A theoretical and methodological framework for analyzing long-term policy and the role played by actors in promoting change and stability

Public policies are often associated with historical legacies constituted by political processes in which actors seek to promote institutional change or maintain stability. This article presents a theoretical and methodological framework that uses process tracing to identify the most relevant processes for the study of the trajectory of policy through a theoretical lens that draws on studies of the public policy process and the theory of gradual institutional change. The framework enabled us to identify changes in medical workforce policy (MWP) between the 1960s and 2010s and understand the influence of different actors and institutional arrangements. Our findings also reveal the importance of the role of ministry of health executives interested in tailoring MWP to the needs of Brazil’s public health system from 2003


Introduction
This article presents a theoretical and methodological framework for analyzing the actors and institutional arrangements that influence long-term policies (c) . The model is used to analyze the More Doctors Program (MDP) within the trajectory of medical workforce policy (MWP) in Brazil; however, the framework is proposed for the analysis of long-term policies in the health sector and other areas.
A review of the literature on health workforce policy (HWP) suggests that most studies, in particular those investigating MWP, focus on human resource issues 1,2 , solutions to human resource problems 3,4 , and government actions over a given period of time 5 . On the whole, the studies employed theoretical frameworks that are widely used in the area of human resources for health, often in combination with elements from the sociology of professions, but rarely drawing on theoretical frames of reference from the field of public policy analysis. Studies investigating specific programs prioritize the analysis of national regulatory norms and standards governing the initiatives, considering their objectives, meanings, and potential or actual effects and the context in which the programs were created.
The analysis of the long-term trajectories of policies such as MWP shows that many apparently new state actions, such as programs, regulatory changes and decisions, have already been proposed, analyzed, and frequently blocked at an earlier stage. They are not actually "novel", but rather historic legacies brought back into public debate at a time when the blocking of the proposal by actors interested in maintaining the status quo could be effectively challenged by actors interested in change. To understand this political process, it is necessary to characterize the actors interested in maintaining the institutional status quo and those interested in change in different historical contexts and to identify their positions, interests and projects for the area of health, ideas, and long-term modes of operation. In addition, it is necessary to analyze the institutional arrangements and combination of circumstances that drive or hold back change over time in order to understand how these factors influence the strategies and resources employed by these actors.
The theoretical and methodological framework proposed here is capable of identifying and treating evidence that refers to long stretches of time and of supporting the analysis of the role of social actors, individuals, and groups, examining the relationship between the results of this analysis with the blocking of or effecting institutional change.
The model draws on studies of the public policy process 6 and the theory of gradual institutional change (TGIC) 7 . The former supports the analysis of social relations between organizations and actors (individuals, groups, and the government) over time; the latter contributes to the analysis of institutional change considering prevailing institutional arrangements, the changes made, political context, and strategies employed by the actors to foster change or maintain institutional arrangements in the face of the restrictions and opportunities created by the circumstances.
(c) This article is the fruit of a doctoral research project analyzing 60 years of MWP in Brazil, the most important changes that took place during this period, and the factors that drove these changes.
In the following section, we present the theoretical and methodological framework, outlining its theoretical underpinnings and underlying concepts and the methods proposed for long-term policy analysis. A literature review was conducted to support, conceptualize, and discuss the model's theoretical and methodological underpinnings. The last section demonstrates the application of the framework by using it to analyze the MDP within the trajectory of MWP in Brazil. The specific methodology used for the analysis is detailed in the second section. The present study was conducted in accordance with the relevant ethical norms and standards and approved by the research ethics committee (reference code CEP-UFRB-05760818.9.1001.0056). We conducted a literature review and performed a documentary analysis of legal and administrative norms and standards governing MWP programs and information published in the mainstream media and on the websites of civil society organizations, focusing on the period 2003 to 2018. We also conducted 19 semi-structured interviews with key informants who played a leading role in the formulation of MWP in Brazil (Frame 1). The interviews were recorded and transcribed, and the interviewees are identified using Arabic numerals.

