competition with a lack of communication or awareness between organisations, together with an unwillingness of some organisations to change and adapt. The overall picture across interviews was one where effective partnerships were needed more than ever, but in many ways were becoming more difficult to achieve.

Developing networks

Through the interviews two development mechanisms emerged to support network enhancement that all groups adopted within the boundaries of their current role:

• referral, signposting, facilitating and linkage

• encouragement and empowerment

We observed that some practitioners, particularly GPs, were cautious in making any network enhancement recommendations. There was a dilemma between giving information that opens up opportunities for individuals and making suggestions that turn out to be negative or harmful. This was further evidence of GP nervousness about being the prime responsible clinician for people with SMI.

I suppose, as a GP, it’s quite hard because you can signpost them to organisations that you might think would be helpful for them, but you can’t always guarantee that (a) they’re going to take up the suggestions, or (b) that it is going to be helpful.

IDSW83, GP

Making suggestions relies on having knowledge and information. The GP group in particular lack detailed knowledge of local resources for people with SMI. ‘I don’t link them into these services because I suppose I’m not really aware of them’ (IDL80, GP). The most common signposting for GPs was via a referral letter if required, or suggestions for the person to reflect on and act on themselves. GPs spoke about often not finding out if their suggestions were useful, and recommended that feedback, individual or collective, would be a positive influence and practice aid.

It would work for me better if I really have the number of patients, the majority of my patients that come in and say to me, look, doctor, I really love that; it’s really helping me.

IDL75, GP

In contrast, care co-ordinators’ and third-sector staff’s core job role is signposting. They reported being comfortable with making suggestions and ‘referring on’. Throughput of clients within services was described as a key driving force, working towards goals that will lead to onward referral out of secondary mental health services. Staff talked about using internet searches with clients to look for information and plan next steps for taking ideas forward.

I look on the internet; I search a lot of things to find classes, or activities, or gyms or . . .

IDL60, social worker

Psychiatrists tended to focus on encouraging and empowering clients to take up suggestions that they or other team members make. The approach taken also depended on a range of factors including how well a person was coping, their relationship with services, illness presentation and resource constraints.

[Some] sort of things would maybe sort of lead me to be a bit more directive, you know, and making suggestions to them, you know, for them to consider. And people that are more resourceful, further on in their recovery, it may be less important and may be more facilitating, you know?

IDSW50, psychiatrist

All groups were concerned about empowering individuals to make their own decisions, preventing the creation of dependency on them personally or the health-care system. All were mindful that individuals’ own agency was central to managing SMI.

We have to be aware that, you know, about creating dependency. So we’re always seeking to, kind of, make sure that the client is, whilst getting the best of the service, aren’t depending on us as, you know, for transport or friendships.

IDSW47, social worker

When considering whose role it was to develop CHNs, the practitioner groups all emphasised the importance of the individual service user themselves. Regardless of current health status, a large part of network development follows the interests of an individual and how engaged they seek to be with others and local resources. Interviewees gave numerous examples of how they were doing this in their own practice.

My role really is to get them, to try to build along with them, you know, the networks which is appropriate for them. But their own role really is to feel and to engage with the networks, which we agree on.

IDL62, support worker

So it’s a collaboration and co-operation of all the services involved in the care of the individual and of course with the individual taking more interest in what they would like and identify as their need. But I don’t think it’s a single person’s role, I think it’s the role of all the people involved in the care.

IDL84, psychiatrist

To my mind, you know, it’s all very well for us to go in and support people, but unless we’ve got an eye on how they’re going to live their lives without that support at the end of our time with them, then we’re failing.

IDSW42, floating support

Roles within the network

For care co-ordinators and third-sector staff, the time they spent with people on their case load included detailed work to extend community connections, tackling risks of social isolation and lack of meaningful activities. This involved direct action within networks. We understood this role as:

• being involved directly in the network – supporting and accompanying people to make new connections

• building skills to support network development (particularly self-confidence and self-esteem, anger management, self-care and managing anxiety), addressing both the impacts of living with mental illness and aspects of illness or symptom management.

How staff understood that role varied from walking alongside someone to being in the centre of the network and controlling it. The focus was always on empowerment, not ‘taking’ a person to an appointment but ‘accompanying’ them; however, the style of doing this varied in practice.

I think our role is, like, if you look at a spider web, we are, like, in the middle and then our responsibility is support them, linking to all the rest of the networks that are available . . . Because they will be quite, maybe, scared to go out there and, you know, join the community or even afraid of being out [of hospital], you know, stigmatised, if you like. So we have to be there constantly, you know, to reassure them. And we also have to make sure they make those small steps. Sometimes it doesn’t happen and then you try again and then eventually they make it.

IDL49, support worker

Make suggestions. Can give practical advice and health advice or dietary advice, you know, we will very much walk alongside someone, you know, I haven’t got the magical key for someone to recover – it’s what their recovery means to them.

IDSW67, team leader

A key role for care co-ordinators was skill-building and developing capacity for the individual to do more network development on their own. They were also likely to refer on to support workers to ‘do’ direct action network functions.

It’s about meaningful things. It’s about occupation things. And it’s about giving people some self-belief that they can reach out and try things that they might not have tried for a while, or they might not have even considered in the past.

IDSW81, team leader

We found that connection-building guided by practitioners mostly had an activity focus. What was notably absent in the practitioner interviews from all groups was work acknowledging the value of friendship within networks as a resource and source of support. In our data, friendship was referred to only as a gap in people’s lives, not as a resource.

Can I just say sadly, patients with mental illness, you know, tend to have less friends, they don’t have support services and things and they may have initially have had it but then people get on with their own lives, they can’t cope with it, they find it too stressful, they can’t bring people towards them – they don’t naturally have the resources to do it either.

IDL80, GP

And so with the nature of mental illness as well there’s like a lot of our clients have quite complex relationships with family and friends, and it’s not actually always in their best interest to try and nurture that network.

IDL63, social worker

However, in more general terms, network development from those most used to working in this way was seen as a positive aid reducing pressure on the practitioner in particular.

But, you know, I mean generally speaking, you know, I always try and look at what resources people have in terms of social networks, and look at how they can be improved, if needed, and you know, how they can build more networks. It is definitely really important, and actually as a worker as well, it kind of eases the responsibility on your shoulders, because, you know, you’re kind of helping people branch out and, you know, get their resources in different areas, rather than just sort of depending on you as a worker.

IDL63, social worker

Through our analysis, the team created a three-level model of practitioner network involvement. This could provide the basis for further research assessing role of practitioners within networks and the development of best-practice guidance:

• network information point – light touch intervention communicated within a consultation session (professional expertise, information, ideas) and also providing leadership for colleagues, advocating for the approach and providing guidance

• network resource that empowers and build skills – therapeutic interventions in groups or one-to-one sessions (so the individual can take more control)

• network resource that co-produces outcomes with the client – embedded resource within person’s network as a network member themselves (accompany and take part in activities).

Each of these roles could place the practitioner ‘on the network map’ as a valued contact and an important resource within a personal network for someone with SMI.

How did practitioners support people with severe mental illness to use their network effectively?

Reflecting on the role of practitioners there was, of course, no one-size-fits-all response as individual service users require varying levels of support, which changes over time. However, there was a role for practitioners in showing that networks were valuable, were part of recovery journeys and were a way of bringing ‘the social’ centre stage in treatment and care dialogues and monitoring processes. Social care workers can help clients build relationships and strengthen links to local communities. 238 As other components of this work show, the practitioner role was particularly influenced by the resources within people’s social networks. Practitioner interviews highlighted that staff view family as an important resource, with safeguards,239 but there was a lack of focus on friends as a recovery resource. This study has not investigated what a practitioner-led CHN network intervention would consist of, but these were under development,125 and relationship-building including friendship would be a core component. At the moment most practitioner strategies involve skill-building or connections to meaningful activities including employment; there appears to be less emphasis on addressing difficult relationships or building social connections.

Poor integrated care and support for people with SMI has recently been highlighted by the Schizophrenia Commission. 240 To facilitate effective working to enhance networks we identify a need for more holistic engagement and planning with individuals. Strategies are required for how practitioners can empower, link, facilitate, support and guide network development. Decisions were required on who was best placed to engage in network enhancement strategies; it was unlikely to be the time-poor GP unless a networking intervention within the 10-minute consultation slot could be created.

How did organisations work together to better support people with severe mental illness?

We found evidence that there were four service ‘silos’ in operation: secondary mental health services, third-sector provision, community organisations (including housing, education and leisure providers) and primary care. There was interagency communication but strategies to improve this were required; communication between primary care and the third sector was very limited and between primary and secondary care often not effective. Community resource information was vital for facilitating network development. Access to resources, through signposting, recommendations or actual referrals, currently rests mostly in practitioner’s own knowledge bank of relevant opportunities.

Several issues arise from our analysis of relevance to the support and development of networks for individuals with SMI:

• Finding a solution for the sharing of up-to-date local network resource information is an important component of interagency support for network development.

• When individual care is provided by different practitioners and teams (shared assessments, individual goals, records, liaison opportunities), ongoing co-ordination is needed.

