Translating the International Code of Marketing of Breast‐milk Substitutes into national measures in nine countries

Abstract The International Code of Marketing of Breast‐milk Substitutes (the Code) adopted by the World Health Assembly (WHA) in 1981 and regularly updated through subsequent WHA resolutions, represents the international policy framework for protecting breastfeeding against inappropriate marketing practices. By March 2016, at least 135 countries had some measures covering provisions of the Code in their legislation. The translation of the International Code into national measures was investigated in the context of the advocacy efforts undertaken by the Alive & Thrive (A&T) initiative with UNICEF and partners. A real‐time evaluation was carried out over 22 months in seven Southeast Asian countries (Cambodia, Indonesia, Lao People's Democratic Republic [Lao PDR], Myanmar, Thailand, Vietnam, and Timor‐Leste) and two African countries (Burkina Faso and Ethiopia). Drivers of policy change and progress were examined. Two theory‐based approaches were used: developmental evaluation and contribution analysis. Data collection methods included participant observation, key informant meetings, in‐depth interviews, reflective practice, and desk review. Overall, countries made significant progress in translating the International Code into national measures and in moving forward throughout the policy cycle. The main driver of policy change was the creation of a strategic group, which engaged key relevant actors and supported the government in the performance of 15 critical tasks, which the analysis reveals is a second driver. Those critical tasks are described in this paper and could help public health advocates to anticipate the stages and challenges of policy change and develop more effective strategies to translate the Code into their legal framework.


| The International Code of Marketing of Breast-milk Substitutes
The International Code of Marketing of Breast-milk Substitutes (hereinafter referred to as "the Code") represents the international policy framework to protect breastfeeding against inappropriate and unethical marketing practices from manufacturers and distributors of products under the scope of the Code. 1 The Code was adopted in 1981 by the World Health Assembly (WHA) (WHO, 1981). It is regularly updated through WHA resolutions, and those should be considered together as far as the implementation of the Code is concerned (WHO, 2017a). In 2016, although at least 135 countries had some measures covering provisions of the Code in their legislation (WHO & UNICEF, 2016), its application remains challenging. Industry regularly circumvents the national measures and uses marketing practices that go against the Code (International Code Documentation Centre, 2017).
In 2014, the global sales value for infant and follow-up formula was about $U.S. 44.8 billion, of which it is estimated that some 10% is likely spent on marketing and promotion (Smith, 2015). This sales value is projected to reach $U.S. 70.6 billion by 2019 . Between 2008 and 2013, the East Asia and Pacific region had the highest volume increase, which was mainly driven by China, Indonesia, Thailand, and Vietnam . Given the profitability of this industry, the use of aggressive marketing practices is likely to increase, requiring governments to enact measures to protect their population against the deleterious effects of marketing on breastfeeding practices (Piwoz & Huffman, 2015;Sobel et al., 2011).
Research on the process of developing and implementing health policy in low and middle income countries has been quite limited (Gilson & Raphaely, 2008). This is also the case for the Code. There has been no empirical research on its full policy cycle, most studies focusing more narrowly on the challenges for its monitoring and enforcement (Aguayo, Ross, Kanon, & Ouedraogo, 2003;Parrilla-Rodriguez & Gorrin-Peralta, 2008;Taylor, 1998). Enforcement is known to be problematic, as seen in the widespread violations documented by International Baby Food Action Network (International Code Documentation Centre, 2017). However, to facilitate the translation of the International Code into national measures, a deeper and broader understanding of the other stages of the policy cycle is needed. The present study aims to fill this gap by presenting the results of a real-time evaluation of the related collaborative advocacy efforts carried out by Alive and Thrive (A&T) in nine countries with UNICEF, development partners and local governments. The overall objective was to document the extent to which policy objectives were (or were not) achieved in each country and to identify the key drivers of policy changes.

