Social Inclusion and Sustainable Development: Findings from Seven African and Asian Contexts

Abstract

Social inequities have widened divisions between diverse population groups. Inequity is associated with social exclusion, structural and physical violence and reduced development, which in turn are linked to civil unrest, conflict and adverse health and social outcomes. Public services are key institutions through which social inequities are created and maintained, but evidence on viable interventions to reduce institutional exclusion is limited for low- and middle-income (LMIC) contexts. We identify common drivers of institutional exclusion across diverse populations in LMICs and inclusion strategies that could potentially work across populations, public service sectors and country contexts. Seven studies engaged with over 385 key stakeholders in healthcare, education and local government settings in Bangladesh, Kazakhstan, Kenya, Myanmar, Nepal, Nigeria and Vietnam. Participatory research, in-depth interviews, policy reviews and multi-stakeholder workshops focused on a range of disadvantaged groups. A multi-sector partnership co-produced recommendations at each site. Findings were synthesised to identify common themes and a framework for social inclusion across disadvantaged populations. The invisibility of disadvantaged communities in public service planning and delivery processes helped maintain their exclusion from opportunities and resources. A spectrum of neglect, restrictions and discriminatory practice reflected structural violence linked to poor life chances, illness, physical abuse and death. Key recommendations include the representation of disadvantaged groups in service staffing and decision-making and the transformation of public service policy and practice to develop inclusive, targeted, collaborative and accountable systems.

1. Introduction

Reducing inequity is essential for sustainable development and social stability that leaves ‘no one behind’ [1]. Failure to address inequity has led to widening divisions between rich and poor, men and women and diverse ethnic, religious and other social groups. This, in turn, is linked to civil unrest, conflict and humanitarian crises, as well as to losses in national productivity and economic development [2,3,4].

Evidence about how to address these issues through more inclusive public services is concentrated in high-income contexts and indicates an urgent need for multi-sector and multilevel strategies [2]. This study aimed to fill current evidence gaps, drawing on a conceptual framework that encompasses social inclusion and exclusion, structural equity and sustainable development. Our research questions explore common drivers of exclusion across diverse groups that experience disadvantage in low- and middle-income countries (LMICs) and strategies that could work across excluded populations, public service sectors and country contexts.

Social Exclusion, Structural Violence and Sustainable Development

The concept of social exclusion is understood to be a key determinant of health and well-being and is strongly influenced by the degree to which individuals and groups can equitably access resources and public services such as healthcare, education and criminal justice systems [1,2,5]. Social exclusion involves processes driven by unequal power relationships that result in a lack or denial of resources and rights and the inability of some social groups to participate on equal terms in economic, social, cultural or political arenas [6].

Dynamic and multi-dimensional processes have been shown to maintain exclusion through unequal power relationships that interact across sociopolitical (macro), institutional (meso) and individual or community (micro) contexts [2]. These processes involve systemic injustice linked to discriminatory policies and practices that have adverse impacts on individuals and society as a whole [4,5,6]. At a structural level, a spectrum of discriminatory practices, ranging from neglect to physical violence, can be understood as structural violence that is used to reinforce inequity in power relations [7,8,9]. As with social exclusion, theories of structural violence identify multi-dimensional processes that generate and maintain social hierarchies in the context of unequal power relationships [7,8,9]. Punitive stereotyping, for example, has been highlighted in relation to lower-quality healthcare and is associated with the higher mortality and illness rates often found within socially excluded groups [10]. Social pain theory provides a further framework through which to view adverse dynamics within excluded populations, such as gender-based violence and criminal activity. Within this framework, social exclusion is conceptualised as a primal threat to well-being that provokes aggression and other heightened emotional responses similar to those caused by physical threats [11].

Public services provide a key mechanism for these harmful processes to operate at multiple levels simultaneously, creating and maintaining inequitable social hierarchies [2,5]. Apart from the physical and emotional pain inflicted on excluded populations, these processes of social injustice have economic costs in terms of wasted human potential and adverse impacts on sustainable development [12,13]. Harvey (2001) [14] links these damaging processes to the capitalist system, which produces inequalities through its focus on accumulation (p. 278) in the context of chronic power imbalances and competing claims between social groups (p. 116). Within such systems, ‘equal treatment’ of people who are not on equal footing produces consequences that are clearly unjust [14] (p. 191).

Our research aimed to identify key processes that maintain the exclusion of disadvantaged communities from opportunities and resources, involving neglect, restrictions and discriminatory practices linked to poor life chances, illness and physical violence. Our working hypothesis was that a transformation of public service policy and practice to develop inclusive, targeted, collaborative and accountable systems is essential to future sustainable development.

2. Methods

Our approach to the research of viable equity interventions aimed to model social inclusion and was based on themes synthesised from existing studies in healthcare, education and peacekeeping across a range of disadvantaged communities, in which co-authors had been involved [3]. We prioritised representation and empowerment of disadvantaged groups within the research process, analysis of the macro-, meso- and micro-level context in which research was conducted, decolonising knowledge to critically review existing approaches and involvement of all stakeholders in co-producing inclusion strategies. In order to build on the knowledge and relationships developed through these existing studies—which covered a range of socially excluded populations—we continued to focus on the countries in which they had been conducted.

Seven separate studies in LMICs focused on excluded populations in relation to health and/or education or local government services and were conducted between 2020 and 2022. Studies identified key drivers of social exclusion and potential scalable interventions that could address these. Two studies specifically focused on addressing the impact of the COVID-19 pandemic on disadvantaged populations.

Each study was designed and led by a research team within the LMIC itself, building on their prior experience of working with socially excluded groups. A range of research methods for data collection and analysis included the use of peer and bilingual researchers, in-depth interviews, policy reviews and multi-stakeholder workshops to co-produce recommendations at each site (see

Table 1. Study methods in each context.