• Joint service planning and operational co-ordination between different parts of mental health provision are needed. This is particularly critical for those discharged back to primary care, given the almost total absence of interaction between primary care and the third sector.

Barriers and enablers to network development

The barriers and enablers that were recognised by organisations in Chapter 5 (see Figure 19) align with those emerging from practitioner interviews (Figure 20).

FIGURE 20.

Practitioner views on network development: barriers and enablers.

Practitioners recognised that networks were important in recovery for people with SMI but reported system-level barriers, including workloads, service reorganisation and bureaucracy of paperwork, that restrict network enhancement. We conclude that even those practitioners who were personally inspired and motivated to focus on the personal networks of people with SMI struggle to do so. There were so many system-level barriers that individual practitioner agency was squashed. We heard from practitioners how, to protect their own mental health and workloads, they could not get too involved in network development, or those were the activities they did in addition to the core role. It was acknowledged that this absence sends an unhelpful message to mental health service users: if the only attention they received from services focused on medication and relapse prevention there was a risk that individuals would not view the social aspects of their lives as important for recovery.

Structural background to the South West site

The site was located in a predominantly rural county, with mostly small communities and limited public transport links. The area also included coastal settlements and one small city. We purposively recruited from 9 of the 109 GP surgeries and 4 secondary care teams; thus the site was one small area within a larger county. Recruitment at this site was from predominantly rural communities. The SW was chosen as a contrast to an inner London borough in order to explore how personal networks and organisational collaborations operated in two geographically and demographically different areas. The generalisability of the findings, in drawing on only two locations, is limited, but such contrasting sites helped provide an insight into if, and how, location impacted on network type and meaning.

The mental health prevalence figures for the whole county from the area’s 2012 Joint Strategic Needs Assessment (JSNA), using data from 2010, estimated 70,000 people had a common mental health problem and 1,700 had a psychotic disorder. The estimated prevalence of psychosis was 0.8 per 1000, the same as the average for the whole of England. A fifth of residents in this county lived in isolated rural communities and a further fifth lived in small or medium-sized towns. The report noted that rurality was a key challenge in providing services, particularly the availability and accessibility of public transport. The SW site had one mental health NHS trust covering the area. At the time of study data collection there were three PCTs in the county; under the new structure there are two CCGs.

How were personal networks composed in the South West?

The spread of network type significantly differed by site (χ²= 21.69, df = 2, p < 0.001) and almost half of SW networks were family and stable. Table 33 shows that network features also differed across network types in the SW site: family and stable networks had the most social ties. All network types reported high levels of time at home.

Table 34 summarises significant network differences from London. In the SW, informal networks were dense, with more longstanding place connections, carers were more common, more waking time was spent at home, connection to places outside the home were less frequent, networks had fewer wider contacts, fewer activities were done with others and there was less engagement in volunteering, training or education or study. GPs featured more prominently, as did physical activities and green and outdoor spaces; however, mental health settings were less prominent.

Site was a significant variable in our regression model, on RGUK, and participants in the SW had significantly higher access to social capital, with mean scores of 15.3 compared with 13.1 (t = 2.07, df = 145.69, p = 0.04). We found this resource to be primarily accessed through family; therefore the increased family dominance in the SW site most likely explains this finding. Yet participants in the SW obtained significantly more social capital resources from acquaintances than participants in London, 16.8% compared with 7.7% of total resources accessed (t = 3.53, df = 146, p < 0.01). It seems that acquaintance contacts involved more resource exchange in the SW than in London, possibly related to longstanding connections within the community.

Networks were also likely to be shaped by differences in participant characteristics by site. As summarised in Table 35, bipolar disorder/manic episodes were more common in the SW site (but not significantly so), ethnicity was significantly more homogeneous and significantly more participants were managed in primary care only. This site was more advanced in a policy of assertive discharge to primary care; the SMI population reflects overall lower ethnic diversity in this county (3.4% BME groups and 3% other white) than in the London borough site (22% BME groups and 20% other white), as detailed in the JSNA reports; and we acknowledge a sample bias in the SW site recruitment towards people with bipolar disorder recruited through primary care. Significantly fewer in the SW were unemployed and, although the difference was not significant, more were on long-term sickness or disability benefit; this may change with future welfare reforms. SW geography may have shaped these networks, with limited public transport links and fewer available opportunities to volunteer or engage in training. The rurality of this site was also reflected in the increased use of green and outdoor spaces, which can be beneficial for well-being.

We have explored the data to assess the impact of diagnosis on the SW network types. In the SW, bipolar networks were more likely to be family and stable; in London, they were more likely to be diverse and active. In the SW, bipolar networks were more likely to be supported in primary care and feature children than bipolar groups in London, but this was also the case for all diagnosis groups. There was no difference between SW and London bipolar groups in terms of HRG scores, RGUK scores or numbers of relationship types.

A detailed consideration of two networks

In this section we provide two illustrative examples, chosen to explore prominent features of SW networks, but they are neither fully representative nor meant as caricatures. In order to protect anonymity, we have changed key features. We conclude each example by considering how each of the participant’s personal network related to key dimensions of interest in the study: social support, recovery, stigma and social inclusion.

South West personal network key characteristics: Matt

The experience of the first participant, Matt (SUSW01), illustrated many of the notable features of personal networks in the SW, in that his was a family-orientated, stable network demonstrating the importance of outdoor spaces. All quotes in this section were taken from Matt’s in-depth interview, and Figure 21 provides a pictorial summary used in the analysis. Matt was a self-employed white British man in his forties, diagnosed with bipolar disorder, living with his wife and two children in a rural town. In his initial interview he described his network as ‘positive’ and ‘supportive’, providing him with ‘stability’.

FIGURE 21.

Analysis using within-case process for Matt. Key: green, past; pink, strategies; yellow, future; orange, activities; purple, building blocks; light blue, change in networks; blue circles, people; red circles, other information, usually statements about health or beliefs; red jagged edges, particularly stressful things in life including voices.

He expressed similar sentiments in his follow-up interview, saying that he was ‘fairly satisfied’ with his network, ‘my friends and family, um, and in terms of the mental health kind of provision’. Matt’s satisfaction with his network at the time of the interview was in stark contrast to past experiences. As Figure 21 shows, his experience of hospitalisation and unmanageable periods of illness caused great distress; subsequently, however, he developed active strategies that have helped him build a stable network with future goals. There was also evidence of connection-building as well as stability.

Family was a central part of his life (as shown in Figure 22) and these relationships suffered in the past when he was hospitalised: ‘the family relationship breaks down and, you know, everything goes completely awry.’ He reflected a contributing factor to his last hospitalisation was irregular contact with mental health workers and particularly his CPN: ‘I became very, very ill, it was too late, intervention happened too late’; as a consequence, he reflected, the recovery period took a long time.

FIGURE 22.

Network map for Matt. CBT, cognitive–behavioural therapy. See Figure 10 for key.

The period up to and including hospitalisation for people with SMI has been shown in Chapter 3 to be a potential network rupture point. Learning from his own experience, Matt, in collaboration with practitioners, worked on creating active strategies to manage potential relapse: ‘We kind of focus mainly on the triggers and assessing . . . self-assessing my condition’. A key part of this involved looking for ‘trigger signs’ of becoming high again to facilitate early intervention. His CPN was identified as an important contact, with whom he had developed a Wellness Recovery Action Plan (WRAP). In depressive phases of his illness after being high, Matt reflected that the mindfulness approaches were helpful. An important aspect that came across was that Matt’s family were more connected to his ongoing treatment and support than in the past. He said that if he became ill again they would be likely to contact his GP or CPN at an ‘earlier stage’, and indeed that they had ‘more of an opportunity to do that now than they did before.’

With respect to Matt’s ongoing sense of well-being, it was evident that the centrality of home life and spending time with his ‘tight knit’ family were crucial. He stressed the importance of keeping up routines with his family both within the home, such a cooking and quality time, and also in activities away from the home, such as walking the family dog and growing vegetables at a local allotment. Spending time boating on a local river was also identified as an important part of family life.

Matt was self-employed and embodied an active outlook to managing his health and well-being, describing himself as a ‘practical, hands on’ person, ‘always kind of involved in some kind of project one way or another.’ His approach to staying well was underlined by ‘maintaining routine on a day-to day basis’, and, as discussed above, family and practitioners contributed to his sense of meaningful structure. Part of developing a routine which supported his health involved Matt finding the right balance and not overstretching himself. In the past he had taken on too much work at times, and such ‘stressful work situations’ were identified as ‘a major trigger’ for becoming mentally unwell.

For Matt, work provided structure, income and a sense of purpose. It also shaped his identity as a skilled craftsman producing work of a high standard. However, as we have seen, work was also a regarded as a source of stress, and awareness of this within his support network provided a useful preventative strategy.

