‘They need to ask me first’. Community engagement with low‐income citizens. A realist qualitative case‐study

Abstract Background Community engagement is seen as key to citizen‐centred and sustainable healthcare systems as involving citizens in the designing, implementation and improvement of services and policies is thought to tailor these more closely to communities’ own needs and experiences. Organizations have struggled to reach out to and involve disadvantaged citizens. This paper examines how if, why, and when low‐income citizens wish to be involved. Methods For this qualitative realist case‐study, 19 interviews (one dyad) were held with (20) low‐income citizens in two Dutch municipalities. Additionally, the results were discussed with a reference panel consisting of professionals and citizens to enrich the results and to ensure the results had face validity. Results The results showed four different ways in which low‐income citizens wished to be involved: (a) in a practical/volunteer way; (b) as a buddy; (c) as a lay expert; (d) not involved at all. The factors affecting citizens’ interest and capacity to participate include citizens’ own experiences of the services they access and their personal situations, e.g. their mental or physical health, extent of financial crisis, family situation, home environment. None of the interviewees was currently involved, but all had ideas for improving health(care) services and policies. Citizens’ experiences of the services they accessed acted as a motivator for some to be involved as they wanted to ensure others would not have the same struggles, while for others their own needs and an apathetic system remained too high a barrier. To enable involvement, citizens need continued support for their own health(care) and financial situation, better communication and accessibility from services, practical support (e.g., training and bus passes) and recognition for their input (e.g., monetary compensation). Conclusion The study shows that citizens’ experiences of the services they accessed influenced if and how they wanted to be involved with health and care services. Despite the fact that all participants had shared solid ideas for improving services and policies, they were hindered by a bureaucratic, impersonal and inaccessible system. Organizations seem to underestimate the required investments to reach out to low‐income citizens and the support required to ensure their involvement. Patient and Public Involvement (PPI) in Study Citizens as well as PPI organizations were members of the reference panel who helped formulate the research questions and recruitment strategy. The local reference panel also helped to interpret and refine the initial findings.

role CE is expected to play in the improvement of communities' wellbeing and in the quality of services and policies and despite the corresponding investments many organizations are placing within CE, important criticisms remain. One of the most important criticisms concerns the disparities in participation rates and that only a select group of citizens are enabled and empowered to participate, that is white, middle-class and often retired citizens. 2,3 Especially young people and low-income citizens go unrepresented in CE initiatives. 8 Previous studies have shown that social inequities in engagement stem from the fact that CE approaches are mostly based on the interests, needs and norms of those designing the approaches and that 'engagement environments' are often built for efficiency in terms of time, money and tighter budgets, instead of citizen empowerment. 9 Such emphasizing of efficiency over empowerment in engagement approaches results in a loss of influence for citizens, especially for those who are socioeconomically disadvantaged. [9][10][11][12] As such, when vulnerable or disadvantaged citizens are engaged, they report feeling shut out and unable to share their experiences and perceptions, their skills and insights-for example, because of poorly timed or advertised engagement activities, a lack of safe and trusting environments or power imbalances. 9,11,12 This, in turn, can lead to citizens feeling frustrated, cynical and apathetic. 7,9,[13][14][15][16][17][18][19][20][21] Partly because only a select group of citizens are involved in the shaping, decision-making, planning and/or governance of health and care services on behalf of whole neighbourhoods or communities, many CE approaches may fail to improve the health outcomes for all citizens and may thus even lead to a deepening of health inequality rates. 3,22,23 Such criticisms are important to consider as without better inclusivity and diversity in the representation of CE approaches, it may be harder to ensure health systems and communities are reflective of diverse experiences and thus to address a wide variety of needs. 7,9 The previous literature has not just highlighted the structural issues with engagement approaches (i.e., organizations' inaccessible structures and processes) but also indicated key areas for improvement. For example, 7,14 developed nine guiding principles for the successful implementation of CE approaches highlighting the importance of, for example providing supportive leadership, fostering safe and trusting environments, acknowledging and addressing power imbalances between citizens and organizations, ensuring citizens' early involvement and creating a shared vision for CE. While, Cyril et al. 3 pointed out key components that can specifically improve disadvantaged groups' ability to be involved, including bidirectional learning and using peer-delivery. Previous studies have also suggested the importance of properly investing in organizational cultures based on levelling power dynamics, giving voice to that previously excluded and true community empowerment, for example, and importantly by skilling-up professionals in community outreach, engagement and empowerment. 3,7,9 Despite previous studies highlighting how and why many 'mainstream' CE approaches exclude disadvantaged citizens, 7,9,13,15,17,18,20,21 organisations still struggle to design and implement CE approaches that either specifically engage disadvantaged groups or which are inclusive of a wider range of citizens. 7,9,22 This is because many studies have highlighted where organizations have gone wrong, for example inaccessible organizational structures and processes, power imbalances, lack of safe and trusting environments, 7,14

