Do social accountability approaches work? A review of the literature from selected low- and middle-income countries in the WHO South-East Asia region

Abstract Governance failures undermine efforts to achieve universal health coverage and improve health in low- and middle-income countries by decreasing efficiency and equity. Punitive measures to improve governance are largely ineffective. Social accountability strategies are perceived to enhance transparency and accountability through bottom-up approaches, but their effectiveness has not been explored comprehensively in the health systems of low- and middle-income countries in south and Southeast Asia where these strategies have been promoted. We conducted a narrative literature review to explore innovative social accountability approaches in Bangladesh, Bhutan, India, Indonesia, the Maldives, Myanmar and Nepal spanning the period 2007–August 2017, searching PubMed, Scopus and Google Scholar. To augment this, we also performed additional PubMed and Google Scholar searches (September 2017–December 2019) to identify recent papers, resulting in 38 documents (24 peer-reviewed articles and 14 grey sources), which we reviewed. Findings were analysed using framework analysis and categorized into three major themes: transparency/governance (eight), accountability (11) and community participation (five) papers. The majority of the reviewed approaches were implemented in Bangladesh, India and Nepal. The interventions differed on context (geographical to social), range (boarder reform to specific approaches), actors (public to private) and levels (community-specific to system level). The initiatives were associated with a variety of positive outcomes (e.g. improved monitoring, resource mobilization, service provision plus as a bridge between the engaged community and the health system), yet the evidence is inconclusive as to the extent that these influence health outcomes and access to health care. The review shows that there is no common blueprint which makes accountability mechanisms viable and effective; the effectiveness of these initiatives depended largely on context, capacity, information, spectrum of actor involvement, independence from power agendas and leadership. Major challenges that undermined effective implementation include lack of capacity, poor commitment and design and insufficient community participation.


Introduction
Improving health systems' responsiveness, quality and efficiency remains an ongoing challenge in many low-and middle-income countries (LMICs) , and enhancing the quality of governance and accountability is increasingly critical in achieving this. Although there are debates on how these objectives should be sequenced, there is growing recognition that achieving them requires managerial 'good governance' models and bottom-up social accountability approaches involving a variety of community actors. According to the World Bank's 'long and short route' framework of accountability, with the long route, citizens influence policymakers and policymakers in turn influence service providers. When this long route breaks down, there are fewer opportunities to ensure service provision is accessible and equitable. Given the lengthy time that the long route of accountability takes in many settings, it is expected that service outcomes could be better improved by strengthening the short route through increasing citizens' power over providers (World Bank Group, 2004).
Social accountability refers to an approach that focuses on civic engagement, i.e. ordinary citizens and/or civil society organizations participating directly or indirectly in policy processes to ensure that their concerns are taken into account and services are responsive to their needs (Carmen et al., 2004). Accountability in this context is the willingness of politicians to justify their actions and to accept electoral, legal or administrative penalties as appropriate. Two of the key aspects of social accountability are answerability (the right to receive relevant information and explanation for actions) and enforceability (the right to impose sanctions if the information or rationale are deemed inappropriate) (George, 2003;Ackerman, 2004). Answerability ensures the compulsion of policymakers or service providers to meet performance goals, while enforceability requires actions with penalty following failure to comply (Bruen et al., 2014). The voice is the instrument of accountability between citizens and politicians, with a range of measures through which citizens express their preferences and influence politicians. Improved accountability requires citizens to have a voice when it comes to locally elected leaders who can hold public servants to account (Lewis, 2006). However, this is not sufficient for accountability; it may lead to answerability, but it does not necessarily lead to enforceability (World Bank Group, 2004). Initiatives seeking to promote meaningful participation and accountability are often considered an important element to improve health system performance (Lewis, 2006;Vian, 2007;Ringold et al., 2012).
Social accountability should be considered as a multi-pronged process that utilizes multiple tactics, encourages collective action and voice alongside governmental reforms that bolster public sector responsiveness and facilitates outcomes and impacts that are more promising (Boydell, 2018). The concept reflects complex interactions between different stakeholders who have varying degrees of interest and power at different points in the service delivery process (Brinkerhoff, 2004;Ringold et al., 2012;Joshi, 2013).
A range of social accountability mechanisms, e.g. participation, watchdog organizations, scorecards and public hearings, have been implemented in South/Southeast Asian countries, with different forms often implemented as packages (Boydell and Kissbury, 2014). Their aims are diverse, including improving access to and quality of care, empowering community stakeholders and improving service efficiency. Thus, the emerging question that we seek to address in this review is: what are the social accountability initiatives that have been implemented, how did they function and what made them successful or not? This review seeks to address this critical gap in the literature and to inform the design of innovative, bottom-up community-based approaches to improve health sector governance.

Scope and objective of the review
This literature review presented here is a subset of a larger review on corruption in Bangladesh, which highlighted a range of participatory and social accountability mechanisms. This review sought to synthesize the evidence specifically on social accountability in selected countries in Southeast Asia, and its implication for strengthening health sector governance, addressing corruption and improving health system performance through citizens' engagement in Bangladesh, Bhutan, India, Indonesia, the Maldives, Myanmar and Nepal.

Case definitions
To answer our question, we developed a literature review protocol using a search strategy identifying the scope and methods for the review (data sources, key search terms and eligibility criteria). The definitions of key concepts used are described in Table 1.
We searched English-language literature papers between January 2007 and August 2017 in the initial phase and later extended the search from September 2017 to December 2019. The search focused on Bangladesh, Bhutan, India, Indonesia, Myanmar, the Maldives and Nepal, which is where the majority of the social accountability initiatives have been implemented and there is a critical mass of studies, enabling us to draw conclusions. For published journal articles (peer-reviewed), we searched electronic databases PubMed, Scopus and Google Scholar. We used the search terms 'Corruption', 'Informal payment', 'Anti-corruption', 'Governance', 'Good

KEY MESSAGES
• Governance failures undermine efforts to achieve universal health coverage and improve health in low-and middle-income countries by decreasing efficiency and equity. • A variety of context-specific social accountability interventions have been tried in Bangladesh, Bhutan, India, Indonesia, the Maldives, Myanmar and Nepal. The evidence suggests that such properly designed and implemented interventions enhance and supplement existing accountability mechanisms.
governance', 'Accountability', 'Social accountability', 'Community', 'Healthcare provider', 'Health service' and 'Healthcare facility', and combined these with the countries' names (Table 2 key search terms). The search terms were used in combination with the Boolean operator 'AND'. In addition, we searched the Internet to identify relevant grey literature. We also searched databases of relevant international organizations, e.g. the World Health Organization, Transparency International (TI) and the World Bank Group, to identify relevant papers and reports (Table 3 search protocol).