Matt reflected on his experience of living within his wider community: both the positive and negative aspects of living in a small town. He was proud of his status as a father and family person, a self-employed craftsman and a member of the local community, which, when he was well, constituted his dominant identity. Living in a ‘tight knit’ countryside community was viewed as a safe, positive environment for ‘bringing children up’; although he had lived in more built-up places in the past, he was adamant that he did not want to raise his children in a city. In terms of his own mental health he saw some positives about the local area. One aspect he touched on was familiarity: ‘you can’t walk up the high street without probably seeing half a dozen people that you know.’ Such familiarity engendered a sense understanding: ‘they [local people] are aware of my health condition’; he added, ‘they are tolerant and accepting to a greater extent.’ However, there have been times when his identity as a skilled, self-employed family man has collided with community perceptions of the mentally ill and he has felt negatively labelled because of his mental illness:

. . . there is a stigma attached to it, but over a period of time and obviously when you’re following on from an episode when I was very low it you know, you are very anxious and paranoid about what people . . . and as a result of what you’ve done and so it does take a long period of time to get around that.

Related to his experience of stigma, Matt considered his recovery not just in the personal terms of his own health and well-being, but social recovery too, as he had gone through the process of rebuilding relationships that had become strained during past ‘high’ episodes.

It’s a huge part of the recovery certainly. Initially you know I couldn’t go out of the door for months, and you know I couldn’t deal with face-to-face contact because of what had happened . . . It just takes a period of time it’s you know rebuilding you know relationships.

Overall, Matt expressed satisfaction with his network and stressed that he wanted the future to be a continuation of the present. He considered himself ‘fairly fortunate’ in that he had a network of family and friends, who complemented his proactive outlook, who ‘steered things in the right direction’; he added, ‘I think a lot of people don’t have that advantage.’

Matt valued the stability of his network, but it is important to emphasise that it was not static and stagnating, but evolving, goal driven and future orientated. Drawing on five areas identified in our literature review (see Chapter 2) that help us understand personal networks, and answer the first study research question, we applied the learning from Matt’s case study in Table 36.

TABLE 36 - Summary points from Matt’s network

Key area of literature	What we learn from Matt
Agency and strategies for staying well	An active, balanced routine comprising meaningful activities, is important for promoting well-being. What these are, and where the balance lies, should be self-directed by the individual
Social support	Practitioners have a crucial role in helping people with SMI to develop proactive, preventative strategies, linking into available family support if possible to maintain mental health
Recovery over time	Effective strategies for managing SMI are learnt over time, and people can draw on past experiences in shaping current decision-making and negotiating these preferences with others. Within family and stable networks the family are often a key part of the recovery journey, and they themselves have a family recovery pathway too
Stigma	Cohesive communities provide a measure of social support and understanding for local people with SMI; however, there can be a perceived stigma relating to public knowledge of someone’s health problems. People feel judged and labelled by their illness and the negative opinions of others at times. How people feel about community knowledge of their health problems changes over time
Social capital	Accessing local resources represented a crucial part of building networks, generating a sense of connectedness and strengthening belonging. Social capital and sense of inclusion can fluctuate, diminishing when unwell and growing during recovery

South West personal network key characteristics: Nick

The experiences of Nick (SUSW51) also illustrated the importance of family networks in the SW. Nick, however, no longer had contact with specialist mental health services and had limited access to wider community resources. He was an unemployed white British man in his fifties diagnosed with schizophrenia, who described his network as ‘positive’, one where he was ‘close to kids’, but felt that he did not ‘see friends often enough’.

Nick explained that hearing voices comprised a dominant feature of his life (as shown in Figure 23) and described most of them as negative. The voices impacted on all his relationships and caused him to experience heightened stress ‘on more or less on a day-to-day basis’, which included inducing panic attacks.

FIGURE 23.

Analysis using within-case process for Nick.

Nick lived alone and had two children, from previous relationships, with whom he was in regular and close contact (Figure 24). At the time of the interview he was proactively involved in the care of his youngest child, and talked at length about the strength of their emotional connection: ‘we play lots and lots and lots together . . . the relationship’s very close, based on quite deep feelings really.’ Nick found child care both rewarding and a source of stress. He discussed a recent experience when they went swimming together when his voices led to him having ‘something of a panic attack’ that ‘went on for several hours’ which he struggled to cope with: ‘I ended up smoking, which I don’t like doing.’

FIGURE 24.

Network map for Nick. Social network size: 13 (second-lowest quartile, mean = 19.91). RGUK score: 13 (second-lowest quartile, mean = 14.2). HRG group: lower score group. SWEMWBS score: 28 (highest quartile, mean = 22.3). See Figure 10 for key.

Ability to manage stressors often changes over time and a person’s stage of recovery; some people with SMI rely on social support, whereas others draw more on their own resources. Nick revealed that his network relationships could be a source of stress, and that he dealt with such life stresses largely alone; he did not have mental health practitioners or peer support groups to call on but developed his own active strategies. When he reflected on the network process it was apparent to him that there is a deficit of friendships in his life: ‘it didn’t seem like I knew enough people; I should have more, certainly more friends, really.’ Hence, his self-reliance was at least in part related to having a limited number of friends to call on, and so was out of necessity rather than choice.

As with numerous other participants who were over 50 in the SW, Nick’s contact with specialist services had ended. His mental condition was managed entirely in primary care, although he reported that he rarely discussed his mental health with his GP. He was, however, receiving medication which he believed was important in helping him to stay well. Mental health services had, in the past, featured in his life and he used to receive counselling, although he felt that it outlasted its usefulness.

Certainly over the latter stages I would keep going to talk to my counsellor about my favourite bands, and, um, there didn’t seem any point in staying there any more.

Nick presented himself as having a strong sense of self-identity and routine; his flat was likened to a well-ordered space from where he organised his life in advance: ‘I normally try to plan my day’. His routine for staying well included meticulous financial planning, as well as keeping a journal, which played an important part in developing self-awareness: ‘you get more in tune with the things that actually make sense to you’. Nick also found that having a ‘creative’ outlet had an ‘improving effect’, supporting his well-being. He explained that he wrote poetry, and had some aspirations to get his work published but was apprehensive about how his voices would react.

Having a balance of physical, creative and spiritual activities was important for Nick. Shortly before the interview he had made a conscious decision to cut down his level of smoking and to also incorporate more physical exercise in his routine, which he believed positively impacted on his SMI: ‘I think that being fitter is important to mental illness, the same as the eating well.’ This was complemented by meditation, which featured prominently in his life:

I think it makes you more calm. I think it calms as a mental activity I think it’s emotionally, it’s a nice experience. I think it takes you outside of simply, um, the external world, experiencing the external world and actually gives you an experience which you don’t get in a, most other ways.

However, Nick was also keenly aware that these strategies for staying well were solitary pursuits, and that his social life was restricted, again not always through choice. He revealed that he was keen to join a meditation group but that social anxiety was a barrier to trying this alone: ‘I’d certainly want to take somebody with me. I wouldn’t want to go on my own.’

This illustrated that, although there were a number of introspective, self-directed activities that Nick engaged in on a regular basis to manage his health and well-being, he ultimately wanted to expand this to encompass wider social resources and social interactions, but required support to achieve this. For individuals like Nick, who have limited amounts and sources of health resources within their personal network, and no specific mental health support beyond a GP, it is challenging to identify where any such resources might come from.

Drawing on five areas identified in our literature review (see Chapter 2) that help us understand personal networks, and answer the first study research question, we applied the learning from Nick’s case study in Table 37.

TABLE 37 - Summary of Nick’s case study

Key area of literature	What we learnt from Nick
Agency and strategies for staying well	The importance of developing a self-directed routine, comprising meaningful activities, was once again underlined; this became increasingly important, but at the same time more challenging to maintain when managing stressful voices. Level of mental distress is a strong driver in shaping appropriate strategies which change over time to compensate for the impact of mental illness
Social support	The role of service absence is highlighted. Some people with SMI need support to engage with community resources, and there is a danger that when service contact is minimised people become increasingly isolated and less likely to engage in activities that may aid recovery
Recovery over time	This demonstrates how over time, especially when support from mental health services is reduced, people with SMI may have a reduced network and find it challenging to make new connections on their own, placing them at risk of isolation, particularly if they are not employed
Stigma	Fear of engaging in new activities, partly because of social anxiety in new situations, in which stigma can also play a part, can be a major barrier for people with SMI developing their network
Social capital	People who struggle with managing their SMI and experience stressors, for example hearing voices, often also have a strong sense of personal identity, in this case their own artistic capital, which can provide potential building blocks to facilitate activity- and place-based connections that in time may lever new social capital opportunities

The commissioner and provider context

The two personal networks discussed above demonstrated contrasting levels of mental health support from services. Here we draw on data from the second round of stakeholder interviews (conducted in summer 2013) to examine how ongoing treatment, support and recovery for people with SMI in the SW site are situated within a wider provider and commissioner context. Particular attention has been paid to challenges and opportunities, and the extent to which there is a recovery-focused culture within that context.

Structural background to the London site

Our second site was an inner-city London borough. The borough’s JSNA for 2012 estimated just under 200,000 people registered with GPs. The population was characterised by a large proportion of young working-age residents, high levels of migration into and out of the borough, and ethnic and cultural diversity. There were also large health inequalities.

