Facilitators, Barriers, and Strategies for Health-system Guidance Implementation: A Protocol for a Critical Interpretive Synthesis

Background: As systematically developed statements about possible courses of action, health-system guidance (HSG) can assist with making decisions and developing policies to address problems or achieve goals in health systems. However, there are conceptual and methodological challenges related to HSG implementation due to the complexity of health systems and policymaking, the diversity of available evidence, and vast differences in contexts. To address these gaps, we aim to develop a theoretical framework for supporting the implementation of HSG as part of an effort to promote evidence-informed policymaking about health systems. Methods: To develop a theoretical framework about facilitators, barriers, and strategies to the implementation of HSG, we will apply a critical interpretive synthesis (CIS) approach to synthesize the ndings from a range of relevant literature. We will search eleven electronic databases and grey literature websites to identify relevant published and grey literature. We will check the references of included studies and papers recommended by experts. Finally, we will conduct purposive searches to identify literature that lls any identied conceptual gaps. We will use relevance and a general ve-items quality criteria to assess included papers. A standardized form will be developed for extracting information. We will use an interpretive analytic approach to synthesize the ndings from included papers, including the constant comparative method throughout the analysis. The literature screening and relevance assessment will be conducted by two independent reviewers and disagreements will be resolved through discussion. Data extraction and synthesis will be extracted by the principal investigator and the sample of extracted data will be checked by a second reviewer for consistency and accuracy. Discussion: A new theoretical framework about facilitators, barriers, and strategies for HSG implementation will be developed using the CIS approach. The HSG implementation framework could be widely used for different HSG with varied topics and in different contexts (including low-, middle-, high-income countries and settings). In later work we will develop a tool for supporting HSG implementation based on the theoretical framework.