but few have investigated what disadvantaged groups'
(support) needs, interests and motivations are to become involved. 2,[24][25][26] The aim of this study was to examine how low-income citizens' experiences of health(care) and income-support services influences if, when, how and why they wished to be involved in improving health(care) and income support services and policies. As such, this DE WEGER ET AL. case-study investigated low-income citizens' perceptions and experiences of the services they access and their interests in being involved with organizations to improve health(care) and support services and policies to ensure. This paper presents the results of a realist qualitative case-study, which explored if, why, when and how low-income citizens and citizens with financial support needs wish to be involved in helping others to improve their financial security and well-being. The study's aim was to investigate the support low-income citizens required to be involved and what the contextual factors and mechanisms were explaining their involvement preferences and needs.

| METHODS
This case-study is part of a 4-year realist qualitative multiple casestudy evaluating the development of CE approaches in six different regions in the Netherlands. The multiple case-study was undertaken in consultation with a reference panel. The panel consisted of stakeholders involved in developing CE approaches within the six different regions, including policymakers, involved citizens, patient and public involvement (PPI) organizations and experts in the field of public health, health inequalities and citizen participation (see Table 1). The panel therefore helped to ensure that the study addressed stakeholders' questions regarding CE and addressed relevant gaps in the literature. The study was informed by the realist evaluation (RE) approach.
RE is especially suitable for complex topics and interventions. To understand how an approach might generate different outcomes under different circumstances, RE uses context-mechanism-outcome configurations (CMOs) to unearth the underlying changes in reasoning and behaviour of participants that are triggered within particular contexts (or by particular contextual factors). More so than other methods, RE seeks to answer: 'what works, how, why, for whom, to what extent and under which circumstances'. 27,28 In this way, RE helps to not just study people's experiences but helps to understand the underlying causation for those experiences. 29 By using RE the study sought to understand the causation behind low-income citizens' involvement preferences and support needs with the expectation that this would enable organizations to better understand citizens and thus tailor their involvement approaches and support to citizens' needs and experiences (see Table 2 for definitions for realist concepts).

| Recruitment strategy and study sample
For this case-study, the authors collaborated with two Dutch municipalities and a PPI organization that wanted to better understand Type of organization Type of function According to the PPI organization, municipality B had developed the best-practice low-income support services within the region. The researchers hoped that the differing contextual factors within the municipalities would help to compare and contrast low-income citizens' experiences accordingly (see Table 3).
Acknowledging that the recruitment of a 'hard-to-reach' group would take more time and effort, the researchers had prioritized the additional efforts it might take and not set any specific timeframes for the recruitment, focussing instead on achieving data saturation and ensuring all those who wanted to participate were given the chance to do so in the manner that suited them and their needs. For this study, purposive and snowball sampling 33 was used to recruit citizens who were living in the two municipalities and were, for example, long-term unemployed, employed but receiving a (too) low income and were therefore living under or around the local poverty line, in significant and long-term debt, receiving or entitled to receive (income) support for their financial instability. Firstly, all citizens receiving welfare support from the municipalities were sent an invitation letter-approximately 200 letters were sent out and eight participants had responded to the letter. Secondly, the municipalities' support workers reached out to clients who were receiving income support and they felt were stable enough and interested in sharing their story (purposive sampling). Thirdly, citizens were also approached by the researchers themselves through local food banks (researchers went to the local food banks themselves and T A B L E 2 CE-oriented definitions of realist concepts 7,14,28,30 Intervention Refers to interventions' implemented activities, strategies and resources, 31 e.g.,: citizen advisory panel meetings, or neighbourhood organized workshops.