The study recruited from across the entire borough, working from 9 out of 29 surgeries and 3 secondary care teams. GP MH8 QOF data for 2012 showed that the prevalence of SMI within the borough was 1.2%, which is higher than the prevalence in London overall and higher than the mean prevalence in England. The male prevalence is higher (1.3%) than the female prevalence (1.0%) of SMI. The 2012 JSNA estimated there were 2,400 people living with SMI in the borough. The London site had one local authority and one mental health NHS trust, but both structures covered a larger region of London than one borough. At the time of study data collection there was one PCT and under the new structure there will be one CCG.

How were personal networks composed in London?

The most common network type in London was diverse and active, while fewer participants had family and stable networks in this site, as summarised in Table 38. We found less stability, but higher numbers of new place connections and wider contacts in social networks, such as colleagues and acquaintances.

London participants spent significantly less time at home than SW participants and connected to places more frequently (Figure 25). The community settings that London participants connected to most were workplaces, green and outdoor spaces, shops and markets, spiritual settings, social recreation settings, physical recreation settings, libraries, and family and friends’ houses.

As shown in Table 34, significantly more mental health settings, education settings and social activity or interest groups were mentioned in London. Significantly more participants also engaged in volunteering, education, training or study, and reading, and a higher number of social activities were mentioned. However, a larger number of unstructured activities were also mentioned. The mean number of physical activities was lower in London, perhaps because fewer outdoor spaces were available and costly gyms were often the only alternative. Levels of physical fitness were not significantly different by site and the number of participants who mentioned at least one physical activity did not significantly differ.

FIGURE 25.

Frequency of connection to places in community.

FIGURE 26.

Ethnicity by site.

London participants had significantly lower access to social capital and this included access to fewer domestic resources (t = 3.24, df = 145.5, p < 0.01) and personal skills (t = 3.38, df = 146.90, p < 0.01). Access to health resources did not significantly differ but in London participants accessed a significantly higher percentage of these resources from practitioners, 20.1% compared with 9.8% in the SW (t = 2.74, df = 144, p < 0.01). London networks were significantly more ethnically diverse, as shown in Figure 26. Psychotic disorder/schizophrenia and other psychoses were more common diagnoses in this site.

We found that barriers to getting the most out of networks differed by site (Figure 27). Participants in London mention mental illness, money, stigma, medication and physical health more frequently, while transport, isolation and internal factors such as lack of personal motivation and self-confidence were more frequently mentioned in the SW.

FIGURE 27.

Barriers to getting the most out of networks by site.

A detailed consideration of two networks

London personal network key characteristics: Raj

Raj (SUL07) illustrated a formal and sparse network, with long-term use of mental health services but limited access to social capital and wider social resources. Raj was an Asian man in his forties, diagnosed with schizophrenia, unemployed and living on his own. Although he identified only eight people in his network (Figure 28) in the initial interview, he described it in largely satisfactory and manageable terms as ‘reliable’, ‘safe’, with ‘zero chaos’.

FIGURE 28.

Network map for Raj. Social network size: 8 (lowest quartile, mean = 19.91). RGUK score: 13 (second-lowest quartile, mean = 14.2). HRG group: lower score group. SWEMWBS score: 23 (second-highest quartile, mean= 22.3)

In the analysis for Raj (Figure 29) the past was shown as a strong influence on the present, with very limited reference to the future. There was a sense that Raj was surrounded by a network, partly shaped by his choices and partly by circumstances; in emotional terms, past and lost relationships constituted key features.

FIGURE 29.

Analysis using within-case process for Raj.

Outside the realm of health-care professionals his network was sparse. He had children but was estranged from them when they were young, but still thought of them as ‘babies’ instead of grown up, and as a consequence found it challenging to reconnect: ‘these types of emotions are not relevant now.’

Raj had been in contact with mental health services for over 20 years, and had found the ongoing level of insight and expertise he received useful: ‘professional expert cognition’ helped to identify things ‘ordinary people don’t’. He explained that he would sometimes e-mail a mental health worker during the night; writing his feelings down was a useful way of working through his ‘racing thoughts’, and collaboratively identifying ‘triggers’.

Raj explained that mental health practitioners supported him to develop active strategies to manage his mental illness on a day-to-day basis. These individuals were a dominant feature in his network, forming five out of his eight contacts (see Figure 28). Taking more responsibility and self-care had become important aspects of Raj’s life; however, he underlined that he required ‘a bit of a nudge’ from health-care professionals to enable him to maintain a sense of routine. Much of this structure came from the Recovery Star approach, which he said empowered him to be more proactive and ‘make my life better’. He explained that he had become motivated to eat more healthily, which in turn ‘impacts on my mental health so that I don’t spiral’. This was in contrast with the past, when his life was dominated by ‘magical thinking, a way of thinking which was theological’, which was ‘the opposite of responsibility’.

If I think I’m feeling depressed or something like that then I see on the star that there’s something that I can do; there are some actions I can take. For me, that self-care this includes, for example, shaving, going to the gym, maintaining a good appearance, washing my clothes.

Support from Raj’s mental health team was complemented by an employment adviser, who helped him to engage in a ‘building-up process’. In the future he aspired to see ‘voluntary work leading to paid employment or something of that sort’.

Although Raj expressed satisfaction at the support he was receiving, and the role such support played in helping him to become more self-reliant, reflecting on his past he cautioned that ongoing contact with mental health services can promote a sense of ‘client-ism’ (dependence as opposed to independence) or even resentment on the part of the service user, an ‘us against them’ attitude.

We found that home space was an important feature across networks. Raj discussed spending much of his time at home, and that his self-esteem had improved since he took charge of his home space, which helped to create more order: ‘I’ve got the confidence, because for example my house looks nice. I’ve got clean clothes, I’m having a hot shower every day, I’m shaving . . .’ This was in stark contrast with the past, when it was ‘a complete tip at home’ and he was ‘going out and keeping a façade’.

In addition to his self-care routine, writing played an important function in structuring Raj’s home life, and provided a key emotional and artistic outlet. He explained that the process of editing was ‘the absolute key’, because ‘it gives me some insight into what I was thinking at that time, what sort of things cause, like a slippage’. He also regularly blogged online, but approached this ‘virtual type of contact’ with an element of caution: ‘you can really see how there are hate mongers out there; there are people who are really not nice’.

While at home Raj also found watching, or having in the background, certain television programmes beneficial for his mental health because of the element of familiarity: ‘it’s calming’ he reflected.

Yes, because I think what’s happening is that my brain is saying that it’s matching, everything’s matching and anything, for example, I like watching reruns, programmes like Friends, I’ve seen the episode 100 times but, for me, that’s a very . . . it happens in the background.

Living alone and having time structured by lone activities in the home were key features of Raj’s network. Because of feeling anxious about the unknown, he faced barriers in accessing unfamiliar places, as well as engaging in new activities outside the home; he needed support from practitioners to achieve these. Raj had been given help to attend a gym, which was funded through direct payments and was an important part of his self-care routine, because ‘it gives me a lot of motivation for appearance, I’m very comfortable in that environment’. He added that being encouraged to go to the gym bolstered his sense of self-confidence, providing him with a sense of purpose and motivation: ‘every morning when I wake up I’ve got something to do and there’s something I can immediately do which will make me feel good about myself’.

We also observed that Raj’s network was not necessarily place-based and he could have potentially lived in another location and yet described similar connections and strategies. It is not clear how far London as a place shaped Raj’s personal network; it was more likely to have influenced his personal strategies for maintaining well-being and illness management.

Drawing on five areas identified in our literature review (see Chapter 2) that help us understand personal networks, and answer the first study research question, we applied the learning from Raj’s case study in Table 39.

TABLE 39 - Summary of Raj’s case study

Key area of literature	What we learnt from Raj
Agency and strategies for staying well	Although people like Raj are often dependent on mental health services for formulating effective well-being strategies, there are practical ways individuals can take more responsibility. In this example, direct payments were used to access a local gym
Social support	The experience of living alone with SMI can be socially isolating and lacking in structure; services need to be mindful, as both features hinder recovery. Where social support networks are small, services can have a role helping people build up their informal support network
Recovery over time	Routines developed gradually over time such as self-care can form important building blocks for recovery. Small steps are important for skill-building and fostering self-confidence and agency
Stigma	A consequence of ongoing support from mental health services can leave people feeling disempowered and stigmatised, together with resentment, sometimes viewed as an ‘us against them’ mentality
Social capital	Home can dominate the lives of some people with SMI; when considering building social capital this can be daunting if the starting point is low. Raj had limited social contacts providing bonding capital, with no family or friend contacts mentioned

London personal network key characteristics: Áine

The experiences of the second participant, Áine (SUL76), illustrated a contrasting experience of someone diagnosed with schizophrenia. Figure 30 provides a pictorial summary used in the analysis. Áine had also experienced long-term contact with mental health services, for over 30 years, but there was greater access to wider community resources, as well as more independently formed friendships. She was a white Irish woman in her fifties who engaged in regular voluntary work, and described her network in positive terms: ‘friendly’, ‘shared interests’ and ‘loyal’.