Background
Health systems can be understood and conceptualized from numerous perspectives. According to the World Health Report 2000 issued by World Health Organization (WHO), health systems have been de ned as "all the activities whose primary purpose is to promote, restore or maintain health" (WHO, 2000), which often were speci cally categorized in six "building blocks": 1) service delivery, 2) health workforce, 3) health information systems, 4) access to medical products, vaccines and technologies, 5) nancing, and 6) leadership and governance (WHO, 2007). To better clarify health systems and support the conduct of related research with a more comprehensive and operational theoretical framework, a panel commissioned by the WHO Alliance for Health Policy and Systems Research (HPSR) systematically analyzed current frameworks about health systems and proposed a new way to de ne health systems, which refers to "the governance, nancial and delivery arrangements for health care and public health services, implementation considerations for reforming or strengthening these arrangements, and broader economic, legal, political and social contexts in which these arrangements are negotiated and operate" (Hoffman et al., 2012). However, only de ning health systems in a sound and conceptual way is not going far enough for achieving the "quadruple aim" (namely, improving patient experiences, improving population health, keeping per capita costs manageable, and keeping providers satis ed) (Bodenheimer and Sinsky, 2014) or the health targets of the Sustainable Development Goals (SDGs). According to the World Health Statistics 2019, there were some improvements in health-related SDG indicators, for example, the increase of global and healthy life expectancy (separately by 5.5 and 4.8 years), and vaccination coverage rates; the decrease in neonatal deaths and deaths in children aged under 5 years, etc. However, there were still some disparities and inequities in heath outcomes among low-, middle-, and high-income countries, some health gaps in different genders, and some room for greater improvements (WHO, 2019).
On the one hand, policymakers, managers, providers, patients/citizens, researchers, and other stakeholders at global, national and local levels have continuously and collaboratively worked together to make efforts to achieve these health goals and targets. Moreover, more and more effective, safe, and cost-effective health interventions are becoming available. However, some key actions still need to be proposed and implemented, such as investments in health information systems, and measures that make health systems more responsive and resilient (WHO, 2019). Multiple studies showed that weak health systems could hinder the delivery and implementation of effective interventions (Bhutta et al., 2010;Travis et al., 2004).
Just as physicians use clinical guidelines in practice, policy makers can develop or revise policies with the help of health-system guidance that generally contain proposed options and their supporting evidence. Health-system guidance has been de ned as "systematically developed statements produced at global or national levels to assist decisions about appropriate options for addressing a health system challenge in a range of settings and to assist with the implementation of these options and their monitoring and evaluation" (Bosch-Capblanch et al., 2012). Given the early phase of health-system research and HSG, there were great varieties in terms that being used in the titles of HSG documents, such as "policy guidance", "policy recommendations", "guideline", "guidance", "recommendations", "policy", "opinions", etc. Therefore, when identifying HSG documents, the recommendations should be reviewed rather than the titles (i.e. whether the recommendations focused on one or many aspects of health-system arrangements). In our study, we adopted the above de nition of HSG from Bosch-Capblanch et al. to identify HSG documents and categorized them according to development process and the proportion of recommendations related to health-system arrangements. Considering the complexity of health systems (its components and their multidirectional interactions), diversity of available evidence about health systems, vast differences in contexts, and multifactorial policymaking process about health systems (Gauvin, 2014), a group of experts identi ed and analyzed the existing conceptual, methodological and practical challenges of HSG (Bosch-Capblanch et al., 2011; Bosch-Capblanch et al., 2012). A growing number of researchers and stakeholders from different institutions and governments, worked together to develop transparent and rigorous processes and approaches to ensure the high-quality process of development (Ako-Arrey, 2015) and contextualization (Alvarez, 2016). The rigorous development and appropriate contextualization of HSG precedes the implementation of guidance at national or sub-national levels, which are certainly prerequisites of successful implementation of HSG.
When HSG is produced at the global level, it is generally used to support decisions or policies by national (or sub-national) governments and by global organizations. For example, WHO produces HSG to be used by its member states as well as its own staff in their work with member states. Lavis et al. argued that HSG can present key features of related health and political systems. Also, they called for a complementary research agenda on analyzing more system and political considerations to ensure policy development is informed by HSG (Lavis et al., 2012). In addition, the HSG development handbook panel proposed that the HSG implementation was not only related to national (or sub-national) needs, priorities and resources, but also need to consider implementation facilitators, barriers, and supporting strategies, even some of which should be early considered by global guidance developers and local guidance users There is a need for a comprehensive, systematic, well-organized theoretical framework and practical tool to support the implementation of global HSG at national or sub-national levels. The rst step to the program of research, and the focus of the paper, is to conduct a knowledge synthesis of the published studies and grey literature to identify facilitators, barriers, and strategies-at the individual, organizational, community and system level-related to HSG implementation. The synthesis results are used for providing the foundational elements of HSG implementation theoretical framework for the research community.
The objective of our study is to develop a theoretical framework by iteratively and exibly examining literature and interpreting results, which makes CIS an optimal approach. The compass question that guides an iterative and qualitative synthesis approach is a key feature of the CIS approach. Our overarching objective is to develop a theoretical framework about the facilitators of, barriers to, and strategies of HSG implementation at different levels (individual, organizational, community and system) based on the following two compass questions: 1. What factors -at the individual, organizational, community and system level -facilitate or hinder HSG implementation processes and outcomes? 2. What strategies -at the individual, organizational, community and system level -can leverage facilitators and address barriers to HSG implementation processes and outcomes?