Context
Pertains to the backdrop of a programme and examples of context include, e.g., pre-existing processes, cultural norms and history of an organization or community in which a programme is implemented, geographic location effects, funding sources, opportunities or constraints. Contexts can therefore be understood as any condition that triggers or modifies the behaviour of a mechanism.

Mechanism
Mechanisms describe how the resources embedded within a programme influence the reasoning and behaviour of programme participants. Mechanisms are usually hidden, sensitive to variations in context and generate outcomes, e.g., citizens feeling more empowered due to learning opportunities.

Outcome
Refers to intended, unintended, or unexpected programme outcomes on the micro, meso or macro level, e.g., sustainability, quality and integration of services (macro), citizens' level of involvement in health and care services (e.g., in designing policies) (meso), citizens' health and well-being outcomes (micro).

CMO
CMO is a heuristic used to explain generative causation. CMOs help to reflect on the relationship between a context, mechanism and outcome of interest in a particular programme. CMOs can be about a whole programme or only certain aspects of a programme. Formulating and refining CMOs is largely how researchers analyse data in RE as it allows for a deeper understanding of which (aspects of) interventions work, for whom, under which circumstances and to what extent. 30 CMOs are also used to generate or refine programme theories, which, in turn, help shape the final product of an evaluation (e.g., recommendations). CMOs are also to generate or refine programme theories.

Programme theories
Is a hypothesis about how a programme or programme component may or may not work, under what contexts and with what outcomes. In this study, the guiding principles (De Weger et al. 14,29 ), which can be seen as action-oriented programme theories, were used as initial programme theories.

T A B L E 3 Municipalities description
Municipality A • Facts about the municipality: rural municipality with favourable unemployment and welfare support rates. The national unemployment average is 3.8% and in the municipality the unemployment rate is 2.7%. The national welfare support rate is 5.2% and in the municipality it is 1.9%. 32 • Reason for including municipality: municipality wanted to improve their employment and income policies and support services by aligning them better to citizens' needs and experiences.

Municipality B
• Facts about the municipality: rural municipality with unfavourable unemployment and welfare support rates. The national unemployment rate is 3.8%, and in the municipality, it is 5.1%. The national welfare support rate is 5.2%, and in the municipality, it is 6.2%. 32 • Reason for including municipality: municipality with, according to the involved PPI organization, best-practice unemployment and income support services Abbreviation: PPI, Patient and public involvement.
collaborated with the volunteers), nonprofit emergency funds (who like the support workers were asked to reach out to clients who were on income-support and they felt would be interested in taking part in an interview), and local churches (who researchers collaborated with to invite anyone who was on income-support and interested in taking part in the study (purposive sampling). Approaching citizens through their support workers had the highest response rate (12). When potential participants reached out to the researchers about the interview, researchers discussed with potential participants that they were under no obligation to take part in the interviews and that

| Data collection
To aid data analysis, all interviews were recorded and transcribed. All study participants had received an information letter and had pro- there was a high rates of recurrence of response). Furthermore, additional interviews and already been planned and were therefore held as conformation and to provide participants with the opportunity to share their stories as well. 34 Ultimately, 19 interviews (one of which was a dyad interview) lasting between 1 and 2 h were held with a total of four respondents from municipality A and seventeen respondents from municipality B. Five interviews were conducted over the telephone and fourteen interviews were held face-to-face interviews. Interestingly no participant chose to do video call interviews (e.g., through Zoom, Whatsapp or FaceTime). The majority of interviews were conducted in Dutch; however, three interviews were conducted in English. The study received ethics approval from the corresponding university (reference EC-2017.96) and the interviews were held between September and October 2020.