A theoretical and methodological framework for analyzing long-term policies
One of the challenges of analyzing the trajectory of a policy that stretches over decades is selecting the appropriate method to identify, analyze, and synthesize the evidence. The model proposed here uses process tracing 8 , which allows the researcher to examine trajectories, documents, interview transcripts, and other sources to determine whether possible explanations derived from theories are valid or should be refined or modified considering the range of variables that may influence outcomes. This method seeks to identify causal chains and mechanisms, understood as theoretical constructs formulated by researchers that focus on dimensions that may theoretically influence or determine a given event or phenomenon. The aim is to formulate middle range theories that explain the event or phenomenon.
Another important element of policy analysis is the definition of the mesosocial space that the study intends to investigate, conceived as a social order influenced by the broader political system and socioeconomic structure, with relative autonomy, specific rules, and certain institutional arrangements that influence and are influenced by the actions of social actors depending on their position in the social order. In the model proposed here the mesosocial space was understood to be a subsystem, defined as a policy-producing sectoral unit, which in turn constitutes a stratified power structure that has unequal implications for political and material resource allocation and provides an arena for debate between actors who advocate different solutions to public policy problems 9,10 .
When analyzing a policy subsystem, a part of a subsystem, or interfaces between different subsystems in any given period, the focus should be the characterization of the trajectory of the policy, considering the historical legacies and other factors that have most influenced this trajectory related to at least two explanatory dimensions: the actions of actors who promote/have promoted stability or change; the prevailing institutional arrangements in the subsystem and the changes these arrangements have undergone. Frame 2 shows the characterization of the trajectory, analytical dimensions, and sources of evidence used in the present study.  A wide range of sources can be used depending on the object of study and the issues and hypothesis being evaluated, including laws, decrees, executive/ministerial orders and resolutions, media reports, and channels of communication used by the government and civil society organizations. Interviews with key actors are another important source, especially when scarce documentation gives limited insight into the positions, ideas, projects, and actions of the actors disputing the policy in question or is insufficient to understand the elements of institutional arrangements that significantly influence political process. To analyze the above documents and interview transcripts, in addition to content analysis, a widely use technique in public health research 11 , we also used critical political discourse analysis 12 , which is particularly suited to the study of the objectives, values, concerns, and circumstances of the actors involved in the political dispute (politics) over a public policy (policy).
In addition, the model also draws on the theoretical underpinnings of studies investigating political process (public policy process) 6 and the TGIC 7 . The former supported the analysis of the role of individuals, groups, and governments, focusing on their objectives, interests, ideas, and involvement in policy formulation in contexts in which rules, conceived as institutions, demarcate the possibilities of defining public policy problems and solutions.
The model employs the concepts of policy community 13 , issue networks 14 , and policy entrepreneurs 15 to identify and analyze the actors that have most influenced the policy's trajectory. A policy community is a more or less cohesive group of individuals and/or organizations with different positions and relationships between its members. These communities share common objectives, interests, and ideas, specialize in certain issues and sectoral policy designs and outcomes, and act in a coordinated fashion to influence decision-making processes and ensure their positions dominate the government agenda 13 . Issue networks are groups that come together to discuss, formulate, and propose solutions to a specific issue. They are characterized by unequal power relationships and few resources, fluctuating interaction, absence of consensus, and presence of conflict 14 . Policy entrepreneurs are key individual actors who exercise leadership and engage in collaborative efforts to create or exploit means and opportunities to draw attention to the problems and solutions they defend. These actors seek to influence the political process in order to bring forth problems and solutions to the government agenda 15 .
The TGIC 7 is the heir to historical neo-institutionalism, which emphasizes historical legacies, based on the assumption that previous events establish parameters and affect decision-making and subsequent events, as well as the dynamics of the agent-structure relationship. This theory was used to support the analysis of institutions, conceived here as rules and distributional instruments laden with power implications and unequal implications for resource allocation and thus fraught with tensions. Consequently, dominant state and societal actors receive more resources to maintain or improve their privileged condition and fulfill their goals. However, to maintain the stability of institutional arrangements these actors need to promote the ongoing mobilization of political support, ensure that those responsible for applying the rules also comply with them, and make active efforts to resolve institutional ambiguities in their favor. Hence, compliance -understood as the effective conformity with and subordination to rules -is a crucial variable in the analysis of stability and institutional change. According to the TGIC, change can be triggered by different mechanisms and driven by both external and internal factors, resulting from a rupture of the prevailing balance. For the authors, the characteristics of institutional arrangements and the political context -marked by the power to veto that some actors and decision-making spaces exercise over potential actions -influence which change strategies have a greater or lesser chance of success.