Search retrieval and analysis of studies
Two researchers screened the list of articles independently. Titles and abstracts of the articles were read to determine their relevance to the topic. After excluding duplicate citations, we excluded nonpeer-reviewed journal articles, articles not in the English language and those published before January 2007. We further excluded editorials, proposals and protocols that contained no empirical findings. For papers meeting the inclusion criteria, the researchers retrieved the full text for further assessment. Any disagreement was resolved through discussion with the project lead. We used framework analysis as an analytical strategy, to identify key themes and divergent findings within the literature, explore their characteristics and synthesize data (Gale et al., 2013). A sample of sources was read and re-read by the study team members (NN, RH, SMA) independently to develop the initial coding matrix of themes. This was discussed, refined and agreed before the remainder of the sources were reviewed and analysed using the agreed coding matrix. Researchers independently extracted data from included studies using the matrix, which helped to develop a working analytic framework. Differences in data extraction were resolved by discussion with the broader team, to ensure quality. After extraction and tabulation, we categorized the data into themes (transparency, accountability and community participation) for analysis. Once data had been extracted and classified according to themes, the researchers read and re-read the papers to identify and confirm the classification. The categories were grouped under three major sub-themes: 'transparency', 'accountability' and 'community participation'. This approach facilitated the analysis; key findings were summarized  from each matrix column, identifying the commonalities and differences.

Results
From the initial search, we identified a total of 30 115 articles from the three electronic databases. The preferred reporting items for systematic reviews and meta-analysis (PRISMA) flow are presented in Figure 1. After removing 8242 duplicates, 21 873 articles were screened, and a further 15 903 articles were excluded as they were not in English, were published before 2007 or full text was not available. From the remaining 5970 articles, a further 4498 articles were excluded as they were not focused on social accountability issues. A further 1389 articles were excluded as they did not refer to the selected countries. Finally, 83 articles were read. Of these, 13 were excluded as they did not include sufficient data for analysis and 30 were excluded as they focused on corruption issues broadly. Out of the remaining 40 social accountability and governance-related papers, we further excluded 23 that discussed possible policies and strategies rather than specific interventions and also those that did not include any outcomes. Out of these, we included 17 papers, of which six were on transparency and governance issues, seven on accountability mechanisms and four on community participation approaches. Apart from the journal articles, an additional 15 documents were also selected that included project and programme reports of national and international organizations, from which finally nine were reviewed. A total of 26 documents (17 journal articles and nine grey literature sources) were reviewed for the paper from the initial search (Figure 1 PRISMA diagram). The updated PubMed and Google Scholar searches identified an additional 39 documents, of which 12 (seven journal articles and five grey literature sources) were identified and reviewed for this paper (Table 4). Thus, a total of 38 documents (24 journal articles and 14 grey literature sources) were included in this paper from the two search phases.

Study design and country of origins
Studies were selected irrespective of their study design, and included a variety of designs. From the initial search; out of 17 journal articles, seven were reviews, five were quantitative studies, four were qualitative studies and one was a case study. Out of the nine grey literature sources, there were six project reports, one case study, one working paper and one blog ( Table 4). Out of the 26 documents from the initial search, six each were from Bangladesh, India and Nepal. Three were from Myanmar, with one each from Bhutan and Indonesia and three from LMICs (Table 5).
From the extended search, out of seven journal articles, four were reviews and three were qualitative studies. Out of the five grey literature documents, three were working papers and two were book chapters ( Table 4). Out of these 12 newly reviewed documents from the extended search, two were from Indonesia, three each were from India and Nepal and four were from Bangladesh.
Out of the 24 journal articles identified during the two-phase search, 11 were reviews, five were quantitative studies and eight were qualitative studies. Out of the 14 grey literature sources, one was a case study, four were working papers, one was a blog, two were book chapters and six were project reports (Table 4).
From the 38 documents reviewed through the updated search, 10 were from Bangladesh, nine each were from India and Nepal, three each were from Myanmar and Indonesia, one was from Bhutan and the remaining three were from mixed LMICs.

Scope and challenges of the approaches practised
The social accountability strategies and interventions implemented to enhance transparency, accountability and community participation varied widely (Table 6). Despite some positive outcomes and process features, most of these approaches had important limitations as well.
The results are presented thematically below.