Site	Setting	Research Focus	Research Methods	Research Sample (Total = 385)	Workshop Attendees
BANGLADESH	Four deprived rural, coastal, hill-tract and urban areas; one affluent urban comparison area	Impact of COVID-19 on healthcare and education outcomes for excluded groups (rural, coastal, informal settlements, ethnic minorities)	Literature review to inform topic guides. In-depth semi-structured interviews and focus group discussions in Bengali. Thematic analysis using ‘constant comparative’ method. Stakeholder validation	Patients with children attending school or college (40) Government, private and NGO healthcare staff (doctors, nurses, health assistants (16)). Government, private and religious education providers (15)	Government health and education policymakers, NGOs, specialists in public health, early childhood and equity (10)
KAZAKHSTAN	Three large cities including deprived semi-urban and rural areas	Impact of COVID-19 on education and mental health outcomes for excluded groups (rural, semi-urban, urban, ethnic minorities)	In-depth semi-structured interviews and focus group discussions in Kazakh or Russian. Reflective fieldwork diaries Quantitative demographic analysis. Thematic qualitative analysis. Stakeholder validation	Teachers (30); parents (30); students (28)	World Bank and government representatives, parent/child NGO, teachers and school leaders (7)
KENYA	Five wards in the deprived informal settlement of Kibera, Nairobi	Impact of COVID-19 on public services for young people in Kibera (ethnic and religious minority, gender, deprivation)	Trained peer researchers produced stories, poems and narrative accounts of their experience. Training and support to conduct thematic analysis by experienced academic researchers and NGO mentors. Thematic analysis by peer researchers, supported by study leads. Stakeholder validation	24 Peer researchers aged 18–35 living in Kibera, including disabled young people	Two workshops with local government, health and education policymakers, NGOs, academics and peer researchers (10 + 36)
MYANMAR	Chin and Kachin districts (ethnic minority populations)	Strategies for women to develop leadership skills and positions (gender, ethnicity, age)	Trained peer researchers conducted semi-structured interviews in Burmese. Thematic analysis involving peer researchers, supported by study leads.	32 community members from Paletwa, Chin State (22 women and 10 men). 72 community members from Kachin (69 women and 3 men)	Informal communication with local government and community leads to develop evidence-based strategies responding to report findings
NIGERIA	Six area councils in Federal Capital Territory, Abuja	Inclusion of socially excluded adolescents girls in the national development agenda (gender, ethnic and religious minorities, rural areas)	Peer researchers were trained and supported to formulate research questions and gather data. Semi-structured interviews and focus groups in a range of languages. Thematic analysis with stakeholder validation	Six peer researchers Adolescent girls aged 9–18 (66) Key Informants—traditional rulers and community members with relevant expertise (12)	NGOs, peer researchers, academics, government policymakers, community representatives (53)
NEPAL	Two municipalities in Kathmandu	Inclusion of disadvantaged groups in health management information data collection and analysis (gender, ethnicity, age, religion, socioeconomic position)	Document review—HMIS policy and guidance documents. Semi-structured interviews with Key Informants (service providers, managers, health planners at two health facilities (public and private) in each municipality Thematic qualitative analysis. Stakeholder validation	Public health facility staff (4); private health facility staff (3); data managers (3)	Feedback on findings following presentations at Ministry of Health policy meetings (3)
VIETNAM	Industrial zone in Ninh Binh province	Impact of COVID-19 on migrant workers and their families	Semi-structured interviews and focus groups with migrant workers and Key Informants (zone manager; provincial and district Labour and Health Bureau officers; Labour Union and Women’s Union representatives. Thematic qualitative analysis. Stakeholder validation	Migrant workers (14); Key Informants (7)	Migrant workers; Department of Health, Department of Labour—Invalids and Social Affair, Labour Unions, Management Board of the Industrialised Zone (14)

). Further details of the research methods and participants for each individual study have been published elsewhere [15].

We adopted a ‘best-fit framework synthesis’ [16] to analyse findings across diverse contexts while also testing the conceptual framework that had informed the studies’ development [3]. The analysis focused on drivers and solutions for social exclusion across the different excluded groups in relation to macro-, meso- and micro-level contexts. Coding of data under these themes supported the identification of common issues across the individual studies that were then validated with research leads and other research team members, including peer researchers. We adopted principles for synthesising evidence that would enhance policy take-up of findings, including the involvement of policymakers in the process, drawing on evidence from the full range of contexts and identifying evidence gaps [17].

Significant methodological challenges also related to the COVID-19 context, including restricted access to the most marginalised populations, for whom digital exclusion was a key issue, and short time frames for data collection. Nevertheless, recruiting multiple researchers and using telephone and socially distanced face-to-face methods where possible helped mitigate such exclusion. The pandemic also redirected the priorities of public service policymakers and practitioners to emergency situations, which affected stakeholder workshops in Nepal and Myanmar. Nevertheless, existing links between LMIC research leads and policymakers helped highlight the relevance of inequalities to emergency responses and feedback was obtained from individual policy actors in these contexts.

These methods also helped mitigate challenges posed by a military coup in Myanmar, where a number of peer researchers were nevertheless able to gather data as members of disadvantaged communities.

3. Results

The studies recruited 385 participants across excluded communities, advocacy groups that represented their interests, public service policymakers and practitioners, academics and representatives from national government and the World Bank. Private healthcare providers were included in Bangladesh and Nepal interview samples, as significant stakeholders alongside public health service practitioners (see Table 1). Numerous common factors, outlined below, were found in these studies that created and maintained inequities for deprived populations, women, young people, ethnic and religious minorities, migrant workers and rural communities.

3.1. Policy and Societal (Macro-Level) Exclusion

3.1.1. Political and Legal Context

Public service policies and legislation reflected wider social norms and policies were under-developed regarding equitable service provision. Disadvantaged communities were mostly invisible in terms of data monitoring or representation in public service staffing and decision-making. This negatively impacted policy and service development and allocation of resources. In Myanmar, for example, ethnic minority women were largely excluded from peace and development processes, and, in Vietnam, policy attention was restricted to the management of migrant worker supply, rather than to the health needs of migrant workers. Concerns about the reliability of population denominators and the effectiveness of health planning systems for excluded groups were raised in both Vietnam and Nepal. Although pupil demographic data collection was routine in Kazakhstan, the lack of data on family circumstances and the lived experience of people from disadvantaged groups was common across these contexts.

Local and national government policies further neglected high levels of unemployment in socially excluded populations. In Kibera, an informal settlement in Kenya, peer researchers linked poor opportunities for employment with crime as residents often had few other options to afford food and shelter for themselves and their families. The lack of educational facilities combined with deprivation made young people particularly vulnerable to criminal activities such as theft, drug dealing and violence, both as victims and perpetrators (see Box 1).

Box 1. Policy and societal exclusion.

Looking for work was one of these challenges I had to face. This is compounded by the little education I have. Even the menial ‘jua kali’ jobs are difficult to get since most employers profile people from Kibera as being thieves and aggressive. And when I get an opportunity, it always turns out that it cannot provide for my rent, food and savings. […] Things are worse now with the COVID-19.The small job I had is now gone and the little surplus I had is already spent. At one time, I was arrested past curfew time because I left the casual work I got from the day that was one kilometre away. I was harassed and beaten up. The police even demanded money from me, and forcefully took the 200 shillings I had earned from that day’s work. I could not go to work the next day since I was nursing my injuries and as a result, I lost the job.

Peer researcher, Kibera, Kenya.

We have no support for migrant workers in our province. Every worker is considered similarly. There is no special remuneration policy for migrant workers outside the province. All workers are the same and are treated through the same mechanisms.

Labour Union Representative, Vitenam

“Those who cannot afford resort to the use of what is commonly termed as flying toilets or use open drainages at night. This poses a lot of health risks […] we pay to use a toilet which is expensive for me considering I live with a lot of people.

That wasn’t the only frustration, we could no longer get an income. This meant, no food on the table, unpaid bills, multiple debts and homelessness for some.”

Peer researcher, Kibera, Kenya

“As I do not know how to speak well, I don’t want to solve problems with government laws. The village elders only practice customary laws. I do not dare to go to government offices. I am not fluent in the Burmese language.”

“There are not any rights for women in my village. Women’s jobs are to give birth and do household chores. Women do not have any opportunity to manage financial matters. When the husband dies, they lose the custody of their children. So, I want to change this law and want to get equal rights with men.”

Ethnic minority women from Paletwa, Myanmar

“Even when extrajudicial killings are covered by international reputable organizations like Amnesty International […] no police officer is held to account for the identified crimes. The youth also mentioned that politicians manipulate them to perpetrate violence against political opponents but politicians are not held accountable for this act. […] the person who shall suffer at the end is the youth in the slum and not the police nor the politician.”