| Programmatic context
A&T is a 9-year initiative to improve infant and young child feeding (IYCF) practices. After a first phase of implementation in Bangladesh, Ethiopia, and Vietnam, during which considerable gains in IYCF practices were achieved Menon, Rawat, & Ruel, 2013;Rawat et al., 2013), the Initiative received additional funding for a second phase (2014)(2015)(2016)(2017). This second phase focused on sharing the successful experience of policy advocates in Vietnam to support governments and development partners in six countries within the region. The aim was to advocate either for the adoption of IYCF-friendly policies or for the implementation, enforcement, or monitoring of existing policies. Countries included Cambodia, Indonesia, Lao People's Democratic Republic (Lao PDR), Myanmar, Thailand, and Timor-Leste. At a regional workshop for Southeast Asia in 2014, each of seven country teams set specific policy objectives regarding three main areas: the Code, maternity protection, and health system strengthening. This phase also included two African countries, Burkina Faso and Ethiopia, which carried out advocacy efforts though their primary focus was on program implementation, and their work on the Code was less salient. In this paper, we focus on the Code and present the following: (a) key drivers of policy changes, and (b) major accomplishments achieved in the nine countries during Phase 2.
2 | METHODOLOGY 2.1 | Evaluation approaches This real-time evaluation used two complementary theory-based evaluation approaches: developmental evaluation (DE) and contribution 1 The Code applies to the marketing and practices related thereto, of the following products: breast milk substitutes including infant formula; other milk products, foods, and beverages, including bottle-fed complementary foods when marketed or otherwise represented to be suitable, with or without modification, for use as a partial or total replacement of breast-milk, feeding bottles and teats. It also applies to their quality and availability and to information concerning their use (Article 2. Scope of the Code, WHO, 1981).

Key messages
• Countries experience challenges in translating the International Code of Marketing of Breast-milk Substitutes into their legislation and in enforcing their legal measures.
• An advocacy approach involving a 4-part process can help set and maintain the agenda for the Code throughout the entire policy cycle.
• Policy advocates have been able to support governments to carry out a set of 15 critical tasks by creating a strategic group and engaging key relevant actors.
• Constant vigilance against industry tactics is required throughout the policy cycle. analysis (CA). DE supports the development and implementation of an innovation by collecting various types of data that help articulate feedback to adapt the innovation to the emergent and dynamic context (Patton, 2010). DE helped track the various actions in the nine countries and guide reflection on the advocacy efforts as they were evolving. CA allows for the assessment of whether an intervention contributed to the observed effects (Mayne, 2001(Mayne, , 2011(Mayne, , 2012presented in a companion paper, Michaud-Létourneau, Gayard, & Pelletier, 2019). CA acknowledges that factors other than the ones directly related to a project are at play and that they may significantly influence the outcomes. This was acknowledged for the evaluation of this Initiative. In sum, although DE helped primarily to collect various data and guide the reflection, CA helped to assess the contribution of various actions to the outcomes observed, utilizing the data collected using the DE approach. The ethics committees from the University of Sherbrooke and from FHI 360, the nonprofit human development organization that implement the A&T initiative, approved the research protocol. Participants were identified either as primary or secondary. Primary participants were those most closely engaged with the advocacy efforts of A&T, UNICEF, and partners. For most of them, data collection involved several in-depth interviews (up to a maximum of 7), which generally lasted between 60 and 90 min. Typically, the secondary
Reflective practice Living documents were developed to stimulate reflection with core actors on strategies to obtain validation (theories of change, exploration of concepts). They helped provide feedback and insights with different actors. They also helped in identifying less tangible outcomes (relationships, ideas, conditions for success) that could have contributed to trigger policy changes.
Desk review A large number and diversity of documents were collected and reviewed to track activities, outcomes, contextual factors, and linkages. Research: opinion leader assessments, legal reviews, media audit reports; strategic documents: policy briefs, Code monitoring reports, one pager, joint letters; A&T resources: workplace lactation toolkit, advocacy guide; A&T working documents: donor reports, presentations, advocacy strategy, trip reports, internal briefs, concept notes, timeline; country teams instruments: roadmaps/workplans, meeting minutes; reports: regional workshops organized by A&T and UNICEF in 2013A&T and UNICEF in , 2014A&T and UNICEF in , and 2016 participants were met during country visits; they provided knowledge about the overall context in the country and were involved with advocacy efforts to varying degrees.

| Data analysis
All tape-recorded interviews were transcribed verbatim. Thematic content analysis (Miles & Huberman, 1994) was carried out through an iterative process in which different types of coding (Saldaña, 2012)  A systematic approach was carried out with most of the data collected. First, numerous elements (strategies, events, challenges, contextual factors, and accomplishments) were tracked throughout the evaluation, drawing upon all the available data at a given time.
A chronology of events was thus developed for each country, which was completed with a desk review. The data collected from the nine countries were then organized according to the policy cycle (Clark, 2002). By taking all countries together, the full policy cycle was represented, and the main active stage, which was the latest stage on which actors focused their efforts during the period of this study, was clearly identified for each country. Finally, through this process, activities were identified that could be linked to certain outcomes and that helped to advance the policy cycle. Those activities were deemed "critical" activities, and the rationale for their identification was based on four criteria presented in Box 1. Throughout the evaluation, preliminary findings were validated with country actors, A&T focal points, or representatives.