FIGURE 30.

Analysis using within-case process for Áine.

It is important to acknowledge that, although some of the data in the report show living with schizophrenia can inhibit people’s networks (such as restricting their size), this is not always the case (Figure 31).

FIGURE 31.

Network map for Áine. Words used to describe network: ‘friendly’; ‘shared interests’; ‘loyal’. Social network size: 30 (highest quartile, mean = 19.91). RGUK score: 17 (second-highest quartile, mean = 14.2). HRG group: higher score group. SWEMWBS score: 29 (highest quartile, mean = 22.3).

Áine attended a wide variety of places in the community; the ‘most important’ was a music club she regularly attended run by one of her friends, which was also where she met her partner. She described him as a ‘very, very supportive strong person’ and he helped to drive her network development; they regularly visited art galleries and museums together. He also assisted her in understanding bus timetables, an area she found challenging, facilitating her access to public transport, which proved essential for her in terms of accessing resources in the community.

I mean, I’m still terrible on the buses, but he knows all the buses, which is mystifying for me, and I’m only just beginning to understand that. Um, but um . . . that has been very, very good, yes. It’s been a real blessing.

Áine also had a number of longstanding friends who had been ‘supportive and sympathetic’ and ‘really understand the illness’. They had been ‘loyal’ and stuck by her when she was hospitalised because of her mental health. She also developed friendships with other people with experience of SMI; when she was in hospital she met some ‘really lovely people’. In the qualitative interview the extent to which she valued the friendships in her life was underlined throughout the session.

When you get to my age . . . the important thing is that your friends are there, and they care about you, and take you for what you are.

Áine’s story showed that the value she placed on friendship was reflected in her network maps, as not only did she maintain contacts with people from the past, but her network-building activities in the present generated new valued social contacts: volunteering, language class, music, socialising, badminton and swimming.

In terms of contact with services she had a close relationship with her social worker and GP: ‘they’ve helped me to assert myself and hold my own’; they also encouraged her to take up productive activities. Áine stressed that she aimed to improve her sense of well-being by focusing on both her physical and mental health: ‘using your mind and keeping yourself fit’. This holistic approach was complemented by ‘positive thinking and doing things in stages’.

At the time of the interview Áine had not been in paid employment for over a decade but had tried a number of activities over the years; some she developed more of an affinity with than others. Such meaningful activities, particularly volunteering, provided her with routine and structure, which, alongside close personal relationships, helped to enhance her sense of positive well-being. One exception was an art course, which she attended but did not enjoy, finding it ‘really difficult’; she reflected that ‘I couldn’t do it. It was for people who had just spent all their time at home, all their time doing art’. She tried other pursuits with more success, which highlights the importance of individuals being open to trying a variety of activities to find out what works in advancing their personal recovery journey. Playing badminton was a past time that she said she benefited greatly from: ‘the sports centre is really beautiful’; she explained that it helped to keep her fit and also helped to develop a sense of ‘real confidence’. In terms of social aspects she had made some ‘great friends’ through the sport. Another important activity in her life was attending a language group, which she said ‘gives me a chance to use my brain’. Áine illustrated how engaging in new activities, and not being afraid of setbacks, helped to generate new friendships formed around shared interests.

I suppose it’s how the friendships formed, some of them, through an activity. Um, yes, I suppose that’s true, and I think it’s more interesting if, um, you . . . you have a common interest. You do things together, that is lovely.

One of the main barriers she cited in her life was tiredness: ‘I used to think . . . I’m never going to be able to get up in the morning’. A worker for a local mental health charity suggested she should try volunteering and helped to facilitate this: ‘they went with me and they were good’. This built on her key interests and skills, and having support to make the first step was the route to success. She volunteered, at the time of the interview, at a school. Initially Áine found this to be a ‘daunting’ and frightening prospect, but had developed strength of character: ‘it gives you that pull’. She also benefited socially from her volunteer work: ‘I’ve made great friends with the teachers there . . . staff there are lovely, and the kids are friendly’. An important aspect was the flexible hours: ‘there’s no way I could do a job where I had to be up in the morning every day’ and she added if you are late ‘nobody minds, if you turn up it’s a bonus’.

Áine demonstrated wide-ranging interests, was content with her network and felt fortunate to have a group of supportive friends and practitioners in her life. There was a strong sense that the desire to engage in the key activities in her life, such as a music club and volunteering work, were self-motivated. However, it was apparent too that she needed help to maintain these structures: in terms of volunteering there was a flexible arrangement where she did not have to adhere to a rigid timetable, and she was also dependent on her partner to facilitate access to public transport. There was an over-riding sense, though, that diversity of activities and friendship groups in her network provided her with a degree of protection, helping her to avoid over-reliance on any particular elements of her network, and there was stability combined with a definite feeling of optimism for the future.

Drawing on five areas identified in our literature review (see Chapter 2) that help us understand personal networks, and answer the first study research question, we applied the learning from Áine’s case study in Table 40.

TABLE 40 - Summary of Áine’s case study

Key area of literature	What we learnt from Áine
Agency and strategies for staying well	People with SMI often have a strong sense of internal motivation, but may need flexible and supportive structures to facilitate meaningful routines. Engaging in shared activities can be a positive way to develop a sense of self-identity
Social support	A balance between informal and formal networks support may be key. Everyday informal support, in Áine’s example being helped by her partner to understand bus timetables, complemented more formal service support. Practitioners should encourage people with SMI to engage in new pursuits and help reduce barriers faced in accessing mainstream community resources; provision of subsidised public transport is essential for this
Recovery over time	Engaging in a range of activities over time is essential in developing a sense of identity and meaning; people with SMI should not be afraid of trying new things or be disheartened by setbacks along the way. The result is a more durable network less vulnerable to collapse in the future
Stigma	The perceived absence of negative stigma can be important for people with SMI, as it means that the barriers to social engagement are reduced. In Áine’s example, not feeling stigmatised enabled her to engage in new activities and she did not feel demotivated, or judged, by setbacks on the way
Social capital	Developing a sense of connectedness and social capital through a diverse range of activities can lead to a more varied and durable network

The commissioner and provider context

Here we draw on data from the second round of stakeholder interviews (conducted in summer 2013) to examine how ongoing treatment, support and recovery for people with SMI were situated within a wider provider and commissioner context in one London borough. In London, the assertive discharge to primary care had not begun systematically and thus more of our study population received secondary mental health care input than was found in the SW.

Evaluative and future-forming active strategies

The extent to which people drive change in their networks, actively engaging in evaluating strategies and forming new plans, will vary from person to person and so too will the degree to which individuals themselves impose constraints on networks. One influence was personal beliefs about the medicalisation of mental illness. We found variation in our study population as to how far people rejected or accepted medical explanations and treatments such as medication: ‘I stick with my labels and don’t use the medical labels and that helps me’ (Michelle, SUSW53). This in turn impacted on the people, places and activities that Michelle chose to engage with, and those that she felt forced to include on her network map. Health and social services were regularly managing relationships with people who struggled to engage with the explanatory models they had on offer to explain SMI, describing their role as ‘giving them options’ (IDSW86, CPN). The balance in many people’s lives on how much choice and control they feel over their networks can relate to health and well-being. When ill-health forces people to engage with supports they did not want, such as medication or inviting health-care practitioners into their home regularly, it can lead to withdrawal if they are particularly concerned to keep their mental illness secret. We heard from people with SMI who actively segment their network to prevent friendship groups learning of their mental health status: ‘[my girlfriend] never looked at me in the same light’ (Kevin, SUL29). That individuals shape their own networks, to varying degrees, was found across this study and for all network types. Often, decisions to engage in new activities which might mean meeting new people and going to new places were based on individuals wanting to improve their life or develop their personal identity; this relates to Emirbayer’s evaluative and future-oriented forms of agency. 23 How far network behaviours were motivated by an explicit desire to increase well-being was harder to disentangle, but individuals gave us clear indications that, however unwell they were, agency in the personal networks of people with SMI is an important force to build upon.

Another example of future-forming strategies related to employment. In London in particular, we found many of those interviewed were on programmes to help them find paid work. Voluntary placements were a stepping stone towards employment; we noted from the network data that those who did voluntary work had fewer health resources. Practitioners might encourage voluntary work for those with fewer network resources to help them venture out into mainstream environments without the emphasis on friendship formation. In the SW case study, Matt (see Chapter 7) was careful to monitor his work, as work-related stress was a trigger for decline in his mental health. Employment builds social capital; it was a very important activity for people with SMI despite the external barriers people faced.

Non-transformational strategies

Individual agency changes over time, moment to moment; an individual’s sense of control over life events, over personal capabilities and over interaction with external structures such as the mental health system or welfare benefits programmes fluctuates. Understanding dimensions of agency in people with SMI was important for helping people set recovery goals.