Local governance
Local governance was a social accountability approach to enhance transparency and accountability at the local level. In three studies (Cleary et al., 2013;Garimella and Sheikh, 2016;, it was found to be effective by enabling space for decisionmaking at the local level. It improved health units' performance by enhancing local authorities' decision-making and enabling them to apply a bottom-up approach. Ensuring accountability was identified as a major role of local governance. A case study conducted in Nepal argued that local government accountability should be envisioned as an ongoing process and should be accompanied by systems of local planning and implementation, and revitalization of local democracy (Dhungana, 2019). Evidence indicates that health system performance in achieving the objectives of efficiency, quality and equity is contingent on the breadth of 'decision space' at the local level. The functional areas of finance, service autonomy, recruitment rules, access rules and departmental rules normally have a very narrow 'decision space' at the local level, constraining the power of local authorities. In Odisha, India, Rogi Kalyan Samitis (RKS), a composite body of decisionmaking in peripheral health units, was formed with a mandate to ensure transparency in health facilities governance. For public service delivery health institutions, such as hospitals and healthcare centres, RKS was formed as an institution of local decision-making to take the public health system agenda forward. The functions of RKS included: (1) governance (accountability, responsiveness and transparency); (2) infrastructure (construction and maintenance, purchase and out-sourcing); (3) human resources management (hiring, transfer and training of staff); (4) financial resource management (cost-cutting measures, resource generation); and (5) quality improvement (supervision, modernization, quality assurance and accreditation). Panda et al. conducted a study on RSK staff satisfaction which showed that the majority (87%) of respondents were 'satisfied' with their current roles. Almost all (98%) noted that local decision-making helped to improve the performance of health units . Garimella and Sheikh conducted a case study to explore posting and transfer at the primary level in Tamil Nadu, India, in the context of the complex governance system of the government health sector. The study emphasized the need for bottomup approaches to address the complexity within the governance context. Moreover, the blurring boundaries between public-private actors needed to be addressed for coordinated efforts towards local governance interventions (Garimella and Sheikh, 2016). Cleary et al. examined accountability in district-level health system governance, and found it important to limit the potential negative impacts of powerful actors to leverage a shift towards well-functioning accountability. A balance between achieving accountability and allowing local-level innovation was suggested as helpful. Findings show that accountability mechanisms could be key tools for ensuring the answerability of public primary healthcare facilities to central bureaucracies through the district health system, while at the same time providing the local decision space that could increase citizen and patient responsiveness (Cleary et al., 2013).

Citizen's Charters
Citizen's Charters are one new social accountability approach to inform citizens about service entitlements, standards and rights. A Citizen's Charter is a document articulating the commitment of government organizations towards citizens through clearly specified measures (Sharma, 2012). In Bangladesh, introducing and enforcing a Citizen's Charter has been seen to ensure transparency and integrity in large-scale public procurement (Shohag, 2018). However, Citizen's Charters have been largely ineffective and have failed to have an impact on enhancing accountability due to the top-down approach of their implementation.
Total number of additional documents reviewed (nine from initial search þ five from extended search) ¼ 14.
Grand total of documents reviewed for this paper (24 journal articles þ 14 additional documents) ¼ 38.
Citizen's Charters are intended to provide information to citizens on the choice and standards of services that should be provided by an institution. During 2004, a Citizen's Charter initiative in Chandigarh, India was adopted with the aim of improving public service delivery. The Charter was supposed to inform citizens about specific complaint centres. The 32-page document was divided into sections. One of its special features was the universal email address for all types of complaints. A case study by Sharma reveals that the Citizen's Charter made little impact as it was not displayed anywhere accessible to citizens. Even the employees of the government agencies were not well informed. The top-down approach resulted in poor design, a lack of awareness or interest among stakeholders, a lack of information and an absence of an implementation strategy or community awareness and participation (Sharma, 2012).
In 2007, a Citizen's Charter was introduced in Bangladesh to improve the delivery of quality services and of transparency and accountability at the local level. Huque and Ahsan conducted a survey in Rajshahi district to evaluate its impact. Findings revealed that only a small number of people were aware of its existence. Information was presented in a way that did not allow citizens to play a meaningful role. Respondents had no stake in the preparation or implementation of the Charter. The survey revealed a number of factors limiting the success of the initiative, which includes similar factors as for Chandigarh. It found that a top-down approach to adopting and implementing the Charter in a haphazard fashion may have contributed to its limited success (Sharma, 2012;Huque and Ahsan, 2016).

Social audits
Social audits are an asocial accountability tool to enhance service delivery transparency and accountability by improving participation (GIZ, 2014). In theory, social audits provide opportunities for mutual accountability by evaluating health system performance via citizens. The literature suggests that they can be effective for enhancing community roles in local healthcare service monitoring by raising service demand and enabling organizational change. Social audits are based on the idea that people's participation in policy processes can become an effective tool to fight corruption, and that this can be achieved when people are aware of the nature and effects of corruption. Civil society groups in India have played a prominent role in raising awareness among citizens about the negative impact of corruption, which has also helped to strengthen government-citizen relations (Kumar, 2019). A social audit initiative in Andhra Pradesh state in India involved poor citizens. But the programme has been found ineffective in redressing and sanctioning functions; while it proved reasonably effective in detecting malfeasance, it did little to reduce it over three successive iterations (Blair, 2018).
Nepal has a long history of utilising social audits. From 2013 to 2014, its government conducted social audits in 602 facilities in 45 districts (out of a total 75 districts) with support from local and international development agencies (e.g. GIZ and the UN) to enhance community participation in decision-making and monitor local healthcare services. In this process, a Social Audit Committee was formed within districts. To disseminate findings within communities, a mass public gathering was organized ensuring the presence of facility service providers and authorities. Finally, a local action plan was developed to enhance transparency and good governance in facilities by assigning responsibilities. Vacant positions were also filled through temporary contracts, the behaviour of health workers improved, facilities were made more responsive to patients' needs and the Health Facility Management Committees were reformed or re-energized. While findings about the impact were only tracked in two facilities, limiting the potential to learn from this intervention, it appears that the use of services increased, staffing shortages were fully or partially filled and drug shortage and infrastructure problems were solved (GIZ, 2014). Social audits have been used as a mechanism to hold frontline health service providers to account-e.g. the audit process in Dang District, Nepal facilitated direct accountability between service providers and the community. Participants reported that the process improved information provision and provided opportunities for dialogue between the community and service providers. While social audits have a positive role in increasing the transparency, accessibility and quality of services, their effectiveness in addressing perennial governance problems has been mixed. Manipulation of the participation process, falsification of information and communities' lack of power have all affected the role of social audits in facilitating accountability. The study authors argued that it is essential to consider these factors while designing and implementing social audit processes and accountability mechanisms between service providers and the community (Gurung et al., 2019). In Nepal, social audits have also been implemented to enhance accountability in maternal health services, but impacts on governance were mainly found at the local level. The problem of corruption in health sector endangers the sustainability ofeffective and quality health care, therefore, Indonesia established an anti-fraud system to protect the universal health insurance fund. Analyze the current anti-fraud system in universalhealth insurance through the lens of international law and principles of good governance. The sociolegal approach is chosen to study therelationship between the State party obligations to international lawand the implementation of concerning universalhealth care and anti-corruption in the designated anti-fraud system. Good governance principles are essential in designing an effectiveanti-fraud system due to the correlation between human rights and anti-corruption that both areas emphasize good governance principles as guiding principles for the realization of human rights and the making of potent anti-corruption strategy.