Depending on the study objectives and the issue in question, other concepts and theories may be combined with this framework as long as they share the same ontological and epistemological assumptions as theories of the public policy process and the TGIC, or the ideas drawn from theories adopting other assumptions are resignified. The following section presents an analysis of the formulation of the MDP using the theoretical and methodological model proposed here. Interface (Botucatu) https://doi.org/10.1590/Interface.220464

Application of the theoretical and methodological framework: analysis of the More Doctors Program within the trajectory of MWP in Brazil
Most studies investigating the emergence of the MDP focus their analysis on the period surrounding its launch 16 and therefore fail to show that the political process of building MWP in Brazil has been marked by tensions between the needs of the country's public health system and resistance to policy changes to address these needs. Likewise, research has paid little attention to the fact that past efforts to effect change, with limited results, have focused on the prevailing elements of previous programs that were incorporated into the Program's design. By examining MWP over time it was possible to reconstitute the process by which these historic legacies were constructed, with the results of the literature review showing that some of the programs and specific actions that have shaped MWP in the country were proposed as far back as the 1960s 3, 16 . This section presents the application of the framework to analyze the stability of MWP overtime, focusing on the period up to 2013, when policy underwent only incremental changes, the main changes in policy between 2010 and 2013, and the period after 2016, when policy largely reverted back to the situation in place before 2013.
The results of the analysis show that health and education policy subsystems, understood as sectoral policy-producing units, are decisive for program formulation, the decision-making process, and the production of rules and regulations governing MWP, which can also be observed in other countries 17 . The use of the concept of subsystem meant that the analysis was not restricted to government actions and the actions and reactions of medical entities, providing insights into the complexity of the policy-making environment in which actors and institutional spaces from the two subsystems participate. We identified a diverse range of actors in both subsystems, including legislators, heads of executive branch agencies and other public institutions, academics, and leaders of civil society organizations. These findings are consistent with other studies [18][19][20] . The model also shows that among the government executives, legislators, doctors, and academics, there are those who advocate maintaining the status quo and those who propose changes to respond to the needs of the Brazilian National Health System (SUS).
The use of the concept of policy community was fundamental to understanding how the members of communities interacted and provided important insights into their ideas and differences. The findings show that three policy communities influenced the trajectory of MWP the most: the health reform movement (HRM-PC), advocates of market regulation (MR-PC), and advocates of liberal medicine (LM-PC). The use of the concept of issue networks allowed us to identify and characterize the medical education network (ME network) and policy entrepreneurs, providing valuable insights into the actions of individual actors who played a leading role in the formulation of programs and in the coordination of the policy communities and issue network 16 .
The HRM-PC advocates that the SUS should formulate HWP in accordance with system needs. The results of the analysis showed that there are three subgroups within the HRM-PC. The first, commonly called "the health movement", brings together specialists, researchers, and academics and is spearheaded by entities such as the Brazilian Association of Collective Health (ABRASCO), Brazilian Center for Health Studies (CEBES), and United Network. This subgroup has defended health reform since the 1970s, prioritizing changes in structural aspects of the system that limit the implementation of the SUS in accordance with principles set out in the country's constitution. The second subgroup is made up of individuals and groups involved in the health sector, mainly linked to the executive branch, but also including the legislative branch. It is led by SUS managers, largely ministry of health executives and officials from the organizational units of the National Council of Health Secretaries (CONASS) and National Council of Municipal Health Secretaries (CONASEMS). They focus more on defending program ideas and solutions to day-to-day management problems in the SUS. The third subgroup is led by union and civil society organization leaders, which, according to Côrtes 19 , since the 1990s, have focused their attention nationally to the National Health Council (CNS) and began to defend proposals more geared towards the specific needs of health workers and patient organizations.