Methods
We will use a CIS approach to develop a theoretical framework that incorporates all facilitators, barriers, and strategies about HSG implementation, explores relationships among the above factors and strategies, and provides an overall explanatory theory for HSG implementation.
CIS, as one type of knowledge synthesis, was developed to address the limitations of traditional systematic review methodology (such as pooling data based on the comparability and homogeneity) by combining with the qualitative inquiry technique (Dixon-Woods et al., 2006). The CIS is used to identify and develop concepts and theories through an inductive, interpretive, and iterative process. The output of a CIS will be the theory grounded in included studies from a large, diverse, and complex body of literature, rather than data aggregation. Therefore, a CIS is conceptual in both its process and output ( . The literature related to HSG is highly heterogeneous with diverse study designs. Also, as a nascent domain, HSG is often mixed in clinical practice guidelines or public health guidance, or included in "general" guidelines (Ako-Arrey, 2015). Second, CIS is used for developing concepts and/or theories based on a detailed inspection of literature, an iterative and exible process of inquiry, and an inductive interpretation for developing concepts and theory (Dixon-Woods et al., 2006). Compared with CIS, traditional systematic reviews prefer to follow a pre-determined protocol to aggregate data or integrate results, and scoping reviews are mainly conducted when examining evidence distribution, boundaries, and gaps (Schick-Makaroff et al., 2016).
The CIS approach will be conducted by following ve steps:

Identifying potentially relevant articles
We will conduct the literature search in phases while being guided by the two compass questions. The search strategy, including the search sources and terms, was developed through multiple consultations with a librarian.
Data, and Web of Science. Also, we will search the following grey literature websites for unpublished literature: BIGG (international database of GRADE guidelines), Guidelines International Network (GIN), Health Systems Global (HSG), National Institute for Health and Clinical Excellence (NICE), National Institutes of Health (NIH), Scottish Intercollegiate Guidelines Network (SIGN), and World Health Organization (WHO). We will check the references of included studies and papers recommended by experts. Finally, we will conduct purposive searches to identify literature that lls any identi ed conceptual gaps (Wilson et al., 2014).
The search terms were seen in Box 1. The detailed search strategy for each electronic database was provided in Appendix 1.

Selecting potentially eligible articles
All reviewers (QW, XJC, YZ, JYZ) will conduct the pilot screening for reaching consensus on inclusion and exclusion criteria. If necessary, the inclusion and exclusion criteria might be changed based on pilot screening results and discussion with the program members. All paired reviewers will independently screen titles and abstracts, full texts of all identi ed literature against our criteria to classify each article as directly relevant literature, indirectly relevant literature, and excluded literature. The direct and indirect literature will be included in our sample pool of literature, from which we will select our purposive sample for the synthesis. Disagreements will be resolved through discussion or consulting a third reviewer.
While a qualitative synthesis approach like CIS does not typically involve the use of a quantitative measure like Kappa, to improve re exivity, we will use Kappa to calculate interrater reliability for measuring the agreement among reviewers. We will accept Kappa value 0.61-0.80 as "substantial" agreement and 0.81-1.00 as "perfect" agreement for literature screening in our CIS. (McHugh, 2012) During the pilot screening phase, if the Kappa result is below 0.61, it shows the reviewers need a meeting for discussing eligibility criteria and pilot screening for another sample before formal screening. After reaching "substantial" agreement, we will start formal screening.
We will include all empirical and non-empirical articles on their study content, i.e. those related to the facilitators of, barriers to, and strategies to supporting HSG implementation at any level (individual, organizational, community and system). In case there is scarce literature directly relevant to HSG implementation processes based on the initial search and opinions of content experts, we will consider including the papers relevant to the implementation factors and/or strategies of clinical practice guidelines (CPG) or public-health guidance (PHG). We will call the former "direct literature" and the latter "indirect literature". The implementation object or research subject of potentially included papers can be categorized as follows: HSG focused: meet the HSG de nition proposed by Bosch-Capblanch et al. (as mentioned above); or all or most of recommendations were related to health systems, i.e. the recommendations can be coded with Health Systems Evidence Database's taxonomy of governance, nancial and delivery arrangements in health systems (hereinafter referred to as "HSE framework", Appendix 2). "Formal" HSG focused: developed using a robust process, such as WHO guidance. Several HSG examples developed by WHO are shown in Table 1.
"Informal" HSG focused: organization documents and/or reports, such as World Health Assembly (WHA) documents. HSG relevant: CPG or PHG that including health-system recommendation(s) or containing a chapter or a section about health systems.
Moreover, to achieve a comprehensive search result, we will not place any restrictions in the following aspects: time frame, context (including low-, middle-, high-income countries and settings), study design (including peer-reviewed publication, conference abstracts, thesis and dissertation, editorials, comments, opinions, letters, correspondence, etc.), and language.
To better conduct the screening, a detailed description (Table 2) about inclusion and exclusion criteria was developed based on pilot screening by the principal investigator (QW) and discussion among the research program members.