Peer researcher, Kibera, Kenya

Social stereotypes characterising such populations as aggressive and dishonest meant that even the most menial jobs were difficult for them to obtain; the resultant poverty was linked to further social stigma as well as to routine police profiling and brutality that targeted young people from these settlements. Crime and violence were described as ‘’normal occurrences” and drug or alcohol abuse was used by many as a means of escaping their stressful living conditions, reinforcing the negative social attitudes that they faced.

Similar policy neglect of housing needs resulted in extremely high and often unaffordable rents for scarce but poor-quality housing in this context, which contributed to overcrowding and gender-based violence. Forced evictions were reported by peer researchers, sometimes linked to urban planning processes that failed to take into account the impact of new developments on existing local businesses and residents.

Policy failure to address the negative impact of COVID-19 in excluded populations was reported in Bangladesh, Kazakhstan, Kenya, Nigeria and Vietnam, where existing inequalities were widened. In each context, people from excluded groups were overrepresented among those with insecure employment and low pay and were very likely to lose their income during the pandemic. They were often not covered by or did not have the required documentation or skills to access government support to mitigate the impact of the pandemic. Migrant workers in Vietnam, for example, were often employed in areas that were ineligible for COVID-19 monetary allowances, reductions to union fees, reduced lending rates or debt rescheduling. A deliberate rationale of keeping the same policies for all workers adversely affected migrants, who were more likely to live in poverty and had less access than other employees to housing, healthcare and education support. There was a failure to recognise that equal treatment for all workers was not equitable when migrant workers faced more barriers to accessing public services and support and had more difficult living conditions (Box 1).

Legal restrictions relating to lockdown were used to legitimise and increase physical violence by police against those living in Kibera. People from this settlement, as well as those in overcrowded and deprived areas in Bangladesh and migrant workers in Vietnam, were also unable to follow government advice relating to social distance and handwashing because of overcrowding and lack of free sanitation facilities or masks, which were not addressed by policymakers. In rural areas of Bangladesh and indigenous ‘tea garden’ regions, awareness of pandemic risk was low and language barriers could also prevent health messages circulated via the media from being understood.

In healthcare and education, the transition to online interactions with public services during the pandemic was easiest for those in urban non-deprived areas as a result of policy decisions. Elite government-sponsored schools in Kazakhstan, for example, were best equipped to make this transition smoothly and quickly because of meticulous planning, prompt teacher capacity development and distribution of digital devices, as well as extensive IT support for both students and teachers. Mainstream schools, in contrast, took much longer to develop staff capacity as a result of lower resourcing and support and less skilled staff. Distribution of devices and pre-paid internet cards to low-income families did mitigate the negative impact to some extent but were inequitable across locations, often omitting those in rural areas. For low-income families and parents with limited education, particularly single mothers, difficulties accessing online education could consequently be insurmountable, reinforcing existing inequities.

3.1.2. Inequitable Resource Allocation

The absence of inclusive policies adversely affected resource allocation for public services, affecting migrant workers, people in deprived areas (particularly informal settlements), locations where ethnic or religious minorities were concentrated and rural areas. An absence of health and education facilities was highlighted by ethnic minorities in Bangladesh, migrant workers in Vietnam and young ethnic and religious minority women in Myanmar and rural areas of Nigeria. Residents in the informal settlements of Kibera complained that all public services were inadequate, particularly waste management, despite good quality services being provided in more affluent areas nearby. The lack of adequate sanitation facilities caused significant health and daily practical problems for inhabitants, many of whom were unable to access even basic toilets and washing facilities without payment. In parts of the settlement where water was available to residents, sewage lines could pass through residential areas and access to sanitation facilities was described as ‘very poor’. Consequently, most villages in the informal settlement were littered with refuse and contaminated with rotting waste.

Most Kibera residents relied on water through standpipes and water vendors, for which they could pay over four times more per unit of water consumed than those living in wealthier areas with city–county water meters. The inability to practice frequent hand washing during the pandemic had obvious public health risks not just for residents of such deprived settlements but for the country as a whole (Box 1).

Similarly in both Kenya and Nigeria, a lack of secondary school facilities in informal settlements and rural areas constrained choices for young people, forcing many to drop out of school because of the costs associated with travel or accommodation outside their local areas, coupled with the financial strains experienced by their parents. As highlighted above, in Kazakhstan, a higher level of resources was deliberately allocated to elite public schools compared to mainstream schools, where the majority of disadvantaged pupils were educated. All teachers received regular salaries; however, those in less-resourced schools were expected to use their own money to spend on stationery that was not provided by their institution.

Other inequities in relation to the provision of housing were highlighted in Vietnam, where migrant workers were ineligible for housing assistance because they did not have the required amount of savings. In Kibera, residents complained that corrupt practices in the allocation of land for housing reinforced inequalities, as only those with additional money to pay officials were able to purchase this. As in other examples above, inequity in resource allocation thus created higher costs for those with the least resources.

3.1.3. Policy Implementation

In some contexts, laws and policies did exist on equitable treatment of vulnerable groups but were often not implemented. In Nepal, health management information systems (HMIS) omitted attention to equitable access, despite the Ministry of Health and Population policy and guidance that aimed to mainstream gender equity and inclusion in social and economic opportunities for all populations [18,19,20]. The publication of guidance on participatory approaches for developing locally tailored health programmes [21] failed to effectively influence health planning and budget allocations, as there was a disconnect between local, district and provincial planning processes. The reporting mechanisms that existed between these different levels failed to ensure that recommendations and disaggregated data submitted by local health facilities were used by the municipality-level planners who allocated healthcare resources.

Similarly, the Nigerian Child Rights Act 2003 continued to fail in preventing the exploitative labour of children [22] in the rural and minority communities with which we engaged. In Myanmar, while women could in theory access national statutory protections, the cost, delays, language barriers, lack of support to deal with the complex processes involved and mistrust of government institutions prevented such access. In contrast, regional ethnic customary law was inexpensive and accessible but often failed to protect women in most areas of their lives, including domestic violence, divorce and child custody, inheritance and access to employment, education or financial resources. These constraints on access to national law meant that women were often denied their legal rights and forced to continue in situations that maintained and reinforced the inequalities they experienced (Box 1).

Political and social instability could similarly reduce or remove the legal rights and protections of vulnerable groups. The formal structures of a national peace process in Myanmar were disbanded following the military coup in 2021 so that routes for women to influence policy and practice were much more restricted than previously. Even prior to the coup, few women were able to participate in the process, which was dominated by male negotiators, as sociocultural norms prevented women from adopting leadership roles, despite the disproportionate impact of ethnic violence on women and children.

Similarly, social tensions in India, resulting from divisive government policies affecting religious minorities meant we were unable to proceed with the planned research. Although our NGO partners in Myanmar were at risk of being targeted for supporting minority rights, they persevered with the project as a result of their close links with affected communities. In contrast, partners in India had closer links with the government, from which their organisation received funding, and the political risks of involvement, as well as apparent support from within the organisation for government policies, appeared to be influence their withdrawal from the project.