Strategies which involved a more limited engagement were often based on choices around regular routines. Although we found evidence of 76% taking regular physical exercise including a short walk, we also heard from people who struggle to leave the house. Home as refuge was an important recurring theme in our analysis: for some people it was a base for the planning or enactment of structured activity; alternatively, home-based networks can represent withdrawal and loneliness. We found 66% of people interviewed spent over 50% of waking time at home; understanding the reason for this is crucial. Our analyses suggest that balance was essential: an over-reliance on unstructured time at home alone could lead to lower well-being and isolation; conversely, a strenuous focus on goal-setting was not appropriate for some people with SMI. Routines in leaving the house for shopping or visiting the library provided important structure for this group; they are examples of activities that provided bridging capital which in time may be cultivated further. For people who experienced social isolation, our evidence suggested that weaker ties or relationships, which often required limited active engagement to maintain, could greatly enhance well-being, serving as potential ‘building blocks’ for developing closer relationships. Practitioners, family members and friends should look for clues to how they can best empower individuals to become more engaged in strategies to manage their networks. We recognised that home can act as a network blockage, limiting people’s ability to move forward with other goals while home remained an obstacle; practitioners could begin with home space, as the network anchor, in any assessment of potential connection-building blocks to wrap new goal-setting around.

Availability of resources

Each individual varies in terms of the number of perceived and actual resources available to them in the form of people, places and activities. The great majority of social ties that individuals had were valued as beneficial to well-being. We found that even those contacts rated negatively can improve access to social capital. Some social networks were extremely small, with features of loneliness and social isolation, and others seemed fragile, with reliance on one key individual. Across all network types the quality of ties, rather than the number, was found to be the most beneficial. Having at least one close relationship was important because this provided access to social resources, as found in our RGUK regression model. Our population show agency in developing positive relationships and removing or limiting negative ones. Activities were used to provide meaning, identity and structure in lives, and those which were social also helped to build an individual’s social network and social capital resources. Places provided access to resources for well-being; they could deliver therapeutic benefit in their own right as pleasurable spaces to be in. We found on the whole that both our study sites had community resources of benefit to individuals with SMI, but individual barriers or lack of awareness of them reduced access.

The most resource-rich networks that participants had developed were those which were either diverse and active, with connections built through hobbies and social activities, as well as employment and volunteering, or family and stable networks where the presence of family offers the majority of resource provision. Neither network type will be ideal for everyone; the former may be difficult to maintain during illness and the latter was possible only where ties to family were available and desired. Formal and sparse networks were those where well-being and access to resources were lowest, and yet at certain points in an individual’s life they may be the only possible, or the least harmful, type of network to maintain. What will be important for practitioners will be working out which network type suits which of their clients at any point in time and the reasons for this. We acknowledge that a larger population sample and in different areas of the country might identify different network clusters and more network clusters. We suggest that a modified version of the CHN approach may offer a useful framework for identifying key features of networks, and we have produced a draft template to illustrate the type of questions that would be useful to ask to understand network indicators and network goals for an individual (Figure 33).

FIGURE 33.

Network development plan (draft).

This is an example of the type of template that could be used to guide networking conversations alongside an emotional closeness map as seen in Appendix 6.

Commissioner, organisation and practitioner level

The health and social care system for people with SMI is undergoing rapid change. Such change offers up opportunities to develop new solutions for how services support recovery but there are risks if there is no clear planning on how changes to community resources might impact on vulnerable groups. In such periods of flux, understanding the resources that individuals have and where they come from is of value, particularly if these are fragile and reliant on services. Offering values-based leadership to people with SMI is a very important role for both organisations and the practitioners who work within them; creating joint mental health strategies to guide these practices across current service silos is vital. We found evidence for a strong vision for delivering recovery-focused services in secondary mental health care and the third sector, but not in primary care. There was a lack of strategy across the service silos.

Practitioners understand and value aspects of social recovery from SMI but many, particularly GPs, and also community mental health practitioners, did not believe they had the time in their job role to focus on ‘the social’. Service systems appeared to thwart the agency of practitioners and created obstacles to person-centred outcome-focused care, even within the third sector, whose practitioners wanted to work in this way far more but found it was not currently feasible within available commissioning arrangements. This identified lack of focus on social recovery was a barrier to working using a networking approach. This study suggests that an alignment between the recovery approach and a networking model using the CHN method could be useful in the delivery of mental health care, although more work is required to adapt a research process into a clinical intervention. This constitutes an area that newly formed bodies such as Public Health England, local health and well-being boards and CCGs may all be interested in: mapping community assets and the potential within individuals to become more active managers in their recovery journey.

Individual network level

A striking issue emerging in our work was how heterogeneous people with SMI were and the implications for working with people to manage their health problems. Yet network types were found which reflected connectedness patterns that could guide the work of practitioners. We found connections to people, places and activities to be an important aspect of individual’s lives which impacted on access to resources used to support well-being, and were not related to diagnosis. Individual agency in developing and maintaining networks was uncovered but most people required some support, even if this support was received reluctantly. Many with SMI were living out the majority of their lives away from mental health services and away from other people using services. However, a minority had significant proportions of their social network formed by mental health-related contacts.

Identifying the building blocks of individual agency that can be enhanced or nurtured with the help of others was a vital aspect of recovery, particularly where individuals lacked belief in both themselves and inner resources. The change agent might be within a person, drawing on their interests, identity and goals. Change agents might also be found in family, friend, wider contact or practitioner relationship groups.

Developing the personal networks of individual’s with SMI was not identified as an organisational priority alongside management of medication and symptoms. While this remains the case it seems unlikely that this population will be able to build personal networks that make use of the full potential of inner and external resources. We note that there was potential in all networks to develop.

Network theory

This pilot study attempted to combine applied SNA to identify form and structure in networks, with qualitative data collection based on a hybrid approach of two traditions. It has merged qualitative SNA, unpicking meaning in network structures, and narrative research, exploring meaning in people’s lives, using the emotional closeness map as a ‘way in’ to the conversations that explored connectedness, well-being and recovery. Our network mapping was guided by SNA theory,24 as it offered a framework to measure connections (or a lack of connections) using name-generating and name-interpreting procedures. 76 The benefits of the approach were that networks provide a top-down view, generated by the individual, of their current social world and meaningful activities. Once mapped, these networks were reflected on and we found this perspective helped conversations about how resources in the wider community impact on, or were used to support, both mental and physical health. This approach allows for consideration of what connections are important to individuals as well as what is missing; the latter provides a springboard to goal-planning. Including place and activity dimensions to create a multidimensional network, although time-consuming, added to the explanatory potential of network types. We found that many participants reported the mapping interview procedure useful in terms of their own thinking and understanding about their networks, and at follow-up interviews some had already begun to make changes. This framework has therapeutic potential, is a whole-person rather than illness-focused approach and could help both individuals and practitioners to consider new opportunities, or steps toward these, that improve well-being, promote recovery and increase access to resources.

The analysis we undertook also took a person-centred focus for understanding the lives of people with SMI, as recommended in the recovery literature,74 through drawing out what was important to people and understanding why in terms of impact on well-being. The qualitative follow-up interviews were particularly helpful in understanding the dynamics of networks, how things change over time and the role of self or others in shaping ‘form’ or ‘content’, and why.

Aside from the individual’s perspective, the structure and composition of networks that this approach provides gives insights for practitioners into stage of recovery for an individual. 9 This can help identify areas of potential to address through goal-setting: isolation or lack of diversity, lack of family or friends can all be identified. The dimension of activity offers an insight into the individual’s identity beyond mental health. Through identifying what is meaningful and important to each person, interventions to build new connections can be targeted. Place is an important dimension because it situates the individual’s network inside the context of their local area; networks are constrained by and enabled by environment (social, cultural, economic). Both opportunities, such as local groups related to an individual’s interests, and barriers, such as transport, antisocial neighbours and safety, can be identified through consideration of place networks. People, place and activity networks interact. A framework can help both the individual and practitioners to build new connections that improve well-being, promote recovery and increase access to resources.

Significance versus importance

The LEEs expressed differing views on how to prioritise the relationship between statistical significance and interpretative importance. The comprehensiveness of the report was itself regarded as an important outcome because of the potential effect it could have on practice change. However, the wealth of quantitative data also presented an interpretative challenge:

The report doesn’t answer that much in itself as there could be many reasons for the same outputs.

One LEE suggested that statistical insignificance had been over reported which made it difficult to focus on the results that mattered. Another found some of the insignificant results unexpected and included them in priorities. Alternatively, all of the LEEs welcomed the statistically significant findings about the relationship of work to well-being as confirmation of existing research but did not rank it as important because these findings are already well known. Thus the importance of work to LEEs did not appear as a theme in its own right but was nevertheless active in the interpretation. LEEs also pointed to the findings around the over representation of ethnic groups in secondary services as both significant and very important but, because all LEEs were white British they struggled to interpret further from their lived experience.

All LEEs struggled with data volume in the network mapping and suggested that key terminology like social capital would need explaining in lay summaries. Despite these challenges, everyone did form a view about what was most important but found it more difficult to say what was least important. This question was only partially answered by LEEs

Data blindness on friendship in social networks

In stark contrast, the LEEs found the qualitative practitioner and stakeholder data much easier to process. Whereas, the network mapping offered multiple interpretative options, what was most important on first impression of the qualitative data was not there at all, that is the almost unilateral absence of evidence to demonstrate practitioner and stakeholder awareness of the importance of friendship in social networking for personal and social recovery.