Assessing relevant articles
The relevant articles will be assessed independently by two reviewers (QW and XJC) and disagreements will be resolved through discussion or consulting a third reviewer. A critical interpretive synthesis does not typically include a quality assessment but instead uses a relevance assessment. We will use relevance and a general ve-items quality criteria to assess the included papers. First, we will use a exible relevance boundary to comprehensively include the papers that contribute to generate concepts and theory. The relevance criteria could be de ned as the ability or the contribution to provide the concepts, theories, or insights to answer the compass questions.

Extracting data and information
Data will be extracted by the principal investigator and the sample of extracted data will be checked by a second reviewer for consistency and accuracy.
We have developed a standardized form (Appendix 3) to extract the characteristics of each included paper, including title, publication year, last name of rst author (if applicable, or development organization), publication form (journal article, or grey literature such as government or organization report), study design (if applicable), country focus (if applicable), and the object of implementation/research subject (e.g. HSG or indirect evidence related to CPG or PHG).
We will extract key ndings from each included paper by writing brief (e.g. one-paragraph) summary and extract the HSG implementation facilitators, barriers, and strategies at four different levels. The facilitators, barriers and strategies will be identi ed if they are explicitly mentioned or referenced in the full text. Sometimes, strategies would be deduced based on the implications of the identi ed facilitators and barriers and the study team's accumulated understanding or insights about HSG implementation eld (Khalid et al., 2020). We will extract facilitators of, barriers to, and strategies to supporting HSG implementation and organize them in the standardized form (Appendix 3), which will further classify ndings based on three frameworks (Appendix 2, 4, 5). For the organizational and community levels, we will extract all implementation facilitators, barriers, and strategies that exit or occur at the organizational or community settings. For example, the organizational level will include all organizations that being involved in health systems, such as organizations for providing care (such as hospitals, clinics, pharmacies), organizations for providing funds (such as donor agencies), and other related international and/or non-governmental organizations (such as WHO, Red Cross). We de ne the community as the local or district health system. Community care programs are different among different communities and/or population groups, and some examples include supportive housing, adult day programs, retirement homes, residential hospices, exercise and falls-prevention programs, etc. (Lavis, 2016).
For the system level, we will separate two sub-categories, including health system and political system. For the aspect of health system, we will use HSE framework (as a set of variables, including governance, nancial and delivery arrangements within health systems and implementation strategies, Appendix 2) to extract and organize the ndings at health-system level.  For the aspect of political system, we will use framework of policy development and implementation (3I+E, "Institutions, Interests, Ideas and External factors"), as a set of variables that can help explain the ndings at the political system level (Appendix 4) to categorize the political-system ndings (Lavis et al., 2012; Lavis 2017; Gauvin, 2014).