| Data analysis
This study took an inductive and deductive approach for analysing the data, using previous literature and theories on CE, poverty and welfare support to form the basis of the analysis approach. This study is part of a 4-year multiple case study that examined how CE was  Table 2). Furthermore, the MaxQDA coding process was based on previous literature and models on the social determinants of health and poverty (see Supporting Information Appendix I and II). [35][36][37][38] Additionally, the nine guiding principles for successful CE were used as the theoretical framework to examine the underlying contextual factors and mechanisms explaining citizens' involvement preferences. 7,14 The clustering followed a deductive and inductive, sequential and iterative process, which has been applied in previous studies and described elsewhere 7,14 : (1) CMOs were coded and clustered into citizens' support needs, their preferred way in helping other citizens improve their financial security and overall well-being.
(2) The authors discussed the clusters and thematically analysed, reviewed and discussed them again.
(3) The final draft of the clustered CMOs was shared with all authors to confirm and refine the themes.

| RESULTS
The following section will first describe how citizens' perceptions and experiences of the support services they access influence their involvement preferences, including their reasons for wanting to be

| Interview participants
The 21 interviewees had a wide variety of backgrounds, employment and educational experiences, family situations and support needs.
Many interviewees had become unemployed due to physical health issues (e.g., musculoskeletal, degenerative diseases) and/or mental health issues (e.g., post traumatic stress disorder, autism), others due to family breakdowns or domestic violence, some due to a history of drug abuse, others were on income-support as they had come to the Netherlands as refugees and were trying to build a new life in the country. Many interviewees were struggling with a complex combination of these issues. The vast majority of interviewees had negative experiences of the support services they accessed and described services as uninterested and unsympathetic and/or as fragmented and bureaucratic. As the section below will show these personal contextual factors and the organizational/systematic contextual factors influenced whether they wanted to be involved or not, but also negatively impacted their ability to do so. T A B L E 4 Summary of participant characteristics, experiences, involvement preferences and required support for involvement  Table 6). It is important to note that there was overlap between the four categories in that some interviewees wished to be involved, for example as a buddy and a lay expert or would have liked to have been involved as a lay expert but currently felt unable to as they were still too much in crisis. In the following section, this paper will further explain the contextual factors and mechanisms underlying the four types of involvement preferences (see Supporting Information Appendix III for more CMO examples and Table 5 for a summary of low-income citizens according to health and care needs, experiences and involvement preferences).
Three interviewees wanted to be involved in a practical/volunteer way as they wished to contribute to their own community in the hopes of building up their social contacts. Examples interviewees gave were for example organizing meetings between health(care) organizations and low-income citizens, or picking up trash and cleaning up their neighbourhood's parks and gardens with others.
One interviewee stated that she had experienced the care and financial support services as unaligned to her own needs and the processes and structures to apply for financial support as too complicated and fragmented (context). She wanted to contribute to the community and support others in financial difficulty but was worried that one-on-one support would become too personal and difficult to cope with (mechanism). Thus she would prefer to be involved in a practical, more organizational manner (e.g., organizing meetings between citizens and the municipality (outcome). However, she stated that she would need organizations to reach out to her and ask her to become involved and to provide clear boundaries for her involvement (outcome). • Citizens who wanted to be involved (regardless of their preferred involvement category) were motivated to do because they wanted to spend their time more meaningfully and to add some structure to their day-to-day routine • There is overlap between the buddy and lay-expert involvement categories due to the desire to improve services and policies for others going through the same issues. However, those preferring the buddy involvement category were focused on one-to-one interaction, while those preferring the lay-expert category were more focused on collaborating with organizations and professionals. • Those not wanting to be involved at all (beyond sharing their experiences on a one-time basis) described not feeling stable enough to be involved on a long-term basis. Finally, a majority of people could not or did not want to be involved in any shape or form (11). The majority of interviewees within this group said they did not want to participate because they were afraid they would not be able to cope with their own situation. A smaller number of interviewees said they did not want to participate because they simply no longer trusted organizations due to their own negative and stressful experiences. Some interviewees who did not want to participate could see themselves being involved if they received the proper support and/or compensation while being involved and thus enabled to be involved in their own preferred way. For example, one interviewee stated that her support workers had advised her to do volunteer work to spend her days more meaningfully; however, she still struggles with her physical and mental health issues (context). Another interviewee, with a history of drug abuse and unemployment, mentioned he would in theory like to be involved as a buddy to help others with a history of drug abuse, but currently felt unable to focus on anyone else but himself and stabilizing his own situation. Another interviewee said she had had purely negative experiences with income-support due to a lack of communication, lack of consistent support and bureaucratic mistakes (context). Because of these negative experiences she feels abandoned by the services meant to support her (mechanism). She is unmotivated to be involved with organizations to ensure others receive better care and support from services and only wants to focus on herself and her own family (outcome