Local governance
Health worker posting and transfer at primary level in Tamil Nadu: Governance of a complex health system function Garimella and Sheikh (2016) Posting and transfer (PT) of health personnel -is a contested domain, driven by varied expressions of private and public interest To investigate policymaking for PT in the government health sector and implementation of policies as Case study of a PT reform policy at primary health care level in Tamil Nadu State, to understand how different groups of health systems The imperative of enforcing rules may need to be complemented with bottom-up policy approaches, including treating PT not merely as system dysfunction, but also as a potential (continued) There is a need to foster greater civic demands on accountability and foster measures for deliberation at the municipal level on a more regular basis.
Overall, local government accountability should be envisioned as a work-in-progress pursuit and should be coupled with systems of local planning and implementation and vitalization of local democracy.

Citizen charter
An Evaluation of a citizen's charter in local government: a case study of Chandigarh, India.
Sharma ( in Bangladesh at the course of many times. This research is based on corruption in the It has scrutinized the overall scenario of corruption and irregularities in the service sectors in Bangladesh and finally it has examined the scenario and experiences of To control service sector corruption appropriate monitoring and oversight mechanisms must be in place in each institution, transparency and integrity has to be ensured in the public (continued) Examine and understand the effect of decentralization at the district health service from the perspectives of service users and providers.
District health facilities.
Decentralization was positively associated with increased service access and utilization and improved service deliv- There are mixed effects of decentralization on corruption and that the types, dynamics of corruption and impact of different anti-corruption approaches may vary in different decentralized settings. Effective decentralization of authority is yet to be established at the (continued) A mix of financial and non-financial incentives, predictable for the CHWs, was found to be an effective strategy to enhance performance, especially of those CHWs with multiple tasks. Performance-based financial incentives sometimes resulted in neglect of (continued) The linkage between citizens and some combination of elected political leaders and those they direct to provide the services.
Explores the paths these three routes cantake and their potential effectiveness in providing citizens a number of institutionalmechanisms to hold political leaders and public service providers accountable,improve service delivery, empower poor people and ultimately enhancewell-being. Civil societies can directly addressing the state bureaucracy in seeking changes in how a policy is implemented.

Public hearings Preventing Corruption in Public Service Delivery in Bangladesh
Ahmed (2016) According to all major global indicators of corruption, Bangladesh is one of the most corrupt countries in the world as per Transparency International according to its Corruption Perception Index (CPI).
The Anti-Corruption Commission (ACC) conducts public hearings at the upazila level for ensuring the accountability of public officials and also transparency of their work.
This study is based on the written complaints raised by 1440 citizens in 72 public hearings conducted by the Anti-Corruption Commission (ACC) in Bangladesh. TIB conducted a study of 13 public hearings with 195 respondents. The reasons for liking public hearings was that it created opportunities for making authorities accountable to citizens (75%) followed by the opportunity to raise complaints before officials (continued) Evidence on impact (69%) and commitment to solve complaints (20%).

Community Action Groups
Women in the lead monitoring health services, Naripokkho, Bangladesh COPASAH (2015) Naripokkho works in all the 64 districts of Bangladesh on empowerment and reproductive health and rights of women specially in rural setting. To strengthen the accountability mechanism in the health service delivery mechanism among rural women. NariDal a village health facility monitoring group by village women's.
Helped to raise awareness on health rights and increase the use of health services in the village women by enhancing accountabilities in service provision at local health facilities.

Participatory Complaints Survey
Local Governance: Accountable Public service in Indonesia GTZ (2009) Challenges of equal access to quality services due to vast geographical area. Increasing the accountability of the public sector, improvement of public services through civil society participation. Complaint survey conducted by service units.
Citizens were more aware and empowered, there was an improvement in service provision; service users also had less complaint than earlier with regards to constraints in service provision.

Social Audits
Making local health services accountable. Social auditing in Nepal's health sector.
GIZ, (2014) Nepal at local level, public health facilities across the country face daunting problems, including insufficient supplies of drugs and basic equipment, understaffing and absenteeism, and low level of accountability to local people.
Social auditing has been introduced on an increasingly wide scale to enhance citizens' ability to participate in decision making about their health services at facility level. In 2013-14, a total of 602 facilities in 45 districts (i.e. the majority of the country's 75 districts) held social audits.
Social audits increased demand for services by informing people. Staffing shortages were fully or partially filled.
The challenge of drug stock-outs and infrastructure problems with buildings and equipment were effectively dealt. On a broader scale, the social audit added value in such as giving facility in-charges opportunities not just to respond to questions and concerns, but educate local community members, The role of social audit as a social accountability mechanism for strengthening governance and service delivery in the primary health care setting of Nepal: a qualitative study Gurung et al.
Kumar (2019) Social audit is a mechanism used to hold frontline health service providers accountable. Using the case of the social audit process in Dang District, Nepal, this study explored the role of social audit in facilitating direct accountability between service providers and community. A total of 39 interviews were held with health facility operation and management committee members, service providers, district level health managers and non-government organization members.
Reviews of records of social audit action plans were undertaken at 10 health facilities.
Participants reported that the social audit process was able to facilitate information provision/data collection, and provided opportunities for dialogue between community and service providers, but the provision of sanctions was found to be weak. While social audit had a positive role in increasing transparency, accessibility and quality of services, its effectiveness in addressing perennial governance problems was mixed.
(continued) Factors contributing to the lack of broader impact were the absence of a mandate for community health volunteers to play an active role in the social accountability process, and limited capacity, including of resources (Hamal et al., 2019).