Box 1: Criteria for the identification of critical activities
The activities meet at least one of the following criteria: 1. The activity has occurred in several countries and has been linked to an advancement within one or more stages of the policy cycle (common positive contributor).
2. The activity has been recognized to help in overcoming a challenge that hindered progress in the stages of the policy cycle (overcoming factor).
3. The activity has occurred in only one country but has clearly been identified as a trigger by several actors (strong contributor).
An additional criterion was used as a counterfactual to reinforce any of the above criteria: 4. An absence of this activity has been observed along with an absence of advancement in the stages of the policy cycle (counterfactual factor).

| RESULTS
Findings from this real-time evaluation based on the country experiences in nine countries are presented here organized in two parts: (a) two main drivers to translate the International Code into national measures, organized according to the stages of the policy cycle and (b) progress achieved on the Code. A terminology for key terms used in this section is presented in an Online Supplementary Material (OSM; Annex 1). In addition to these primary results, a composite case is provided in Annex 2 to illustrate in a more holistic way the experience of actors working to advance the Code, the challenges they face, the strategies they try to put in place, as well as some intermediate and major accomplishments. It also illustrates the non-linear nature of the processes.

| Main drivers and triggers of policy change according to the policy cycle
The different stages of the policy cycle observed included agenda setting, development of the Code, adoption, preparation for implementation, monitoring and enforcement, and finally, evaluation, learning, and adaptation. Although the stages of the policy cycle do not typically take place in a strict linear manner, this real-time evaluation revealed that there was a certain sequence in terms of how progress from one stage to another was achieved. At the onset, the creation of a strategic group represented the first driver of policy change. At each subsequent stage, several activities were identified as critical to trigger advancement within the policy cycle. Later, those activities were termed as critical tasks: none of them was sufficient by itself to move forward the whole policy cycle. Nonetheless, the more those tasks can be carried out, the more likely a country is to progress within the policy cycle. Therefore, this set of critical tasks represents the second key driver of policy change. Those two drivers are represented in Figure 1 with an emphasis on each critical task according to the stage of the policy cycle.

| Agenda setting
The main driver of progress was the creation of a strategic group and the subsequent engagement of key relevant actors.
The creation of the strategic group was a direct result of the inputs from the advocacy efforts engaged by A&T, UNICEF, and partners The ongoing advocacy efforts of the strategic group helped to place the Code on the agenda and to keep it there to achieve progress at any subsequent stage through the realization of the 15 critical tasks described below.

| Development of the Code
The Code can remain in development and/or stalled for several years. replace an existing one that had been developed and approved in FIGURE 1 Critical tasks to translate the International Code into national legal measures the past but was found to have several weaknesses. In such a case, the strategic group worked on the revision of the Code or on various pieces of legislation, but the process appeared similar.

| Approval/adoption
Seeking approval of the Code often was long and arduous depending on the internal processes of the country and nature of the opposition.

Groups of actors
A domestic group was composed primarily of local actors who represented different government entities and who were often directly involved with the law-making process. It played a crucial role in the legal processes. A group of development partners was present in the country. A strategic group of actors was created as the convergence of these two previous groups. Its main function was to create and carry out strategies to advance the Code. It acted as a resource for the government: to connect with international experts, to have good relations with international organizations, to engage the right people at the right time. Existing technical groups such as the ones related to the Scaling-Up Nutrition movement sometimes provided a platform for work around the Code, but this was not always the case. The collection of actors united their efforts throughout the processes to act together and counter various aggressive techniques by the industry.

Individual actors
Champions were instrumental in building consensus with other actors and carrying out certain activities. For example, strong advocates for the Code within the government were not afraid to stand up and defend the Code from tactics used by industry. Dedicated persons specifically in charge of follow-up on actions to advance the Code in a country made it possible to remain alert and act quickly. Internal high-level allies informed members of the strategic group on the internal processes of Code approval and alerted them when there was interference by the industry. Domestic legal experts who were well versed in the legal system facilitated the advancement of various stages, especially early on. They helped the various actors navigate through legal processes to work on the Code effectively or defend it whenever necessary. When legal actors did not have a background in health and/or on the Code, sharing evidence on the importance of breastfeeding and supporting them on problems related to the Code and its application appeared key. International Code experts (from UNICEF and International Code Documentation Center) frequently acted as external resources to support local actors. They were regularly mentioned as having played an important role in the development of Code drafts. In several countries, a consensus building or policy dialogue was organized and such experts were invited to: guide the staff of certain ministries on legal aspects related to the Code, share experiences from other countries, review the regulations related to a draft Code, and guide the local actors in finalizing the Code. The use of international legal experts has often proved decisive during the approval stage to present the facts and respond to criticisms from the industry. Actors from diverse sectors, for example, the Ministry of Culture or Ministry of Information were involved in work with the media, which proved to be effective. Researchers played a key role in generating evidence, providing supportive arguments and sometimes became champions. Civil society organizations helped raising awareness and monitoring violations; they also undertook advocacy activities. The population of some countries was engaged to encourage involvement with monitoring violations; however, such an activity was at an early state.