LEEs acknowledged that social capital had a wider application than friendship but, given the importance of friendship to their recovery, they were disappointed by lack of recognition of this key area of human relationships. Although the qualitative reports showed realisation amongst practitioners and stakeholders that people with SMI had fewer friends, this insight appeared compartmentalised from practice and at distinct odds with the wealth of data within the network mapping. While it was difficult to conceptualise an absence as a concrete result, the LEEs were unanimous in prioritising it as a key if not the most important study finding.

Further themes from all reports were as follows:

Diagnostic differences on lifestyle and opportunity

For two of the LEEs, diagnostic comparisons were the differences that mattered most in the network mapping.

What I first noticed are the differences between the diagnostic groups on home ownership. That’s the thing that stood out for me. I concentrated more on these differences rather than differences between the 2 locations.

One LEE reported unease on first impression about organising people’s experiences under diagnostic labels as a continuation of the medical mode of interpreting and categorising experience:

When I read the report I don’t particularly see it in relation to diagnosis, rather in relation to people. I think listening/reading transcripts of their lived experience was very important. I did find it a challenge seeing the diagnosis at the top of each transcript in the same way that seeing the breakdown of data into schizophrenia and psychotic disorder, bipolar disorder and manic episodes and other psychoses.

Another LEE stated that diagnostic differences were the least important as the focus of the report was what people do to support recovery regardless of symptoms. For two LEEs the practical and educational implications of the report were most important although all had something to say about this. Nevertheless, for three LEE there was also a clear sense in which diagnostic differences were statistically significant findings which made some important health inequalities in social capital visible.

LEEs commented in particular on the diagnostic differences in lifestyle. For example, ‘only 6.8% of participants in the schizophrenia and psychotic disorder group owned a home . . . compared to 38.5% of those in the bipolar and manic episodes group’. This could be to do with severity of illness. A far higher percentage of participants with schizophrenia had been inpatients compared with those with a bipolar diagnosis and this may have had an adverse effect on their accommodation status. The impact of cognitive deficit for people with schizophrenia on lifestyle planning could also be higher. Crucially, well-being did not differ across the three diagnostic groups implying that it is not the severity of illness that mattered most to well-being. Rather, differences in well-being were linked to social factors such as employment, children and gender. Although some of these statistics made difficult reading to the LEEs, it was also suggested that the majority of these factors were open to change and could be acted on within a recovery focus to living well with continued symptoms of SMI. The LEEs were unanimous in suggesting that more support for accessing and maintaining employment was needed for people with SMI to reduce these lifestyle inequalities. For example, services open outside working hours to reduce time taken off and education for employers on how to get the best from staff with SMI.

Diagnostic differences on network stability and closeness

There was broad consensus that the quantitative evidence showed that network stability over time (10 years) was most important to subjective well-being. However, network emotional closeness was also seen as of equally high priority as people are more likely to confide in people they trust.

Pets were singled out as a potential source of network stability and closeness in so far as they provide non-judgemental companionship. Activities associated with pet ownership such as walking, buying food and visiting vets could offer purposeful opportunities to expand social networks. It would be interesting to see further research into whether those who are not in a relationship are still on a more positive recovery journey because they have pets.

LEEs were keen not to emphasise clinical factors when it came to developing targeted supports for people with SMI to improve their social capital. Paranoia, depression and social anxiety were also prioritised as obstacles to developing network stability and closeness. Although there was no significant difference in well-being linked to diagnosis, LEEs suggested that some symptoms have a social impact on network stability across the range of diagnoses. For example, it is harder to feel motivated to invest time and effort in other people if one is very depressed and it should go without saying that paranoia and social anxiety might make relationship-building more difficult than for people without these symptoms.

Diagnostic differences on network size, closeness and duration

Network size was also seen as an important factor in so far as participants who had larger networks and more people in their inner circle of closeness were significantly more satisfied with their networks. However, it was the number of contacts in the inner circle that was significantly correlated with higher well-being, suggesting that quality rather than quantity was most important. Participants with a diagnosis of schizophrenia also had lower number of contacts in the inner circle, suggesting that they might require more help to maintain and develop relationships over time. Cognitive deficit was singled out as a particularly challenging obstacle to relationship-building for this group which could require targeted support.

One LEE found these results particularly important. They confirmed expectations that network size of people with bipolar would be larger than for people with schizophrenia. However, the differences in size between weekly and daily networks suggested that diagnostic differences might also point to different kinds of social disablers to routine contact, perhaps greater effects of diagnostic labelling and stigma on daily contacts for people with schizophrenia and possibly more duty-driven weekly contact from people around them. People with a diagnosis of bipolar, alternatively, might find it easier to strike up daily short term relationships but could more easily sabotage them.

The most puzzling finding here was the larger inner circles for people with more inpatient stay which seemed at odds with findings for people with schizophrenia. LEEs suggested that possibly network stability could also be maintained through the social effects of having a serious problem but that such stability could also be a function of a highly medicalised social world. Exceptionally, people with bipolar managed in secondary care also had larger social networks than in primary care. Nothing else showed that secondary care improved the size of social networks for other diagnoses which seemed to have been reduced in secondary care. This could be an effect of greater monitoring and compassionate support in secondary services for people with bipolar.

Age and health inequalities

Another striking finding was the negative correlation between older age on social network activity. Against the general population, age was more statistically important than mental health in the reduction of social networks over time, suggesting that older people with SMI too could be at greater risk of becoming more socially isolated than peers. It is already known that there are greater levels of bereavement for the ageing population which would inevitably impact on network stability in both the inner and outer circles of closeness. However, LEEs also suggested that living with SMI at any age made it more difficult to access appropriate physical care which could also stack up into greater comorbidity in later life, potentially reducing physical capacity to maintain social capital earlier for people with SMI.

The mean age of participants was in their 40s which prompted some curiosity from LEEs about what would have been found with a younger population. Although this study did not include social media, LEEs suggested that the increasing importance of networking sites, like Facebook, to young people’s social networking activity and sense of social capital should not be underestimated in future research. Further, virtual networking via e-mail and websites of common interest had been of real value to three LEEs as an in between recovery space, that is, connectivity they could maintain at home between outside networks or employment even when they felt unwell.

Gender differences on network composition

Two LEEs noted that men had subjectively better quality ties but women had greater network size. This challenged gender stereotypes of women’s social networks as more co-operative and nurturing. It was speculated that men might demand more from their environment and/or might be perceived as more helpless, thus eliciting a more nurturing response from others. It was suggested that it might be more striking to perceive a man as helpless in relation to gender stereotypes of masculinity thus eliciting a greater sense of need. Alternatively, the gender stereotype of women as primary caregivers could work against them as not in so much need of nurture and support within their networks and against men as non-nurturing. The reports did not support further interrogation of these ideas. Nevertheless, it was suggested that a finding that was significant for so many people with SMI should be investigated further and that gender awareness around social networking should be part of any training developed as a result of the research.

Family, home life and parenting

LEEs welcomed recognition of the importance of family in stakeholder reports. Families could be a valued part of a person’s social capital but could also actively impede it or need extra support and education from services to most ably help the person with SMI. LEEs suggested that, unless the person lived alone, time spent at home needed to be cross referenced to quality of family life and impact on social networking behaviours. One LEE suggested that if home life was very poor this could lead to increased social networking as a form of escape. However, families that ‘wrapped the person in cotton wool’ could also be unhelpful. Broader discussion about the effect of home and family life on mental health suggested that this could be negative if someone turned into a ‘couch potato’ but could also be a refuge or place of safety which could resource social networking.

Two LEEs prioritised the findings about negative correlation between well-being and parenting as particularly surprising:

Do we know what proportion of those interviewed had children and were allowed to keep them? The ability for parents to stay with their children in a parenting role when they have a diagnosis of SMI is something that interests me . . . It’s interesting that, in my experience and talking with other mums, having children is quite often the reason given for preserving one’s life.

There was further curiosity about the role children played in their parent’s networking. Children give many reasons to develop social capital, e.g. contacts with other parents through schools, hobbies, etc. However, children may also view parents positively or negatively as part of their social capital The LEEs were interested to know more about how SMI is explained to children and their friends and how children describe their parent with SMI to their friends.

Location differences on network composition

Diagnostic, gender and age differences were more statistically significant than location differences. This led one LEE to suggest that perhaps the social difficulties faced by people with SMI were common. For two LEEs, differences in location were still more important. In group discussion, it was suggested that the urban/rural divide was also not straightforward in comparing samples as there were also a small city in the rural location. Nevertheless, some statistical differences between locations were intriguing. For example, there was a correlation between network size and well-being in the SW but not in London. Potential reasons for this were discussed in relation to cost of living in London, feeling anonymous in a crowd and not belonging anywhere. There were more long-term relationships in the rural sample which could support a greater sense of belonging and network stability. It was suggested that familiarity growing up around the same people could engender more community responsibility. Curiously, network stability in the inner circle correlated with well-being in London but not in the SW and people in the SW were less satisfied with their social networks overall. It was suggested that in a tight knit rural community there is no escape from being known which, when stigma is also present, could disable social networking for people with SMI from the start. Further, everyone knowing everyone’s business can be stifling and anonymity in a crowd could be viewed as exciting if a person were seeking to develop new networks or leave poor relationships behind.