Score cards
Score cards are a quantitative approach typically involving surveys of citizen satisfaction, which include a facilitated meeting between providers and beneficiaries to discuss results and agree on follow-up actions (World Bank Group, 2012). Work on community scorecards is intended to be a participatory, community-based social accountability approach to evaluate and improve public services, and to inform and empower local actors. The use of score cards was found to be effective for monitoring quality in service provision in one study. A facility survey using a score card conducted in Bangladesh reported it as useful to better understand various aspects of service delivery through gathered data, which also helped to strengthen the management information system itself (Khan et al., 2013). In Myanmar, the 3MDG Fund trained its implementing partners (IPs) on the use of community score cards as a participatory approach for communities and service providers to engage in dialogue on the delivery of services (3MDG, 2016). A score card health survey was conducted in 80 health complexes (upazila) in Bangladesh in 2009. A list of basic medical equipment was used to calculate equipment availability for health facilities. More than 60% of the facilities were found to have at least 75% of the basic medical equipment. In terms of human resources, both physicians and nurses were in short supply at all levels of the healthcare system. The overall job satisfaction index was <50 for physicians and 66 for nurses out of 100, with 100 being very satisfied. The score card approach was found useful for monitoring quality (Khan et al., 2013).

Participatory complaints surveys
Participatory complaints surveys are a participatory social accountability tool to enhance service provision accountability. Complaints surveys empower citizens to hold authorities accountable for given services. They are effective for identifying gaps in service provision, and can improve local-level planning.
As a vast country with a population of over 240 million, spread out over about 6000 inhabited islands, Indonesia faces enormous challenges with the provision of equal access to quality services. Since 2000, the State Ministry of Administrative Reform has implemented a Support for Good Governance initiative. A representative patient complaint survey was developed through workshops with service users (80%) and service unit staff (20%) led by trained facilitators. A complaint survey was conducted at 60 service units with a minimum of 80% of service users at the lowest level of service provision. Typical complaints about the services of Local Health Centres (PusKesMas) included lack of medical personnel, lack of discipline/ skills/information sharing of medical personnel, lack of medication, variety in pricing and finally exorbitant costs for medication. A number of districts and municipalities (74 out of approximately 500) applied this participatory method, reacting to the complaints of 380 000 respondents. A number of districts and municipalities continue to expand this approach into new sectors, often at their own cost. A few service units at the provincial and national level (the customs bureau) have also successfully applied the method. As patient compliant surveys were repeated, citizens' became more aware and empowered, and there was an improvement in service provision; service users also had fewer complaints regarding service provision (GTZ, 2009).

Community volunteer/participation
Community participation is defined as the involvement of people in a community in projects to solve their own challenges (World Bank Group, 2012). Community participation in the form of community volunteers or health workers was found to be effective for enhancing system performance in resource constraint scenarios. It was also effective in bridging the community and systems by identifying community needs. It positively impacted local resource mobilization and helped organize activity planning.
Community health workers (CHWs) are an effective way to enhance performance in preventive, curative and promotional primary healthcare services in LMICs with a given mix of financial and nonfinancial incentives (Kok et al., 2015;Wangmo et al., 2016). A systematic review of 140 quantitative and qualitative studies identified factors related to the nature of tasks and time spent on delivery, human resource management, quality assurance, links with the community, links with the health system and resources and logistics having an influence on CHW performance. Good performance was associated with intervention designs involving a mix of incentives, frequent supervision, continuous training, community involvement and strong coordination and communication between CHWs and health professionals, leading to the increased credibility of CHWs (Kok et al., 2015).
Myanmar faced critical resource constraints, creating major gaps in access to and coverage of health services. Recognizing the benefits of community-based health workers, Myanmar trained, deployed and integrated auxiliary midwives (AMWs) in the health system, to deliver maternal and child health services (MCH) services to hardto-reach and remote areas. A quantitative cross-sectional study was conducted in 2013 to assess the extent of AMWs' contribution to addressing the shortage of midwives. AMWs were able to provide essential maternal and child health services including antenatal care, normal delivery and post-natal care, and had a comparative advantage due to longer service in hard-to-reach villages where they lived, speaking the same dialect as the locals, understanding the socio-cultural dimensions and being well accepted by the community. Despite these contributions, challenges remain; e.g. 90% of AMWs stated that they had received no adequate supervision, refresher training, replenishment of AMW kits or reimbursement of transportation costs (Wangmo et al., 2016).
In 2005, in India, the National Rural Health Mission launched to revamp the rural health system. An ethnographic study was conducted in Odisha to gather evidence on community interaction. It showed that for health workers, the notion of integration goes well beyond a technical lens of mixing different health services. Crucially, 'teamwork' and 'building trust with the community' (beyond trust in health services) are critical components. Evidence shows that highly hierarchical health bureaucratic structures, which rest on top-down communications, limit efforts towards sustainable health system integration (Mishra, 2014).
Unlocking community capabilities through self-help organizations (SHOs) was helpful for bridging between the community and the system. An international research organization in Bangladesh, 'ICDDR, B', implemented a project on 'self-help for health' to work with existing rural SHOs. SHOs are organizations formed by villagers for their well-being through their own initiatives without external material help. Following a self-help conceptual framework, the project focused on building the capacity of SHOs and their members through training on organizational issues, imparting health literacy and supporting participatory planning and monitoring. Villagers and members of the SHOs actively participated in the self-help activities. SHO functionality increased in the intervention area, in terms of improved organizational processes and planned health activities. These included convening more regular meetings, identifying community needs, developing and implementing action plans and monitoring progress and impact. Between 1999 and 2015, while decreases in infant mortality and increases in utilization of at least one antenatal care visit occurred, increases in immunization, skilled birth attendance, facility deliveries and sanitary latrines were substantially higher in the intervention area than in the comparison area (Bhuiya et al., 2016).
In Jhenaidah, Bangladesh, a community-driven initiative helped to mobilise 46 additional workers from the local community for the Chowgacha health complex (sub-district health facility) by involving communities, particularly including locally influential individuals (SHARE, 2016). The 3MDG fund is working with existing IPs to strengthen their engagement with communities, including poor and vulnerable populations, by providing support to improve their approach to participation, inclusion, information sharing and responding to community feedback in Myanmar (3MDG, 2012). Community participation helped to mobilize local resources and to raise collective demands through community empowerment (3MDG, 2012) as well. It needed the active involvement of citizens, as well as the facilitation, communication and a change in mindset of the administrators. A high turnover and a lack of incentive limited community-led interventions. To practice social accountability approach understanding of the local context and perspective of both community and provider, building community capacity and ownership, linking the interventions to the formal system and meeting community demand were all found to be crucial for such interventions (3MDG, 2016; Bhuiya et al., 2016;Wangmo et al., 2016).