| Support citizens require to enable their involvement
First of all, it is worth noting that none of the interviewees were, at the time of interviewing, involved in any of the organizations or volunteer work. It could therefore be argued that by not providing the support low-income citizens require to be involved, citizens miss out on chances to be involved and organizations miss out on ideas to improve their services and policies. The results indicate that improved accessibility and communication is vital to enable low-income citizens' involvement as this will help tackle the constraining systematic contextual factors facing citizens for their involvement. To start with, interviewees mentioned the need to be approached by organizations and to have their participation options explained. Interviewees also highlighted the importance of ongoing support for their own (mental)health, well-being and financial status and that any potential involvement on their part would not put any more stress on these factors. Furthermore, interviewees discussed the importance of feeling heard and listened to by organizations and feeling valued and recognized for their (potential) contributions. Some interviewees mentioned that they would need training, for example to be a lay expert or buddy, for refugees in learning Dutch. Finally, some interviewees mentioned that (monetary) compensation for their involvement would also help them feel more motivated and to enable them practically to get involved (e.g., for transport costs) (see Supporting   Information Appendix II for more CMO examples).
For example, one interviewee experienced that financial support services were very 'business like' and that the professionals had very little empathy for people presenting to the services in crisis (context).
She really wanted to be involved as a lay expert to support people in crisis, to improve organizations' communication to low-income citizens and to ensure organizations took a more person-centred approach, but she feared she would not be able to cope with the responsibility and added pressure (mechanism). To become a lay expert, she would need support for her own financial status and wellbeing, she would need the training to help her fulfil the role of lay expert and mentioned a small monetary contribution (outcome). Only one interviewee had said that she had approached the municipality and told them she wanted to be involved to build up her social contacts and to help her spend her time meaningfully (mechanism). However, the municipality had never gotten back to her about the different ways she could get involved and is still waiting to hear back from them (outcome). This means the municipality missed the opportunity to involve citizens (outcome). All interviewees had shared useful ideas and suggestions for improving services and policies and for aligning these better to their own needs and lived experiences.
This includes participants who had said that they would not or could not be involved because they were too much in crisis themselves,

| Panel deliberations
Both professionals and citizens within the panel highlighted that they found it a challenge to involve harder-to-reach groups. Professionals within the panel recognized that their communication and accessibility should be improved, while citizens highlighted that organizational communication and accessibility was also often a barrier to their own involvement. In an effort to improve accessibility, professionals recognized the need for them to develop trusting relationships with (low-income) citizens. They thought that using a neutral partner or community figures could help them to build such relationships. They also discussed the compensation they could provide citizens for their involvement. Finally, based on these results, the professionals expressed a wish to develop a broader spectrum of involvement opportunities for citizens to try and align their CE approaches more closely with citizens' interests and needs. However, professionals expressed that to improve their accessibility and to broaden citizens' involvement opportunities, they would need to be given more resources and space to innovate their approaches.  Table 6). However, none of the interviewees had been enabled or asked by organizations to get involved in any shape or form, further underscoring the systematic/organizational contextual issues that negatively impacted their ability to be involved. This highlights that by excluding low-income citizens and preventing them from discussing their experiences and needs and the ways in which they want to be involved and the support they require to enable them to do so successfully and sustainably, organizations are missing out on important ideas to improve their services, policies and organizations. This suggests organizations also miss out on opportunities to ensure the health and care system is more inclusive and representative of a wider variety of citizen needs and experiences. On an important note, while neither the interviewees nor the organizations would describe participants sharing their experiences and ideas during the interviews as 'involvement', such activities could be seen as a first step to enable low-income citizens to be involved in the manner best suited to their own interests and needs, thus enabling them to improve services and policies.