There was debate about whether greater network size afforded more choice about who was accepted into the inner circle which could also engender more selectivity about the quality of these relationships. In group discussion, it was considered whether there might be a sense in which the urban streets could be viewed as paved with gold from a rural location, as promising more choice from a larger pool. Interestingly, there were no significant differences between sites on the overall number of place connections made and no correlation between size of place connection and well-being. This finding appeared to weigh against the greater choice theory but, as more than one LEE pointed out, satisfaction and well-being are not the same. Perceived difference in choice could account for dissatisfaction whether this affected well-being or not. LEEs also pointed to location discrepancies on availability of paid work, suggesting that differences between voluntary and paid opportunities might explain some of the higher dissatisfaction with networks in the SW.

Importance of individually generated place connections and individually structured activities

One of the biggest surprises was that the majority of place connections were built by the individual, most likely through their own interests and proximity to a place and everyday routines. At face value, this finding appeared to suggest that people with SMI did not need support with developing social capital. However, this was not borne out by the diagnostic, gendered and age differences. Further, the amount of places attended outside the home had no bearing on well-being suggesting that it may be what people get out of going to specific places that is more important. Only individually structured activities were significantly positively correlated with well-being, i.e. activities need to have some structure or a goal to be beneficial. Unsurprisingly, the data showed that those who worked full time had significantly higher well-being that those who did not. Thus LEEs speculated that paucity of practitioner structured activities and/dominance of the medical model could also account for the high prevalence of individually generated activities which might not by themselves always be linked to well-being or any kind of meaningful activity.

Current levels of service support for maintaining and developing social capital in primary and secondary care

All LEEs agreed that recovery orientated practitioners would be ideally placed to encourage connection building that worked in support of people’s personal choices and interests and to help with goal setting to develop structured activities and self-management techniques. The qualitative data showed variable recognition of this. There was a strong investment from stakeholders in recovery as a management and strategic ideal but something of a bypass in translation between this and practitioner behaviour where highly medicalised language and practices often dominated.

The LEEs also identified a mismatch between recovery rhetoric and application across both Chapters 6 and 7. There was a strong feeling that practitioners who really understood recovery as living a meaningful life with continued presence of symptoms would not need the importance of friendship pointing out. However, it was also clear that sometimes recovery was misunderstood by professionals, for example as setting goals for rather than with service users. The LEEs observed that quite often reports demonstrated a superficial understanding of recovery. For example, frequent references to ‘the recovery model’ in the stakeholder’s report showed lack of understanding of the most basic point; that is there is no model that can adequately capture what recovery means to each individual. The closest thing to a model is person-centred practice. One LEE suggested that sometimes a practitioner could implicitly understand this, for example, a GP might ask questions about friends, family and other relationships as part of their routine ‘bedside manner’ but not have conceptualised this as part of a recovery approach. Practitioners also appeared to recognise the loneliness of people with SMI but did not often appear to take the extra step of acting on this information.

The LEEs agreed that the reports showed that mental health services do not do enough proactively to help people keep in touch with contacts and maintain and develop meaningful activities in the community. This may be particularly important for those who need inpatient treatment or have lost contacts because of their mental health. Differences between secondary and primary care were also key to LEEs in making sense of the lack of awareness of social networking amongst practitioners:

This is interesting – do practitioners look at people’s networks. Certainly when I was unwell, no-one made even the slightest attempt to examine my social networks. I think somewhere in the report, it mentioned that participants receiving care from just primary services were less likely to ‘get their networks looked at’. This surely has implications for those receiving primary services.

LEEs expressed concern that cuts and closures of voluntary services would result in an even poorer, crisis- driven mental health service. The reports showed that financial constraints and difficulty obtaining resources often led to earlier discharge of significant numbers of people from secondary services back to GPs to provide services to the most needy. Again, people receiving services from primary care were seen as particularly vulnerable to falling through the net of finite resources.

The data showed people with SMI did not necessarily go to GPs with an awareness of the importance of social networks. There was concern from two LEEs about the high numbers of people in the SW who did not know why or did not say why they had visited their GP. It was also a source of comment that only 10 per cent of people’s social networks in primary care were with other people using mental health services. There were mixed views about this. One recounted the detrimental effects on her mental health of being referred to a mental health support group. However, LEEs also spoke about long-term friendships with other people with SMI as key sources of network stability and closeness. LEEs talked about getting the balance right for each individual. Joining groups explicitly set up for people with SMI could be helpful in facilitating an initial sense of belonging somewhere for people who are very isolated, paranoid and/or socially anxious which could then be used as a platform to access community groups and activities with greater confidence. However, joining mental health groups could be less helpful if that is all a person does, fostering a sense of dependence and sick role.

The qualitative data showed considerable hesitance on behalf of practitioners to refer to community groups as they had no way of knowing whether they would help or not. There was a risk that a community group would have little understanding or acceptance of SMI, potentially making things worse. However, there was also limited local knowledge of community resources for people with SMI including voluntary services. LEEs suggested that people with SMI evaluated their experiences of participating in community and third-sector groups and contributed on-going feedback to an electronic resource for practitioners to keep the resource in date.

Communication between secondary and primary services was high priority, with discharge again seen as a point of increased vulnerability where GPs did not always feel knowledgeable enough to help or felt they lacked understanding of SMI. Good ideas in the report for managing this more effectively included locating a CPN or link worker in GP surgeries. Advance statements in primary and secondary care could also help with maintaining existing relationships during a period of crisis rather than leave it to the person with SMI to start again on discharge.

Face-to-face contact was also flagged up as key to relationship building. Stakeholders could be more proactive here by inviting key members of voluntary organisations to meetings, where they could talk about their services and hold an annual ‘getting-to-know-you’ event and allow mental health service staff time off to attend. LEEs also suggested producing guidelines and leaflets for all relevant agencies.

Recommendations

1. Further research would help to understand more rigorously the statistical differences on diagnosis, age, gender and lifestyle prioritised above, curiosities expressed about home, parenting, family life and the ethnic differences on secondary and primary care not examined by this report.

2. LEEs also suggested that many results were sufficient to ground training focused on implementing network mapping sensitive to common symptoms and problems faced by people with SMI and developing databases of community resources for practitioners.

3. Peer support could help people with SMI develop strategies for navigating social relationships whereas service user-led recovery training could help practitioners see the value of working with clients on their social networks to develop self-management strategies and support lifestyle goals such as employment and a decent place to live.

4. Stakeholders need to take a whole-systems approach to optimising best use of finite resource between primary and secondary care. Link workers in GP surgeries and encouragement of advance statements could be piloted as an approach to communication break down between secondary and primary services, maintaining and developing continuity of social network and as a way of sharing knowledge about SMI for people managed solely in primary care.

Micro mapping

It was also noted by one LEE that many of the everyday networks and brief interactions that support well-being in a community were not included in the report but could usefully be drawn on when supporting someone with SMI. For example noting regular interactions with a local shopkeeper, persons met frequently walking a dog, taxi drivers or people using the same bus route or train and the virtual relationships described above. It was suggested that the micro effects of these kinds of interaction (or lack of) could greatly contribute to person’s sense of well-being or impede it in connection with a specific place, meaningful activity or regular geography. The style of interaction in these micro relationships could be important to the on-going expansion or limitation of a person’s social world. For people who choose to disclose, it was suggested that there are ways of talking to less close people about a mental health diagnosis which can be more or less stigma busting which could be supported by mental health and/or peer workers.

Toxicity mapping

Most LEEs pointed out that all relationships and activities contained the potential to be enabling or disabling. As well as mapping for strengths, people with SMI could be supported to map for toxicity. That is, unhelpful social influences and relationships. LEEs were divided on how far toxicity was a useful term. Two suggested that this accurately named some of the less helpful social influences in their lives and two suggested that this term could be stigmatising of the person identified as a toxic influence. It is rarely black and white and could be more helpful to identify toxic behaviours within networks and develop assertiveness strategies for dealing with them. Ultimately, it would be down to the person with SMI how they named less helpful relationships in their networks.

Finally, all of the LEEs agreed that social network mapping would have been helpful to them if they had accessed it when first unwell:

Looking at the mapping of people’s social networks exercise has been really interesting. It made me realise that something was missing from my mental health treatment. I really wish someone had done this for me when I was unwell and then taken steps to help me expand my networks.

Ultimately, it does not cost much to sit down with someone and find out what their human resources already are and how they could improve on them. In this sense, social network mapping holds out real hope in these gloomy economic times for services to improve social capital and become recovery orientated in practice as well as strategic vision.

Diagnosis checking

Place

Activity