3.1.4. Participation and Representation

The absence of effective policies to address social exclusion was maintained through the non-representation of people from disadvantaged backgrounds in positions of power. Women, those from deprived or rural areas, young people, ethnic and religious minorities and migrant workers were routinely excluded from decision-making processes that affected their lives. In Nepal, programmes at the municipality level that aimed to target disadvantaged groups failed to follow detailed guidance on reaching out to these groups to address community health issues [21], replicating social exclusion even in activities meant to benefit these groups. In Vietnam, migrant workers faced difficulties in becoming involved with the Labour and Women’s Unions that could influence working conditions because of the need to move around in search of employment. Labour Union representatives, who were employees of the company, were unwilling to advocate for migrant workers, despite being the key mechanism to pass on comments and complaints to managers.

Intersectional exclusion compounded disadvantage and was also not addressed by relevant public service policy or practice. Both within their own communities and in wider society, lower social status constricted opportunities for women and young people in multiple ways; for example, in Nigerian rural settings, families could give higher priority to funding boys’ education compared to that of girls, whose contribution to families was expected to end after their marriage. Fears that cultural values would be undermined in school settings, where staff did not reflect local student populations, were also particularly acute in relation to adolescent girls. In Nigeria, Kenya and Myanmar, poverty and the absence of trusted local institutions could thus exacerbate existing constraints on education for young people, particularly girls, from ethnic and religious minorities.

In Myanmar, education for women could be seen as unnecessary by some community members who stereotyped women as naïve and lacking intelligence. Those who were supported through this research to take on leadership roles in peacemaking could face pressure from their families, including other women, not to become involved in community affairs, and could lack support with household and caring responsibilities. A lack of policy attention to addressing these issues prevented women and young people from minority backgrounds from challenging structures and attitudes that discriminated against them within their own communities and in wider society. Education and support that could help develop capacity for leadership roles were thus restricted at community, institutional and policy levels.

3.1.5. Exploitative Practices

The lack of policy in relation to social exclusion enabled those in positions of power to exploit the disadvantage of communities to create further benefits for themselves. Young people in Kibera described a regular cycle of being employed by politicians during each election to harass and commit physical violence against political opponents. Many who had no other income felt forced to commit such crimes and could themselves be injured, arrested or even killed in the process, while others were drawn into drug abuse as a coping mechanism. Whilst politicians were not called to account for their use of gang violence, extra-judicial killings by the police and enforced disappearances were described as part of the human rights violations to which the community was subjected on a regular basis, with no protection from abuse of power (Box 1).

The exploitation of labour was common in many of the study settings; for example, adolescent girls in Nigeria, those in informal settlements in Bangladesh and Kenya and migrant workers in Vietnam had limited employment options involving low wages and little or no protections for their health or family welfare. Urban planners in Vietnam urban planners focused solely on managing the availability of migrant workers for urban industry. In Nigeria, young people were sometimes drawn into child labour or prostitution in order to cover daily living costs or to pay for their education. Political and societal acceptance of these restrictions constrained individuals to choose between severe deprivation and undertaking criminal activity or leaving education to support their families.

3.2. Organisation (Meso-Level) Factors

A disadvantage was further created and reinforced for these communities through inadequate access to public services as a result of policy neglect, unfair resource allocation and non-representation amongst service staff.

3.2.1. Access to Public Services

Public service staff did not reflect the communities they served and discriminatory practices by staff were commonly reported. In Bangladesh, the absence of health services in informal settlements forced residents to rely on NGOs, which were unable to deal with critical medical conditions.

Apart from the distance of health facilities, which many excluded groups experienced, discrimination and even violence by healthcare staff towards Kenyan women from such settlements was reported during antenatal care and labour as a further disincentive to seeking help from where they lived. Young people from Kibera’s ethnic and religious minorities similarly complained about being profiled and beaten by the police because of their residence and clothing. Enforced curfews during lockdown could compound such discrimination, as residents of informal settlements reported being delayed en route to healthcare facilities during childbirth or beaten by police, with their earnings stolen on returning from workplaces located at a distance from the settlement (Box 2).

Box 2. Public service factors.

“Anytime we saw a police vehicle we used to run, not because we had or have anything to hide but simply because they threatened us that they ‘will come for us’ without any explanation about our wrongdoing.”

Peer researcher, Kenya

“The doctors slap you if you do not cooperate. The midwives are ill intentioned. The sanitation facilities are too dirty. If you do not guard your baby well, it can be exchanged or even taken away. Some women just give birth on the floor as the nurses look on without offering any help.”

Vivien, Peer researcher, Kenya

‘‘Getting treatment is always tough for us. We need to buy our own medicine due to lack of supply in the health centres. It is a burden for us. The situation is the same both before and during Corona.’’

Indigenous respondent, Bangladesh

“Responsibility [for data input and analysis] is not assigned. When I came here, there was no one who knew how to use computers.”

Local healthcare provider, Nepal

In Vietnam, the Labour Union, a civil society organisation that implemented government policy, made no allowances for the poverty experienced by migrant workers in their housing support scheme, eligibility for which required proof of ownership of the existing land estate. Similarly, the Women’s Union restricted support for migrant women to obtain a loan to start their own businesses by requiring permanent registration documents. These same unions gave charitable donations or gifts during national festivals to migrant workers, however, and recognised they had “difficult lives”; some companies also provided travel or housing costs over the holiday period on proof of original residence. Rather than advocating for migrant workers to have better long-term conditions and equitable support, the Labour Union and employers thus helped to maintain the disadvantage they experienced.

High out-of-pocket expenses relating to travel, medication, loss of work or internet costs were an additional barrier to accessing public healthcare facilities for migrant workers in Vietnam, and, for all excluded groups in Bangladesh, the better quality services from private healthcare providers were not affordable for groups experiencing exclusion. Access to primary care and essential services such as maternal and child health and immunisation provision were all affected by these issues; further restrictions relating to the pandemic exponentially increased access problems, as excluded groups had little alternative to public healthcare, particularly for acute care needs (Box 2).

In Vietnam, despite the high levels of deprivation they experienced, migrant parents were often obliged to use private schools with higher fees for their children, as they could not access public education without household registration documents, which required long-term residency. Children’s education was also restricted by their parents’ mobility in seeking employment opportunities at new locations [23]. No provision was made by education providers or employing companies to address this deficit. Fears about health risks during the pandemic further reduced migrants’ access to healthcare and led to additional housing costs, as migrant workers stayed in rented accommodation rather than travelling home during public holidays to reduce the risk of transmitting the virus to their families.

Deprivation and digital exclusion were common amongst all the excluded groups with which we engaged and the pandemic created further difficulties through the need for equipment and a stable internet connection to access online education and healthcare. In Kazakhstan, both teachers and students could be affected, severely disrupting children’s progress, particularly in rural areas. Some parents were consequently forced to send children to live with relatives in urban locations so that they had the opportunity to study. The digital divide also affected access to healthcare and to public health guidance on COVID-19 in Vietnam, as well as in the informal settlements, ethnic minority locations and coastal and rural areas of Bangladesh, where provision was already traditionally low, exacerbating existing inequities.

The impact of school closures in Bangladesh disproportionately affected those from lower socioeconomic backgrounds, who were overrepresented amongst excluded populations. Whereas affluent urban families reported access to digital and interactive forms of education, children from excluded groups were limited to lessons via television broadcast or self-learning methods, with very high rates of dropout from any form of education.

In Nigeria, ethnic and religious minority adolescent girls in rural areas were unable to access secondary-level education without incurring costs for travel and accommodation in locations distant from their families. Their lack of capacity to cover these costs could often drive them to prostitution or dependence on men who could help them financially. This reinforced negative attitudes about educating girls within minority communities, so it was difficult for girls to access education and for the communities to feel this was possible in a safe environment that valued their cultural norms.