Community action groups
Community action groups (CAGs) are another form of communityled social accountability found to be effective in ensuring the inclusion of marginalized communities. They can raise collective voices and increase service demand, especially among poor women.
Experiments with community-led approaches have been practised to raise collective voices in this region (Asia Pacific Network, 2011;COPASAH, 2015;Bhuiya et al., 2016a). For example, 'Naripokkho' in Bangladesh is a national membership organization working on women's rights since 1983 to empower women (COPASAH, 2015). To strengthen accountability mechanisms in 2003, Naripokkho initiated a 'Women's Health Right's Advocacy Partnership' (WHRAP) in five districts with 16 NGOs and 640 active members. Under this WHRAP initiative, marginalized women were organized into groups named 'NariDal' in villages to monitor health services. As the poor and marginalized women of the community are reluctant to use available healthcare services, the NariDal members advocate with marginalized women for their health rights. In these meetings, issues like availability of healthcare services (e.g. facilities, medicines), women's health conditions, health rights, entitlements and obligations of providers are discussed. Though the NariDal members faced challenges from within the family and community for their involvement, through regular meetings they raised awareness on health rights and increased the use of health services by women in villages by enhancing accountability in service provision (COPASAH, 2015).
Between 2009 and 2011, Bhutan established CAGs in four districts (Asia Pacific Network, 2011). The group members included local government representatives, village health workers, religious group members and representatives from different sectors ensuring female representation. The group discussed priorities and developed a local action plan. They met quarterly and sent reports to the central level every six months. The initiative was reported as helpful and having improved village sanitation. Although the high turnover of village health workers was a major challenge, this approach was helpful to create community ownership of health activities, stimulate decentralization and build capacity of local leadership. CAG members receive a three-day training course covering sanitation, community motivation, nutrition and child care. CAGs were successful in improving sanitation in the villages (Asia Pacific Network, 2011).

Public hearings
Public hearings are defined as formal meetings at the community level where citizens express their grievances on matters of public interest to public officials who try to address these grievances (Ahmed, 2016). Public hearings are expected to provide a platform through which citizens can call authorities to account. They were found to be effective for exposing corruption and mismanagement in public service provision, but there is limited evidence on their impact (Ahmed, 2016).
Public hearings have been practised as a means of empowering citizens with information on public services and raising collective voices. This involves public officials and citizens of the same locality, and allows citizens to question the authorities directly on irregularities in public services. The Anti-Corruption Commission in Bangladesh organized 72 public hearings by 1440 citizens in 61 upazilas of 51 districts and in two metropolitan cities up until 2017. As per the public hearing findings, systematic corruption prevails in public service delivery and health was identified as one of the most corrupt service departments. An absence of citizen engagement was mentioned as a reason behind the corrupt practices (Ahmed, 2016). A follow-up survey conducted by TI, Bangladesh in 2017 found that 75% of respondents liked public hearings as a platform to make authorities accountable to citizens. Sixty-nine per cent thought that public hearings provide the opportunity to raise complaints before officials. The study also revealed that as a result of holding public hearings, authorities have taken measures like putting out more information boards, providing complaint boxes, improving filing systems and monitoring through CCTV to improve public service delivery (Ahmed, 2016). The steps taken after public hearings demonstrate that these mechanisms appear to be effective instruments in corruption prevention, this is not supported by the evidence.
'Ayauskam': A classic case of practising different social accountability tools The 'Ayauskam' project in India led by a civil society organization is a classic example where different social accountability tools were used to reduce corruption and improve service delivery responsiveness (PTF, 2012). Initiated in 1993 in the State of Odissa, Ayauskam conducted a baseline survey in 64 villages to explore corruption. It established community-based organizations and organized public hearings. It also organized a broad campaign against corruption, holding rallies and demonstrations and using media to protest corruption. Ayauskam faced several challenges throughout the process. The service providers, government officials and local politically influential people were not supportive, and obstructed efforts. They even filed criminal cases and made false claims against Ayauskam. The project staff made frequent efforts to have discussions with the authorities, service providers and local influential people. Through gradual cooperation between the them, Ayauskam was able to make it clear that it was combating corruption and not individuals. The authorities recognized the strength of community and thus started initiatives like village health committees, which increased community participation in the decisionmaking and monitoring processes. The intervention helped to improve child nutrition and antenatal and postnatal services. There was a reduction in corruption practices in government hospitals in the project area, with 80% of those surveyed not needing to pay a bribe for giving birth at local hospitals (PTF, 2012).
Broader reforms to enable social accountability In addition to the above initiatives, our review identified evidence on social accountability as a part of comprehensive health system reform packages-presented in the 'Good governance' and 'Decentralization' sections of this article. In many of these, there were implicit measures to involve end users and citizens and to ensure feedback loops and responsiveness.