The MR-PC is made up of actors linked to the medical-industrial complex, financial capital in health, private university holding companies, and private higher education institutions (HEIs), and their supporter in the media, academia, and executive, legislative, and judicial branches. This community defends that private health services and education should be regulated by the market with minimal state intervention and without outside controls on workforce pay and medical services. According to the MR-PC, distribution, pay, scope of practices, and health professional education and training, including curriculum content and the number and location of under and postgraduate degree places, should be regulated by the market.
The LM-PC is made up of the following actors: medical entities (primarily the Federal Council of Medicine and Brazilian Medical Association); actors (generally doctors) involved in health and education management at federal level, particularly collegial bodies, committees, and administrative departments tasked with discussing "medical issues"; legislators, also generally doctors; health service managers, especially hospital directors; and doctors working in HEIs. This community aims to maintain its dominant interests and benefits [21][22][23][24][25][26] and opposes changes in prevailing MWP. The group defends the use of non-market based mechanisms to control pay and positions itself against state and market regualtions [21][22][23][24][25][26] . The LM-PC defends that medicine should be a liberal profession, calls on the state to control the private education market and prevent an increase in the number of undergraduate degree and residency places. While defending on the one hand legislation that protects -or broadens -the prerogatives of the profession, the group believes that the government should play a minimal role in professional regulation, the definition of the scope of professional practices, the provision of doctors, and the definition of curriculum content.
The results of the analysis also enabled the identification and characterization of policy entrepreneurs that played a leadership role in the policy communities and issue network, promoting both the maintenance of the status quo and changes in MWP. These actors played a decisive role in creating various programs, including the degree validation system (Revalida), the Primary Care Professional Recognition Program (PROVAB), and the MDP. In contrast, the lack of support from these individuals for mandatory civil service is one of the reasons why the bill was not approved 16  The theoretical and methodological model showed that the policy entrepreneurs responsible for changes in MWP, especially between 2011 and 2016, were members of the ME network, former members of the DENEM, and adversaries of members of the LM-PC since the 2000s. Acting in different institutional spaces, they had similar professional backgrounds, shared the same ideas about medical education as a priority for improving the SUS, and followed the same course of political action. The latter was marked by various conflicts: with leaders of the LM-PC in the ME network; within the CINAEM; during the elections held by medical entities; and related to positioning with regard to the Medical Act and the creation of PROVAB and the MDP.
The framework also enabled us to analyze the institutional arrangements of the education and health policy subsystems, particularly those related to HWP and MWP. With regard to HWP, the literature shows that even in countries with a strong tradition of state intervention like France, the state delegates certain regulatory functions to professional organizations 17,21,22,25 . In Brazil, it is the National Congress that legislates on professions, defining the scope of practices, licensing, exclusive tasks, and wage ceilings. Professional education and training is regulated by the Ministry of Education, which, via the National Education Council (CNE), establishes prerequisites, curriculums, and general norms and standards. The National Medical Residency Commission (CNRM), a collegial body belonging to the Ministry of Education, does the same for medical residency programs. The Ministry of Health is not actively involved in decision-making concerning health professions, with its role being limited to issuing reports, which the National Congress and Ministry of Education may or may not accept, and participating in Ministry of Education decision-making forums on medical education. Even the CNS, attached to the Ministry of Health, plays a lesser role than established by the legislation. The law governing each profession generally defines that regulation and oversight of professional practice are carried out by the respective professional councils, as is the cases with the Federal Council of Medicine (CFM). Although professional councils are autonomous state bodies whose directors are elected by council members, they have major difficulty understanding this and behaving like an agent of the state, tending to protect their own interests and benefits 21 .