Most of the above issues constraining access were not visible through the reporting mechanisms used to monitor healthcare services. In Nepal, standard reporting formats for both public and private providers at local levels did not report population differences apart from gender. Sometimes even this category was absent in private facilities, where data were often incomplete. More detailed data, such as patient ethnicity and education level that might be collected by local public facilities, were not reported to municipalities and there was no evidence of data on socioeconomic status being collected. Data collection and analysis for the management of health systems was further constrained by limited capacity to input and analyse data and a lack of internet access or dedicated staff for this purpose. This was consequently a superficial exercise, with reduced amounts of data at higher levels, that had little to no impact on equitable health planning or budget allocation processes. At the national level in Nepal, data were only disaggregated by province with few exceptions (Box 2).

3.2.2. Service Outcomes

Lower resource allocation for mainstream public schools in Kazakhstan had an impact both on the quality of teaching compared to elite schools and on the capacity of teachers to adjust to online teaching. Many staff members, particularly those who were older or living in rural areas, had less capacity, guidance and training to adjust to online education during the COVID-19 pandemic than those in better-resourced institutions. Parents often felt compelled or were encouraged by teachers to pay for private tuition to make up for the low quality of education that was being delivered online, increasing financial pressures on low-income families. In Bangladesh, similar problems with school infrastructure and supplies as well as staff capacity were highlighted during the pandemic, particularly affecting disadvantaged groups.

3.3. Community/Individual (Micro-Level) Outcomes

Study participants described how poor access to education, healthcare, sanitation and other public services, which was exacerbated during the pandemic, created a self-reinforcing cycle of exclusion from positions of influence that could help change their lives (Box 3).

Box 3. Community and individual level outcomes.

[my daughters] used to go to school by themselves. Now that school is closed they are helping with work. We are now thinking of marrying them off.

‘Monira’, 26 yrs, indigenous mother, Bangladesh

I joined a small group of young men who made me feel that I belong to the street I lived in and protected me from other boys whenever fights began. To fit in, I had to put on dreadlocks, put on a tattoo and wear casual clothing. In order to commit some of the heinous acts without feeling guilty, most of the youth started abusing drugs. Unfortunately, these young people were injured, some died in the process and others were arrested.

Francis, Peer Researcher, Kenya

As artists we could no longer recite our poems, sing our songs or dance to our tunes.

Peer researcher, Kibera, Kenya

in our [rural] area there are parents with extremely low social status; the whole family lives in one room. It was exceedingly difficult for them as they eat, sleep, and live in the same room.

Female teacher, Kazakhstan

Oh, so many people died in our village due to this pandemic. At that time, people were terribly stressed.

Teacher in Almaty, Kazakhstan

“those parents have an average level of education; it was too difficult for them to work with the computer.”

Female teacher, Kazakhstan

“In my opinion, the role of migrant workers is important and they should know their rights and how they should act to fully access their rights. But they themselves do not have enough knowledge and they do not understand their rights or have the skills to raise their voice. If you do not cry, your mother will not feed you.”

Labour Union representative, Vietnam

As at the macro- and meso-levels, the COVID-19 pandemic exacerbated existing dynamics of exclusion within excluded communities. Already high levels of unemployment increased during the lockdown as many individuals lost their jobs when businesses closed down or interactions moved online.

3.3.1. Deprivation

Peer researchers in Kenya wrote about their experiences of growing up and living in poverty and their efforts to remain resilient and resourceful despite a lack of food or clothing and footwear. People with dependents, including young carers of elderly relatives, often lived “a life of hand to mouth” [15]. Many tried to sell small-scale services and goods to make a living, but others could find themselves forced to join local gangs as the only source of income on which to survive.

At the same time, housing could be very expensive, with eviction being a constant threat, and those living in the informal settlement could even have to pay for basic sanitation such as toilet facilities. In this context, the pressure to earn income from any available source could be immense. Even those with work were often unable to fully cover their daily living costs, let alone save for future needs.

Participants felt they were excluded or forgotten by policymakers on issues such as housing and education but were deliberately targeted by law enforcement services, who regularly threatened or committed violence against them for no explained reason. Overall the peer researchers conveyed a sense of wasted potential and a daily struggle to survive.

Oscar, a peer researcher in Kibera, described how he became a voluntary football coach to young boys and girls using a football pitch that was affected by the lack of sanitation facilities. One corner of the pitch was used as a waste dump and the whole area became a health hazard when a sewage pipe close by burst, spreading waste on most of the pitch and resulting in two-thirds of the children falling ill. The restricted opportunities for education and social activity exemplified by this situation were linked to young people becoming involved in crime.

Similar reports about the impact of deprivation were received from adolescent girls in rural areas of Nigeria and migrant workers in Vietnam. Deprived populations were exposed to more health risks as a result of strenuous and unhealthy working conditions and struggled to pay for healthcare where this was not free, resulting in reliance on herbal remedies rather than formal healthcare services.

Access to education was also prevented by unaffordable school fees or lack of educational facilities as a common factor in Kibera, rural areas of Nigeria and for migrant workers in Vietnam. Children’s education could often not be prioritised and the need for basic necessities such as food and shelter was linked to child labour in Kibera. Young people in Kenya and Nigeria often mention family income needs as a reason for seeking work and giving up education.

Access to secondary education for deprived communities in rural areas of Nigeria involved spending more to access education than their less deprived urban counterparts, as schools were located at a distance necessitating accommodation and living expenses. As highlighted above, adolescent girls from rural areas whose families supported them to attend these schools often found themselves in financial difficulties and would either drop out of school or “meet boys for money” in order to cover rent and food to continue their education. This, in turn, could lead to unwanted pregnancies that reinforce communities’ unwillingness to educate girls beyond primary school.

Social inequities that made it more difficult for deprived communities to observe social distancing or isolation as a result of overcrowded accommodation or working conditions increased their exposure to the risk of contracting and dying from COVID-19 in Bangladesh, Kenya, Nigeria and Vietnam (Box 3).

In informal settlements and rural areas, many participants felt devastated by the destruction of lives and livelihoods. The fear of catching the illness and seeing neighbours and relatives die of COVID-19 greatly impacted the psychological and mental well-being of study participants in Kazakhstan, for example, as well as their children. Most respondents from Kibera had lost their jobs during the pandemic, causing considerable financial and mental hardship (Box 3).

Participants in this context also reported that some residents had lost employment as a result of not being able to afford COVID-19 tests, which cost double their monthly salary. This contributed to a domino effect of other stressful consequences, such as forced evictions, overcrowded homes and inability to pay for adequate food or essential commodities. In Vietnam, financial consequences involved adaptations to daily spending, such as skipping or reducing meals, prioritising children’s food or selling assets, to provide additional income. Anxieties expressed by migrant workers about the impact of the pandemic are likely to have been compounded by restrictions imposed after the study ended when migrant workers were forced to remain on company premises throughout the lockdown period and unable to travel home.

The confined living arrangements that affected deprived families during the lockdown created tensions over resources and within relationships. In Kazakhstan, although lockdown reduced spending on travel and uniforms, families with multiple children could struggle to manage the various needs for computer access. In Kenya, an increase in domestic abuse during lockdown was reported by some study participants (see Intersectional Disadvantage below).