Good governance
Good governance, defined by Huss et al. (2011) as the exercise of power through institutions to steer society for the public good, has been practised as a model to enhance transparency and accountability within systems. Indonesia established an anti-fraud system within its universal health insurance, and a study sought to analyse its operation through a good governance lens. Findings indicate that good governance principles are essential in designing an effective anti-fraud system due to the correlation between human rights and anti-corruption; both areas emphasize good governance principles as fundamental for the realization of human rights and the making of a viable anti-corruption strategy (Juwita, 2018). Good governance approaches were also aimed at enhancing accountability and improving service delivery in Karnataka, India (Huss et al., 2011). A public complaint agency (KLA) was created in Karnataka state in India in 1986, which played a prominent role in controlling systemic corruption. KLA had the authority to investigate complaints from citizens about public maladministration and to initiate prosecution for criminal offences. In the initial phase, KLA was criticized by the Karnataka High Court and Karnataka Administrative Reform Commission for its failure to hold governments accountable, ensure effective redressal of grievances and improve public administration governance. Later, the post of Vigilance Director for Health, Education and Family Welfare (VDH) was created and under strong leadership it became widely known, gaining a reputation for independence and a strong will to fight maladministration. Thus, the change in leadership in 2001 and the creation of the position of Vigilance Director for Health improved the effectiveness of the KLA. The Karnataka experience showed that a shift towards good governance requires the interaction of leaders, followers and system changes. An effective accountability mechanism requires a committed and powerful leadership, adequate resources, robust capability to investigate and deal with internal governance issues and the authority to propose institutional reforms (Huss et al., 2011).

Decentralization
Decentralization is defined as a socio-political process of powersharing arrangements between central government and local authorities in planning, management and decision-making (Regmi et al., 2010). Decentralization was aimed at enhancing accountability with resources via power transformation at the local level. It was effective in terms of bringing services close to citizens.
Decentralization has had a positive impact in improving district health service provision, especially in planning and management with a clear local agenda. For example, within the decentralization framework of government, the Ministry of Health (MoH) Nepal initiated the decentralization of primary care services closer to citizens in 1999. Regmi et al.'s (2010a) study on decentralization revealed that service users considered decentralization as a means of transferring authorities and accountabilities with resources (both human resource and finance) from the centre to local authorities and community healthcare facilities. This was viewed by respondents as a possible advantage of decentralization for the local government. Decentralization was positively associated with increased service access and utilization and improved service delivery. Most of the districts' health service facilities (health institutions) were handed over to local committees. The main purpose of restructuring was to involve community people and bring health services closer to citizens. The study reported decentralization impacting positively on the district health services in terms of service provision, community participation and empowerment, service planning, management and coordination. It also identified the barriers to implementation, such as difficulties in developing capacity, monitoring and accountability systems, clarity in roles and responsibilities and also fund allocation and distribution (Regmi et al., 2010a).
In Nepal, decentralized human resource management was practised by handing over the health facilities of 28 districts to local bodies (Devkota et al., 2013;Gurung and Tuladhar, 2013) and forming village development and district health development committees. The initiative promoted ownership at the local level as a result of resource sharing to equip health facilities, and to improve recruitment and retention of staff locally. However, weak monitoring has been seen as a key obstacle in promoting local leadership (Devkota et al., 2013). Nepal's experience revealed that successful implementation of decentralization requires a broader context of institutional capacity building and resource management and underlines the need for consideration of these during implementation processes (Regmi et al., 2010b). The active involvement of service users, providers and policymakers in the process of decentralization and clear local level agendas were reported crucial for such initiatives (Regmi et al., 2010b). The effects of decentralization on corruption are mixed, and the types, dynamics of corruption and impact of different anti-corruption approaches vary in different decentralized contexts. Conflicting provisions regarding administrative authority at the different levels affect the effective functioning of local-level institutions and their ability to govern, and this requires strengthening of the upward and downward accountability mechanisms (Bhattacharya et al., 2018).