The results of the analysis of institutional arrangements throughout the course of MWP since the 1950s shows that the influence of the medical profession stretches beyond functions legally delegated to the CFM and regional councils of medicine. For example, in decision-making concerning professional undergraduate degree curriculum guidelines, it was shown that greater weight is traditionally given to the guidance provided by professional organizations and technical advisory bodies are made up of people recommended or not vetoed by these organizations. Within the CNE, there are bodies that are responsible for analyzing and deliberating over health issues. The same does not occur however in the Ministry of Education, which employs "committees of experts" with advisory functions that conduct studies and issue opinions on a diverse range of educational topics and issues. These committees also provide input for decision-making concerning curriculum guidelines, evaluation instruments, and criteria and measures for creating and closing courses. When the question is deemed to be a "medical issue", ministry of education decision makers generally rely on these committees and specialist government health agencies for expert advice, as was the case with the former Department of University Hospitals (incorporated into the state-owned Brazilian Hospital Services Company in 2011) (interviews 1; 2; 3; 6; 9; 16).
With regard to medical residency programs, while regulatory decision-making is the responsibility of the Ministry of Education, this process takes place through the CNRM, a collegial body whose function is to regulate, supervise, and evaluate programs and the institutions that run them. Between 1985 and 2011, the CNRM was composed of nine members, four of which represented the government, and five medical entities -CFM, ABEM, Brazilian Medical Association (AMB), National Association of Resident Doctors (ANMR), and National Federation of Doctors (FENAM). The CNRM is an institutional space controlled by medical entities, whose decisions tended to protect the interests of the profession, paying little attention to the goals of the SUS 4,16 . As a rise in the number of professionals over and above a set level tends to reduce the price paid in the market for specialist services, the control and limitation of specialist training is among one of the LM-PC's main concerns, hence its interest in maintaining the status quo [22][23][24][25][26] and resisting changes in the composition of the CNRM proposed by the HRM-PC.
The results of the analysis of MWP show continuities resulting from historical, ideational, and institutional legacies, as well as incremental changes in institutional arrangements, blocked attempts to effect change, and strategies developed by key actors in the two subsystems interested in influencing policy. When attempts to effect change are made the forces that seek to maintain the status quo are forced to reveal themselves more than in periods of stability and act to maintain the prevailing balance. Hence it is important to focus on the moments when new topics find their way onto the government agenda, the formulation and decision-making processes that culminate in institutional change, the attempts -successful or not -to block change, and the strategies adopted by individuals and groups during this process.
The analysis of MWP between 2003 and 2010 reveals the roles played by actors and institutions, providing a better understanding of why, after the creation of the SUS in 1990 and with members of the HRM-PC taking the reins at the Ministry of Health in 2003, significant change in MWP only took place 2013, when the MDP was created 16,28 . During this period, ministry of health executives identified four points of veto of proposed changes: the National Congress, Ministry of Education, the president of the republic, and the LM-PC. The active opposition of the latter was majorly responsible for the resistance from the other points (interviews 1; 2; 3; 5; 6; 9; 10; 12; 13; 14; 16).
The power of veto of the National Congress manifested itself in the changes that required the authorization of the legislative branch. The Congress tends to be resistant to changes in legislation governing professions when the organizations that represent these professions are against these changes (interviews 1; 3; 4; 6; 9; 10; 12; 13; 15). In addition, the LM-PC is very influential in the Congress. Its members include doctors who are legislators and the community has the capacity to leverage positive mass media coverage, which greatly increases the political cost of the legal changes they are opposed to (interviews 1; 3; 4; 6; 7; 9; 10; 11; 12; 13; 15; 16). Indeed, the only initiative opposed by the LM-PC that was approved during the study period was the MDP law.
In view of the strong influence that the LM-PC also had on the Ministry of Education, ministry of health executives, which from 2003 were predominantly members of the HRM-PC, tried to transfer roles and responsibilities from the Ministry of Education to the Ministry of Health in the first years of the Lula government (2003)(2004), when the Department for the Management of Health Work and Education (SGTES) was created to broaden the role of the Ministry of Health in decision-making related to HWP. This effort was frustrated largely due to resistance from leaders of the education subsystem and opposition from the LM-PC (interviews 6; 9; 10; 14). The strategy adopted by ministry of health executives between 2005 and 2010 was to accept the limits set by ministry of education executives and formulate, in conjunction with the Ministry of Education, actions that envisaged the joint participation of both ministries (interviews 1; 5; 9). Between 2011 and 2016, the Ministry of Health's strategy was to strain these limits, modifying and creating new institutions in which power was shared between the Ministry of Education and Ministry of Health, sometimes involving, the CONASS and CONASEMS. An example are changes in the composition of the CNRM in 2011, including representatives from both councils, and the creation of the PROVAB and MDP in 2011 and 2013, respectively. The latter included governance forums with representatives of the Ministry of Health, Ministry of Education, CONASEMS, and CONASS (interviews 3; 11; 12; 15).