3.3.2. Capacity and Skills within Communities

Lower capacity in disadvantaged groups to adapt to the greater reliance on technology during the pandemic, in addition to lack of funds for equipment and data, disproportionately affected access to healthcare, education and employment. Migrants in Vietnam reduced their support seeking for health, which could have life-threatening consequences. Despite experiencing illness symptoms, some people avoided contact with the health system by choosing to remain at home or take medicine without prescriptions. In some cases, symptoms deteriorated and patients were rushed to hospital emergency units, which could result in a higher economic burden for both patients and the health system and require longer treatment times [23].

Education providers in Bangladesh and Kazakhstan reported that students from lower socioeconomic backgrounds and rural areas suffered more as a result of the pandemic’s consequences than other students and, in Bangladesh, a significant number of these students were forced to drop out of school, increasing the number of child marriages and perpetuating the vicious cycle of low employment in their communities. In Kenya, the pandemic made worse an already severe situation in which 43% of girls and 29% of boys in Kibera have not attended school at all [24]. Similarly, migrant workers in Vietnam reported that they stopped sending their children to school during the pandemic because of reduced incomes.

Parental involvement was pivotal for the effectiveness of online education in primary school students and it was clear that parents found this extremely stressful. Those who could not afford private tuition had to learn subjects such as mathematics themselves in order to support the learning of their children. Without the sustained support of parents, primarily mothers, ensuring children’s progress through effective online teaching was challenging, if not impossible. Students whose parents needed to be at work or lacked either subject matter knowledge, teaching or digital skills were likely to fall further behind students who had access to better resources.

Although mechanisms sometimes existed for study participants to complain about the issues they faced, these were often not used because of unfamiliarity with the processes involved. For example, migrant workers lacked the knowledge and skills to participate in planning processes via the Labour Union and Provincial People’s Committee, which in turn appeared to think it was the migrant workers’ own responsibility to initiate action (Box 3). As highlighted earlier, these organisations made no effort to reach out to this group, despite being aware of the difficulties they faced and their specific remit to advocate for workers.

3.4. Intersectional Disadvantage

Multiple layers of exclusion relating to gender, age, ethnicity, religion and geography created additional impacts to those described above for women and girls, young people and members of ethnic or religious minorities. Many participants fell into a number of these groups, compounding the adversity to which they were subjected. Dynamics at macro-, meso- and micro-levels of society further constrained their ability to escape from deprivation and violence or even challenge the status quo.

3.4.1. Gender and Age

Gender-based violence (GBV) was reported in many of the study contexts and was linked to a number of factors: social acceptance of GBV and lack of GBV policy or implementation; overcrowded housing; financial pressures; and substance abuse and child marriage. These factors often accumulated and could be linked to mental health problems in excluded populations that increased the high rates of risk of domestic violence and rape for women. Unwillingness by parents to report crimes where close relatives were involved, along with the costs and delays involved in legal action and threats from perpetrators to survivors, compounded the lack of action on GBV. The risk to mental health in young people who witnessed or were subjected to such abuse and the heightened risks to children of online abuse during the pandemic were also highlighted as key issues to be addressed.

In Myanmar, women who were involved in activism relating to GBV complained of being targeted by the military junta and excluded from peacebuilding activities that could help change the social norms that perpetuated this kind of violence. The junta itself had a history of GBV and was seen by peer researchers as threatening women’s rights and the organisations that documented this issue.

The federal nature of Nigeria was seen as a major impediment to the implementation of laws that could prevent violence and social exclusion of women and children, as legal protections that existed at the national level were not adopted by many states within the country. This included provisions such as the Trafficking in Persons (Prohibition) Law Enforcement and Administration Act 2003 and the Violence Against Persons (Prohibition) Act, 2015. Some Nigerian states did have legislative measures to secure the rights of women and children, but these were limited to particular regions [25].

The existence of policy on GBV and gender equality in Nepal was similarly not routinely implemented. For example, although hospital outpatient registers collected data on patients who had experienced GBV, this was not reported to those responsible for health planning and service development. Women were also more likely than men to have caring responsibilities and those living in poverty were under considerable pressure to also provide an income, especially if they were single mothers, sometimes going without meals themselves in order to feed their children or elderly relatives. Children brought up in such conditions were often forced to work to help raise money for basic needs, particularly if healthcare costs needed to be covered.

Where education involved costs, it was common to find girls who were not educated beyond the primary school level because of the costs or distance involved in accessing secondary education, increasing the chances of marriage at a young age. As highlighted earlier, Nigerian girls who were supported by their families to attend secondary schools were vulnerable to exploitation because of their precarious financial situation.

Peer researchers in Nigeria and Myanmar highlighted that, in the context of a patriarchal society, girls would not be empowered to speak up, challenge or make demands on the existing systems either within their communities or in society more broadly without education. For women who did take on political roles in Myanmar, their engagement could be further constrained by a lack of childcare support, the competing demands from domestic responsibilities, an overall lack of public safety and the threat of GBV that such roles could attract.

Adolescent girls in the study contexts were often not involved in key decisions that affected their lives relating to education, work and marriage. Girls in rural areas of Nigeria pointed out how important it was for them to be involved in decisions concerning their education, marriage or trade in order to enable them to commit fully to these. The links between community and political dynamics were also recognised—the participation of adolescent girls in decision-making at the family and community level had implications for future participation in political and leadership roles and broader societal development (Box 4).

Box 4. Intersectional disadvantage.

Without being educated, you cannot be seen or known in the community…when a female is educated, she can express herself and she is well recognized in the community. It is important for the development of the communities… when I want to marry, my husband will respect me, and I will be able to express myself and be respected.

Adolescent girl from Kuje, Nigeria

I don’t follow [two daughters’] education much. They used to go to school by themselves. Now that school is closed they are helping in work.

Indigenous mother, Bangladesh

The customary laws for Khu Mi Chin women make women to suffer and they do not have any rights. Women are nothing. I faced domestic violence many times and my parents didn’t do anything about that for me because of the bride price they took when I married. My husband is having an affair though I cannot ask for punishment because I am afraid that I will not get guardianship to my children and I cannot live away from my children. I am suffering domestic abuse.

Female participant, Myanmar

During the pandemic, some adolescent girls in Kazakhstan, Bangladesh and Kenya faced additional barriers to education from the lockdown, as they were expected to contribute to household chores through this period with negative consequences for their learning. Women in Kazakhstan and Bangladesh could also struggle with the expectation, which fell mainly on them, to provide educational instruction to children during lockdown. Those who needed to work or had limited education themselves were unable to meet this expectation effectively, so children did not receive the support they needed for online learning. Existing disadvantages were thus transmitted to the next generation within these communities (Box 4).

3.4.2. Ethnicity and Religion

All studies included ethnic or religious minority population groups, many of which were overrepresented in informal settlements and/or deprived or rural areas. Evidence from each context highlighted the reasons behind worse outcomes for these groups from public services such as failure to analyse routine data on ethnic and religious minorities in Nepal or address their needs in health planning processes. Language and caste discrimination could influence access and outcomes in all study contexts.