Synthesizing and integrating ndings
We will use an interpretive analytic approach to synthesize the ndings from included papers, and to develop an explanatory framework of HSG implementation processes. We will conduct a constant comparative method (Glaser, 1965) throughout the analysis, until reaching the saturation in identifying new concepts and themes (i.e. facilitators, barriers, and strategies) and new connections among them.
We will conceptually map the relevant papers. Speci cally, we will use a 3 × 4 matrix to categorize the ndings of included papers, which cross links three aspects (facilitators, barriers, and strategies) and four levels (individual, organizational, community and system). The conceptual mapping exercise will be used to categorize the literature into domains and topics of interest for further assessment, abstraction, and coding. After getting the initial results of conceptual mapping, we will probably identify some conceptual gaps, and further conduct purposive sampling to identify relevant literature. The sampling process will be stopped when reaching the saturation, i.e. the new literature cannot provide additional new content to our framework. However, the purposive searches and sampling could be considered as dynamic and iterative, i.e., the new conceptual gaps could be being identi ed and lled in the synthesis process and integration phase of ndings (Wilson et al., 2014). During the process, the iterative discussion and feedback among the program members will be conducted.
As noted above, we will use the TDF, HSE framework and 3I+E framework to organize the ndings. The analysis, synthesis and integration processes will be iteratively conducted with constant feedback from the consultation with our program team with interdisciplinary expertise and experience (Wilson et al., 2014;Khalid et al., 2020). We will use the Microsoft Excel 2010 and NVivo version 9 software during the analysis and synthesis process. This will involve: 1. identifying common themes and concepts based on included papers; 2. developing theoretical constructs based on the emerging themes and concepts; 3. critiquing the emerging theoretical constructs as a whole and with our full sample of literature to identify conceptual gaps in the available evidence in relation to our principal aims; 4. conducting additional purposive sampling of included papers and/or conducting additional purposive searches to ll conceptual gaps (if needed) until theoretical saturation is reached; and 5. integrating the theoretical constructs into a "synthesizing argument" about HSG implementation processes (i.e., an explanatory framework).

Discussion
We will use a CIS approach to develop a theoretical framework that incorporates all facilitators, barriers, and strategies about HSG implementation at four different levels (individual, organizational, community and system), explores relationships among the above factors and strategies, and provides an overall explanatory theory for HSG implementation. The study will have the following strengths and challenges.

Strengths
Firstly, a CIS approach is appropriate for developing a theoretical framework based on diverse and complex body of literature (Dixon-Woods et al., 2006). Given the characteristics of literature on policymaking and implementation about health systems, a CIS approach could be used for generating a theoretical framework about HSG implementation. Secondly, the output of our study will be the rst implementation framework specially for HSG, given the complexity of health systems and policymaking, the diversity of available evidence, and vast differences in contexts, which will ll the research gap about HSG implementation. In addition, this framework will be widely used for different HSG with varied topics and in different contexts (including low-, middle-, high-income countries and settings). Thirdly, comparing with a descriptive conceptual framework, a theoretical framework for HSG implementation that includes different factors and strategies, could provide clear relationships and connections among these elements. Finally, the development of an HSG implementation framework will not only provide strong theoretical support for developing an HSG implementation tool, but also contribute to the research community of HSG and/or implementation.

Challenges
The anticipated biggest challenge might be the process to synthesize the ndings from a complex and diverse body of literature. Our response to this challenge is to enrich the CIS approach with ongoing input from interdisciplinary study team. Also, according to our pilot search and screening, the literature directly relevant to HSG might be scarce and seems not provide the comprehensive ndings for the framework because the research of HSG implementation is still in the beginning stage and is just the small part of health systems research or implementation science. One response to this potential challenge is to expand to include the literature not directly relevant to HSG implementation but very closely or similarly to our focus. For example, the "indirect evidence" includes the papers relevant to the implementation of CPG and/or PHG. Again, we will rely on our study team to re ne the framework based on their interdisciplinary expertise and experience.

Next step
In later work we will develop an HSG implementation tool based on the theoretical framework using adapted Delphi method and validate the tool.  ✓ The implementation process (any facilitators, barriers, strategies, etc.) of HSG (generally or speci cally) · Implementing health systems guidance: a workbook to support the contextualization of recommendations at the national or subnational level · Handbook for developing health systems guidance: supporting informed judgements for health systems policies