Strategic decisions made by ministry of health executives who were members of the HRM-PC culminated in changes in the trajectory of MWP and were influenced by positions taken up by the president's office and its environs (the Executive Office of the President and Department of Institutional Relations). Ministry of education executives had shown themselves to be contrary to the creation of degree validation measures between 2003 and 2009. In 2010, president Lula demanded that ministry of health and education executives resolved the issue. In the same year, Revalida was created, which, despite opposition from the LM-PC, established a centralized system for the validation of medical degrees awarded by universities outside Brazil (interviews 3; 5; 12; 13). The same happened with the change in composition of the CNRM and implementation of PROVAB and the MDP. The support provided by president Dilma and her advisers was an important factor in overcoming resistance from ministry of education executives (interviews 2; 3; 6; 11; 12; 13; 16). Recognized by the literature, the power of the presidential agenda is a factor in institutional change in sectoral subsystems, which, especially in Brazil, influence legislative agenda-setting, and moreover, the agenda of the executive branch itself 29 .
The period of greatest change in MWP -2011 to 2016 -brought together a series of factors that combined to rupture the balance that had provided notable stability to the prevailing institutional arrangements thus far. The main changes were as follows: changes in the direction of the federal government, positioning a group that was very interested in changing MWP; worsening of the doctor shortage in the SUS, giving even greater emphasis to this issue in the policy system; poor results of programs designed to tackle the problem; the political conjuncture in 2013; strategic actions developed by policy entrepreneurs supported by the president; and lessons learnt and ideational and institutional legacies produced during the formulation of previous programs, whether they were implemented or not 16,30 . The design of the MDP may have been influenced by ideational legacies stemming from studies investigating experiences in Brazil and around the world of initiatives to improve staffing levels and health professional education and training. Furthermore, the ideas, rules, and resources used to shape the MDP arose mainly out of actions resulting from the regulation of the law governing student loans (FIES), the implementation of PROVAB, both in 2011, and the National Medical Education Plan, created in 2012. The design was also influenced by historic international agreements between the Ministry of Health and Pan American Health Organization (PAHO), and an agreement made in 2013 between the Ministry of Health, PAHO, and Cuban Ministry of Health 31 . This agreement was strategic because it boosted national and international recruitment of doctors, which was previously imagined by local health managers to be insufficient to attract the necessary number of professionals (interviews 2; 3; 7; 12; 13; 15; 16). However, with the change in "destabilizing" conditions due to the impeachment of president Dilma, including the exit of members of the HRM-PC from the Ministry of Health, unprecedented strengthening of the influence of the LM-PC within the Ministry of Health in the Temer government, and the advent of the Bolsonaro administration, stability was restored and various changes that were made in the period 2010 to 2016 were undone or neutralized. PROVAB was discontinued in 2016 and Revalida was interrupted in 2017, only resuming operations in 2020, due to the imposition of legal measures by Law 13959/2019. Several measures of the MDP were administratively suspended, such as the expansion of public medical schools in 2016 and the international recruitment of doctors in 2019. Changes to the law governing the MDP made by the Temer and Bolsonaro administrations revoked provisions that sought to tailor medical education and training to the needs of the SUS, inducing the implementation of new curriculum guidelines for medical courses and creating a new roadmap for specialist medical training in Brazil. In 2019, much more influenced by the LM-PC than in previous moments during the study period, the Ministry of Health proposed and approved the creation of the Doctors for Brazil Program with the explicit support of this community, hailing it to be a replacement for the MDP. However, the government only began implementing the program in 2022, the last year of the Bolsonaro administration. Nonetheless, important historical, ideational, and institutional legacies have been built and may be used in new change processes that seek to reorient MWP and HWP towards the demands of the SUS and health needs of the Brazilian population.