Discussion
This review of the literature on social accountability and its impact on governance in the health sector sought to explore alternative, bottom-up and community-engaged interventions to improve governance and combat health sector corruption in selected LMICs in South and Southeast Asia. Findings reveal a multitude of experiments in different countries of the region to strengthen health sector governance through 'social accountability' initiatives. Our review demonstrates that different countries (e.g. Bangladesh, Bhutan, India, Indonesia, Myanmar, the Maldives and Nepal) have implemented a broad variety of social accountability mechanisms, e.g. social audits, score cards, participatory complaints surveys, public hearings, community volunteers and community actions groups.
Key themes emerging from the review were: transparency, accountability and community participation. Across the papers on transparency approaches, strong political commitment, appropriate policy design and active participation of citizens appeared to be key to effective implementation of such interventions; however, the link between these factors and the success of social accountability programmes is only tentative. Similarly, the absence of these factors undermined the application of the approaches based on transparency. Transparency intervention such as Citizen's Charters often failed due to poor design following a top-down approach, poor political commitment and lack of citizens' involvement through a participatory process. Approaches based on strengthening local governance, on the other hand, allowing space for decision-making at the local level, were found to have potential provided there was strong leadership. However, the balance between monitoring accountability and allowing a local-level decision-making space was a challenge to such interventions. Thus, the evidence on whether incorporating these key elements into the interventions can ensure an impact is inconclusive.
Across the papers reviewed under accountability approaches, we found that effective design and implementation capacity were considered crucial for interventions like social audits, participatory complaint surveys and score cards, while the absence of robust guidelines or skilled facilitators of the process were constraining factors. Moreover, the active participation of citizens and mechanisms to follow up on complaints were a clear prerequisite. There was reasonable evidence that the interventions were effective in enhancing local-level accountability, monitoring service quality and empowering citizens through active involvement in shaping service provision. However, there was limited evidence that accountability approaches such as public hearings were effective beyond providing a platform to raise citizens' voices. Evidence suggests that only coordinating citizens may fall short in achieving governance and service delivery improvements unless these processes are institutionalized and linked with systematic reform. Overall, the accountability approaches emphasized the mutual responsibility and participation of both service providers and communities, with similar findings reported in a social accountability study in Malawi (Gullo et al., 2017).
Apart from specific accountability approaches, broader reforms like decentralization and good governance can play an important role in promoting social accountability, but they largely depend on institutional context and capacity. Governance is recognized as a cornerstone of a well-performing health system, and initiatives aimed to improve governance are especially important in lessdeveloped countries whose health systems face numerous constraints (WHO, 2007). For this reason, a number of initiatives have sought to improve governance with a focus on strengthening legal frameworks, regulatory capacity and enforcement powers (Mikkelsen-Lopez I et al., 2011). However, despite being effective at the individual programme level, evidence demonstrates that top-down approaches remain insufficiently effective and often create new areas for rule-breaking and governance failures (Khan et al., 2019). The proliferation of social accountability (bottom-up) approaches takes a different perspective; it argues for involving and empowering relevant key stakeholders and a broad range of policymakers, ensuring that there are formal mechanisms to channel their concerns into actions and achieve changes in the health system. Khan et al. (2019) argue that, by aligning incentives and motivating at least some powerful sectoral organizations, sectoral strategies can be applied by organizations to address sector-specific problems.
Across the papers reviewed under community participation interventions, we found that participation in the form of community volunteer/worker/action groups is common in many social accountability programmes. Understanding and building on the perspective of both communities and providers and creating community ownership through identifying roles were a prerequisite of community-led interventions. The papers reviewed suggest that these approaches helped in increasing healthcare demand by raising collective voices and ensuring the inclusion of marginalized communities. However, the success of participation-based strategies depends critically on local power relations and on citizens' capacity for collective voice. Thus, some of the interventions remain effective to an extent, achieving their intermediate goals by establishing inclusive and participatory processes; however, evidence on how these can be scaled up to a broader population and sustained, or how they can improve substantively health outcomes, is limited.
Most of the papers reviewed the technical preconditions for implementing effective social accountability models. Many of these are based on good intentions but may underestimate the importance of strategic design-not only of the interventions per se but also of their fit into the overall health system. As highlighted in the last section, social accountability initiatives are often part of packages of managerial interventions to counter the effects of poor governance-e.g. to improve human resources management and increase transparency. Another set of papers showed that broader decentralization reforms were mainly aimed at improving efficiency in quality service delivery and enabling local autonomy in decision-making; these were not clearly linked to social accountability approaches. Therefore, an important lesson may be that social (bottom-up) accountability has to be implemented in conjunction with top-down approaches (e.g. potentially using co-design and co-production) (Manikamet al. 2017;McAuliffe et al., 2017;Beran et al., 2018;Ward et al., 2018). Implementing social accountability as a standalone intervention may be ineffective if institutional and contextual factors are opposing it (UNDP, 2013). Thus, the review showed that creating spaces for decision-making at the local level was useful but was constrained by administrative needs, capacity and other contextual factors.
Another critical point that is little discussed in the literature is the role of power (Shiffman, 2014). The review showed that the citizens and their communities played a central role in the accountability mechanisms, giving them voice via awareness building and provision of information, and this was frequently practised for each of the social accountability mechanisms, from the local to the national level. While many of the papers explicitly acknowledge that support from powerful actors was a necessary precondition for implementing the interventions (Hickey and King, 2016), the underlying premise--that the intervention is often designed and implemented in a way that may not challenge or that may even enhance the underlying power structures-was not explicitly examined.

Conclusion
This review synthesized a diverse type of social accountability initiatives implemented in selected countries in South and Southeast Asia to enhance transparency, accountability and citizens' participation. The initiatives and interventions differed from context to context, were initiated by a range of actors and occurred at various levels. Findings indicated that success (perceived or measured) largely depended on context, capacity, information, spectrum of actor involvement, independence from power agendas and leadership. Overall in different contexts, social accountability mechanisms are reported to have enhanced efficiency in service delivery, increase responsiveness and establish and strengthen links between citizens and the system. However, there is no common blueprint to ensure accountability mechanisms are viable and effective. Therefore, conclusions should be cautious as many of the papers do not reliably assess the outcomes of social accountability interventions (final or intermediate) using sufficiently rigorous qualitative or quantitative methods, often focusing on the process indicators or observations of those involved in these.

Policy implications
The learning from the bottom-up accountability approaches reviewed in this study suggests that they are promising and often lead to positive effects locally; however, institutional and policy support can be important for implementing the approaches in a sustainable manner. This is especially helpful when the traditional 'carrot and stick' approaches to contain poor governance and corruption in the health sector appear to be ineffective and inconsequential, and instead more conciliatory and participatory approaches involving multiple stakeholders are considered.

Strengths and limitations of the study
The review provides comprehensive information on the processes and experiences of social accountability interventions practised in selected countries in Southeast Asia. Strengths include the approach adapted to searching the literature, which included published peerreviewed articles and grey materials and included a broad range of study designs, allowing us to provide a detailed map of the evidence on governance and social accountability. While offering insights by identifying key themes in this research area, it highlights the need for a full systematic literature review to arrive at conclusive findings. The review was restricted to English language documents only, and only three search engines were used. No formal quality assessment of the included sources was conducted. However, three researchers screened the papers and materials for appropriateness to the research question, working individually and then in a group, to review and synthesize data, maintaining consistent application of the inclusion and exclusion criteria. This article mainly focused on presenting innovative approaches practised, but not all articles provided sufficient information to report. Resource constraints prevented us from conducting a wider search, especially covering country-specific literature produced by non-academic and civil society stakeholders at the country level, which would undoubtedly be relevant. However, the review enabled us to explore a broad range of innovative approaches practised for improving governance and accountability in health systems, which have relevance to LMICs worldwide.

Funding
This publication is an output of the SOAS Anti-Corruption Evidence (ACE) research consortium funded UK Aid from the UK Government [Contract PO 7073]. The views presented in this publication are those of the author (s) and do not necessarily reflect the UK government's official policies or the views of SOAS-ACE or other partner organizations.