| DISCUSSION
This study's interviewees' negative experiences of apathy, bureaucracy, fragmentation and never being asked about their experiences or whether they would like to be involved shows the importance of tackling constraining systematic/organizational contextual factors. 3,9,14 The lack of involvement and outreach from organizations towards low-income citizens is in line with how citizens had experienced the services they had accessed: as unpersonal and apathetic, as bureaucratic and fragmented. Previous literature suggests that consecutive (western) national governments' policies regarding benefits and income-support have been focussed on improving efficiency and effectiveness. This has made it more difficult for organizations to deploy resources to alleviate poverty and social exclusion. 12,39,40 Previous authors like Cortis 12 suggest that this is a major reason why organizations tend to over-provide to those easiest to reach and assist and where results are more demonstrable and at the same time underprovide to more disenfranchised citizens who are more 'challenging' and costly to assist. Such a policy environment makes it more difficult for organizations to assist more marginalized groups and to promote more equal involvement. 39,40 The reference panel discussions underscored such findings as organizations had not seemed to prioritize the outreach of 'harder-toreach' groups. They stated they were still searching for outreach and engagement activities more specifically tailored to disadvantaged groups' needs. Such upfront support may well differ from the support and investment already-engaged citizens require (as highlighted within the original guiding principles).
Finally, by examining citizens' involvement preferences, this study highlighted the importance of examining and addressing citizens' motivation to be involved. 7,14,42 By investigating whether citizens are (a) 'intrinsically motivated'-like the interviewees who wanted to do practical volunteer work to build their social networks or to have a daily routine-or are (b) 'identified regulated'-like the interviewees who wanted to be involved as buddies or lay experts because they felt it was important to improve services and policies for others-and/or (c) 'externally motivated'-for example like the interviewees who were motivated to take part in this study because of the gift vouchers-organizations can become more sensitive to citizens' motivations and needs. 43 It will also help organizations to understand when to provide rewards (e.g., training, stipends) to trigger 'external motivation' when involvement approaches are too far removed from citizens' own personal motivating factors. Understanding citizens' motivating factors is especially important when trying to reach out to harder-to-reach groups whose feelings of distrust towards (local) government and public sector services negatively impact their motivation to become involved. 44

| Study limitations
One limitation is the fact that this case-study only included citizens from two rural municipalities. Presumably, the rural context of this case-study influenced several important factors, including the availability and accessibility of health, care and financial support services available to interviewees. This limitation was mitigated by the reference panel's workshop discussions as this confirmed the validity and applicability of our interview findings in other contexts, thus further validating and enriching the interview findings.

| Future studies
This case-study seems to indicate that low-income citizen involvement is possible if organizations were to prioritize reaching out to harder-to-reach groups based on citizens' own needs and experiences, rather than on organizations' interests. The study also suggests that low-income citizens have valuable insight and suggestions for the improvement of services and policies. Future studies could investigate if and how low-income citizens' involvement would affect services and policies and ensure they are indeed better aligned with citizens' needs and lived experiences.

| CONCLUSION
This study investigated how low-income citizens' perceptions and experiences of the services they access influence if, why, when and how they wished to be involved with health and care services to ensure all citizens in their communities are financially secure. The study shows that citizens' experiences of the services they accessed influenced if, why, when and how they wanted to be involved with health and care services. All citizens, whether they wanted to be involved as (a) in a practical/volunteer way; (b) a buddy; (c) lay expert or (d) not at all, all citizens had solid ideas for improving services and policies and aligning them better to citizens' own needs and experiences, rather than organizations' interests. However, despite these solid ideas for improving services and policies, participants were prevented from sharing these by a bureaucratic, impersonal and inaccessible system. Organizations should not underestimate the required investments to reach out to low-income citizens and provide the required support to ensure their involvement (including e.g., person-centred and empathetic approaches, transparency and accessibility, training, support for their complex needs). Such investments will help reduce the bureaucratic barriers citizens experience and will help them feel more listened to. With low-income citizens' involvement, organizations can improve their services and policies and ensure these are more aligned with all citizens' needs.