Women from ethnic or religious minorities could experience additional gender disadvantage through a lack of access to legal frameworks available to other women. In Myanmar or rural areas of Nigeria, women living under customary laws described these as a reason for gender violence and exclusion from decision-making processes that affected them. As highlighted earlier, the existence of parallel legal frameworks in Myanmar could fail to protect women in relation to GBV, inheritance rights and child custody (Box 4).

Politics in Kibera, and Kenya more widely, were seen to involve mobilisation of ethnic groups and ethno-regional patronage linked to historical grievances and political violence, with residential areas often divided along ethnic lines. Both religious and ethnic divisions could also adversely influence perceptions of minorities in Nigeria, creating and reinforcing social exclusion. In Myanmar, a combination of history, culture and politics were seen to have resulted in a fractured society linked to communal violence involving groups that are collectively termed Ethnic Armed Organisations.

Allocation of resources for public services in areas where ethnic and religious minorities were concentrated was shown to be deliberately unjust in many of the pilot reports, constraining access to healthcare, education and sanitation facilities and creating costs for these services that were not present for the more powerful populations in each context. At the same time, these populations were overpoliced and coerced in Kibera, Myanmar and Vietnam, indicating that public service resources were focused on controlling rather than supporting these groups.

3.5. Strategies for Inclusion

Recommendations to reduce the public service inequities identified were drawn from co-production workshops involving policy, practice and community stakeholders across the diverse contexts of the studies. The following strategies were considered to have the potential for scaling up across the relevant country context.

3.5.1. Visibility of Public Service Inequities and Community Needs

A key first step was to make inequities visible and raise awareness about the urgency of necessary action. Training and equipping public service staff to generate, analyse and report good-quality evidence about public service access and outcomes for diverse social groups was promoted. Routine data collection was recommended, along with periodic surveys on key health indicators that were disaggregated by gender, age, ethnicity, religion, migration status, education, occupation, disability status, location and other relevant stratifiers. Equity-related outcome measures integrated into public service programmes and interventions through a combination of supportive and regulatory mechanisms would increase accountability and transparency. The involvement of private providers in routine data collection and reporting was also seen as essential where they contributed significantly to the sector involved.

Quantitative and qualitative research evidence, to supplement these data and to identify group-specific as well as intersectional disadvantage, wasalso seen as empowering to these groups, particularly participatory research that is directly linked to policy and practice development.

The use of such data in planning processes was seen as key to equitable allocation of resources. Identifying good practice examples, establishing leadership, accountability and independent oversight were considered essential for equitable planning processes and resource allocation.

3.5.2. Representation of Excluded Groups

The need to employ and involve those from excluded groups was seen as essential to developing inclusive and detailed solutions to current inequities and improving awareness and critical thinking about inequity, both in wider society and within excluded communities themselves. Financial and practical support for disadvantaged groups to contribute to collaborative decision-making would help ensure that barriers to equitable participation, such as involvement costs (including lost employment) and childcare, were reduced. Capacity building and involvement were closely tied to better access to education and employment opportunities.

Involvement in local, regional, national and international development programmes for public services was seen to require transparent planning processes and regular channels of communication on a long-term and equitable basis. Independent oversight of police authorities to ensure policing enhanced community security and safety rather than diminishing this was seen as essential. The routine involvement of existing NGOs that advocated for and empowered excluded communities, such as Kituo Cha Haki Kibera, a youth-led community-based grassroots organisation in Nairobi, was promoted. This organisation employed community-based paralegals and drew on evidence-based research to challenge human rights violations. Women-led organisations in Myanmar played a similar role in improving family and community outcomes.

Public services were urged to make specific provisions for the representation of those experiencing intersectional disadvantage, such as women and young people from ethnic and religious minorities. Identifying, supporting and promoting positive role models and mentors who could promote routes out of deprivation and social exclusion was seen as an important aspect of the increased visibility needed.

3.5.3. Multi-Sector Partnerships

Approaches that crossed multiple sectors were seen to require strong leadership that could address the discriminatory social context in which public services operate and the complex intersectoral mechanisms that create and maintain inequity.

Linkages between data systems in different sectors were promoted to obtain a fuller picture of unmet needs and inequities. Identifying good practice examples of interventions crossing sector or service boundaries that had been evaluated and scaled up was also promoted. In Nepal, for example, a system of social audit to monitor service use and ensure accountability in primary healthcare along with the use of subsidies for vulnerable populations to access hospital care were highlighted. Hospital-based One-stop Crisis Management Centres also provide coordinated services across diverse sectors to survivors of gender-based violence [26].

3.5.4. Targeted Support

Alongside generic recommendations, specific policy support for increased agency by women, young people and ethnic and religious minorities on matters that directly affected them was promoted. Representative Community-based Management Committees were recommended in Nigeria as an acceptable way for elected community members to drive development issues forward within their communities and to formally engage with public service providers and policymakers. The need for adequate support and resources to run these committees was highlighted.

Participants recommended that public service development particularly addressed issues affecting women, young people and ethnic or religious minorities in deprived areas, including gender-based violence, access to formal legal processes, police brutality, substance abuse and public service discrimination. Such initiatives needed to link to programmes on social sustainability and economic development.

A need to ‘build back better’ following the COVID-19 pandemic was highlighted to mitigate the greater impact on excluded groups and transform current policy and practice so that public services worked well for everyone.

3.5.5. Focus on Education and Economic Inequities

Increased access to free good quality education and employment opportunities were seen as key routes out of deprivation and exclusion. This would involve redistribution of education funding and resources to benefit disadvantaged schools and marginalised learners and investment in digital inclusion education and devices for staff, as well as parents and students from excluded backgrounds. Involving universities in capacity-building initiatives, work–study programmes and research to generate and evaluate policy and practice initiatives were recommended.

Financial support for families to eradicate child labour and promote greater employment opportunities through multi-sector initiatives was prioritised. For example, Kazi Mtaani in Kibera, Kenya, employed youth and helped reduce crime in the informal settlement during the pandemic when most residents had lost their jobs and closed their businesses [27]. Participants recommended extending this model to public services and ensuring salaries were sufficient to meet employees’ living costs.

4. Discussion

We believe this is the first study to explore solutions to social exclusion across public service sectors for diverse disadvantaged populations at macro-, meso- and micro-levels. This focus has previously been highlighted as an evidence gap in LMIC contexts [2,3].

Our findings can be understood in relation to theories of social exclusion and structural violence outlined earlier, involving processes that generate and maintain social hierarchies in the context of unequal power relationships [7,8,9]. The theory of social pain also helps our understanding of how such processes can lead to violence that exists within these communities [11]. Findings confirm that these processes are rooted in public service institutions [2,5] and demonstrate that common mechanisms operate across a range of such services to adversely affect diverse disadvantaged populations in similar ways.

4.1. Social Exclusion, Structural Violence and Social Pain

The kinds of structural violence detailed in our studies ranged from the direct and indirect experience of verbal abuse, racism, coercion, restrictions, criminalisation, and physical and sexual violence to the consequences of these injustices such as mental and physical illness and premature death. Such violence is the means by which inequitable access, representation and outcomes are maintained within public services.

At macro- and meso-levels, symbolic and actual violence in public services was reflected both in the neglect of these groups—in terms of where services were located and the absence of adequate or appropriate provision—as well as in the punitive targeting of excluded communities through harmful policies and practices. Neglect was also manifested through the invisibility of such groups in public service data, service staffing and representation in decision-making. Contrary to international guidance on addressing the determinants of health, our findings confirm that members of disadvantaged populations were often treated as second-rate recipients of services, whose needs and human rights could be systematically denied, rather than as equal partners in planning how public institutions should be run [28,29].