| Monitoring and enforcement (implementation)
This stage, which includes Code surveillance, is essential to achieve its ultimate objective: to regulate the actions of industry and ensure the compliance of health workers, media, retailers, and others for the well-being of children. However, tremendous challenges were experienced. WHO, UNICEF, and other agencies had jointly developed NetCode to guide countries in creating their own monitoring system. 2 Though few documents/tools related to NetCode were then available, agencies provided direct assistance in setting up this monitoring system, which appeared to enable progress.

Develop tools for monitoring and reporting violations:
Checklists were developed and used for ongoing monitoring.  Table 3 presents the policy objectives set by each country team prior to this evaluation and some insights on progress. Several high-level observations are noteworthy. First, prior to the regional meeting in

| Progress on the Code in nine countries
March 2014, most countries had a Code in place. However, this version of their national Code had insufficient provisions or low ranking in the hierarchies of laws, which explains why many participants at the regional meeting said they wanted to strengthen it. Second, the policy objectives were broad and ambitious, for example, in proposing "revision, monitoring and enforcement" of a national Code. Third, none of the national Codes seemed to be enforced, allowing industry the freedom to continue advertising their products.
Countries made significant progress in translating the Code into national measures and advancing through the policy cycle. Three countries were working on the revision of their Code (Indonesia, Lao PDR, and Burkina Faso) after putting it on the agenda of various organizations (governmental and non-governmental). Thailand had progressed through the internal approval process for its Code, which was finally approved early in 2017 after the end of this evaluation.
Cambodia, Myanmar, and Vietnam progressed with Code monitoring, including creating guidelines, setting up working groups (oversight board, control committee, executive working group), or adopting a deadline for compliance. In Vietnam, a decree was approved regarding the marketing and use of feeding products (feeding bottles, teats, and pacifiers) for young children, to further specify the Advertisement Law. Timor Leste had faced numerous challenges but was gaining momentum in the critical first step of setting the agenda and engaging key relevant actors. The differential progress across countries is due in part to the long timeline typically required for policy change as well as to variation in the opportunities, facilitators, and barriers.
Examining major accomplishments to assess progress is essential, Finally, Vietnam became a model, showing it is indeed possible to make progress using various strategies. The experience of Vietnam, shared at the regional workshop, catalysed movement in all six countries in the region by motivating actors to embark on a similar journey, which is a major accomplishment in itself.

| DISCUSSION
In this paper, we present a broad range of findings on the progress achieved in nine countries in which the A&T initiative took place to improve IYCF policies in collaboration with UNICEF and other partners in supporting governments. We highlight two major contributions and several implications.  relevance as well as their role in protecting breastfeeding. However, much of the literature draws attention to problems and focuses on one stage of the policy process (especially monitoring and enforcement) because of the many reports of industry violations of the Code (Aguayo et al., 2003;Costello & Sachdev, 1998;Parrilla-Rodriguez & Gorrin-Peralta, 2008;Pereira et al., 2016;Perera, 1999;Pries et al., 2016aPries et al., , 2016bSweet et al., 2016;Taylor, 1998  For Southeast Asian countries, the policy objectives were agreed at the 2014 Bangkok meeting.
*Countries in Southeast Asia that have modified slightly their policy objectives after the 2014 Bangkok meeting.
For the African countries, policy objectives came from their Detailed Implementation Plan.
advocacy and policy change efforts can be informed by diverse theories from political science, sociology, and communications (Stachowiak, 2013), the findings presented here add to a recent focus on generating and using "practice-based evidence" (Green, 2008;D. Pelletier et al., 2013).
The findings have implications on how progress can be achieved in other regions. As a result of the advocacy efforts, most of the countries in Southeast Asia in this study have now prioritized work around the Code. This has triggered important momentum in the Southeast Asian region as a whole. It also explains why the great majority of the countries followed in this real-time evaluation have made significant progress in their policy process regarding the Code.
One important driver that appeared to be an achievement by itself was the creation of a strategic group. Actors from the governments of those countries have really taken on the responsibility of improving the Code in their country and collaborate with development partners to achieve their common goal. African countries could benefit from the experience gained in Southeast Asia regarding the Code.
Finally, it is important to remember that the Code was intended to be adopted by all governments as a minimum requirement. Although