These discriminatory processes in multiple public service sectors combined to create an everyday struggle for survival, affecting all the disadvantaged groups in our studies. Policy failures in education, local government and policing, for example, were linked to high levels of unemployment and stress that were in turn linked to crime, drug abuse and domestic violence. Neglect and punitive targeting by public services also helped to maintain the lower capacity and skills that could prevent people within disadvantaged communities from challenging the dynamics that created and reinforced these inequities. While the privilege of those in affluent areas was maintained through more resources and support, as well as cheaper and more adequate provision, excluded populations struggled with higher out-of-pocket costs to access basic services and higher exposure to health risks and employment insecurity. Our findings confirm that under highly stressful conditions such as the COVID-19 pandemic, existing social inequities, including those rooted in race or religion, have the potential to resurface intensively [30]. Excluded communities were less able to mitigate the impact of the pandemic on their finances, health or education and were penalised through even lower access to public services and punitive policing during lockdown that led to further financial loss.

At the micro-level of individuals and communities, social pain theory provides a framework through which to view violence within excluded communities, such as gender-based violence (GBV) and criminal activity. Within this framework, social exclusion is conceptualised as a primal threat to well-being, provoking aggression and other emotional responses similar to those caused by physical threats. GBV and child abuse were linked to the pressures of domestic life in excluded communities and the high levels of mental stress caused by structural violence. Inadequate legal and policy frameworks also failed to challenge harmful social norms affecting women and young people. Levels of crime and violence were thus an indicator of social and political inequity [12,13,31] and community-level trauma and stress [32,33]. Our findings show that social crime and violence were often initiated and supported by those in power as a means of maintaining their privilege. Gang violence, drug abuse, GBV and prostitution within excluded groups also acted to reinforce negative stereotypes used to disempower these populations and to legitimise even more punitive public service responses, including police brutality.

The economic costs of exclusion and restrictions on capacity development were apparent at individual, family and community levels, through pressure to work instead of being educated, the struggle for everyday necessities and illness and premature death. These dynamics reflect not only the physical and emotional pain of exclusion but also the wasted human potential and undermining of sustainable development [12,13]. The lack of support to thrive has enormous financial costs to economies in terms of lost productivity [12] and a higher healthcare burden related to crisis management [23]. A further cost is the reduced capacity to mobilise against inequity, precisely because of the restricted access to resources and constant social pressures on which participants had to instead focus. The negative impact of such structural violence and social exclusion was intergenerational, passed on to children through mental and other health impacts and the transfer of disadvantage via parents’ lower capacity to support their children’s education and the family’s housing and other health-related needs.

4.2. Macro-Level Interventions

Workshop recommendations called for macro- and meso-level interventions in four key areas, providing a policy and practice framework for social inclusion and the reduction of structural violence. This framework may be understood in relation to addressing the sociopolitical context; increasing transparency and accountability; leadership; and capacity building.

Changing the political and social context was considered essential in recommendations about how to develop, implement and sustain progressive policies. In places such as Myanmar and India, this context has led to communal violence with extreme consequences, experienced mostly by excluded groups but also causing social and political instability for wider society [32]. Reducing the social acceptance of discrimination that leads to such extreme conflict requires initiatives that address cultural norms, historical grievances and ethno-regional patronage, all of which influence the unjust allocation of resources and constrained opportunities for disadvantaged groups [3,13].

Greater transparency and public service accountability, particularly in relation to political governance and its enforcement through policing, has the potential to remove the impunity with which such discrimination can occur, increasing trust, democracy and efficiency in governance and legislative support [13]. Action to remove the privileges to which dominant groups may consider themselves entitled is likely to require engagement at an emotional level; leadership from within dominant groups themselves is most likely to reduce resistance from these groups and undermine their acceptance of structural violence that affects excluded populations [34].

These macro-level recommendations lay the foundations for social inclusion that transforms current systems, as opposed to the inclusion of disadvantaged groups within inequitable systems that maintain existing privilege. Without such systemic transformation, the violence propagated by public services is unlikely to change [35]. Adopting these recommended approaches to develop a shared vision and iterative policy process, however, has the strong potential to disrupt such violence and to benefit the collective well-being of society as a whole [35,36,37].

4.3. Meso and Micro-Level Initiatives

Alongside macro-level restructuring, workshop recommendations promoted the voice of disadvantaged groups through their involvement in public service staffing and decision-making. Policy implementation in this area requires regulation, incentives and resources to support staff training, representative employment and effective leadership that is needed both within public services and in excluded communities themselves [13,38,39]. Workshop recommendations recognised that public service inequities had cumulative impacts and interacted with each other, with implications for how ways forward might be achieved. Increased capacity for leadership within excluded groups was understood to require improved access and outcomes in education and healthcare, particularly during early childhood. A recommended focus on education and employment reflects their fundamental importance to capacity building that would support the increased influence of excluded groups through improved economic and social status and early prevention of gender inequalities affecting boys as well as girls [40]. The wider literature on race equality suggests that career progression in public services would also need to be addressed to ensure influence at the highest levels [38].

Equitable employment opportunities are also linked to the macro-level issues highlighted above: societal awareness, legal and social protections and visible role models have the potential to reinforce each other and support social inclusion in public services and society more broadly, reversing current dynamics [38].

Our findings confirm that multi-sector and multilevel initiatives are simultaneously needed to create the pathways through which such outcomes might be achieved [2,41]. The call for such approaches has also been made by other studies that focus on excluded groups not included in our research, such as people with HIV, refugees and those who misuse drugs [42,43,44,45,46]. This suggests that the overarching frameworks to address social exclusion and structural violence presented in this paper have global relevance for disadvantaged populations. These generic frameworks may need to be combined, however, with evidence about the specific needs of each group in terms of social protections, anti-discrimination measures and supports needed [47,48,49,50,51]; training and education materials [52]; ways of reducing the criminalisation of communities [45,53]; and learning from the experience of non-governmental organisations [54,55]. Further research to evaluate existing initiatives and explore areas of uncertainty is also needed to support the scale-up of interventions, particularly in relation to implementation in specific contexts [42,56].

5. Conclusions

Our findings provide detailed insights into how inequalities are created and maintained at macro-, meso- and micro-levels for the populations involved in our studies and how these processes may reinforce each other to maintain social exclusion. Recommendations developed with the input of all key stakeholders take account of the social and political context in which studies were conducted and are grounded in existing evidence to suggest potential solutions that contribute to policy, service and social development. Follow-on research to test the implementation of these suggested ways forward and to evaluate interventions in practice is now needed. Particularly urgent is the need to evaluate interventions addressing the “poverty penalty’’; the political aspects of exclusion; the cyclical reinforcement of exclusion through structural abuse, neglect and violence; and intersectional disadvantage experienced by young people and women from disadvantaged populations. As highlighted by workshop participants, such interventions would provide policy and practice frameworks for social inclusion and the reduction of structural violence. Within the wider literature, interventions that involve equitable collaboration with excluded groups are suggested as being most likely to address the key areas highlighted by our studies of sociopolitical context, transparency, accountability, leadership and capacity building [2].