*2.2. Methods for the Expanded Mapping*

The expanded mapping sought to include data from 48 additional Gavi-supported countries, as well as incorporate more recent data from phase I countries, when available. We also aimed to interpret data by applying additional equity filters, considering genderrelated barriers to immunization and inequities in immunization for target populations (urban poor, remote rural, conflict). Lastly, we sought to organize strategies thematically and identify opportunities to make findings accessible to countries, development partners and other interested parties

## 2.2.1. Data Source Selection

The mapping considered Gavi Annual Country Joint Appraisals (JA) reports between 2016 and 2019, aligning with the Gavi 5.0 strategy timeline. JA reports describe country-level implementation progress and the performance of Gavi funding support [14]. Countries were selected based on eligibility for Gavi support in 2020. In view of the COVID-19 pandemic in 2020, countries did not submit JA reports. Instead, countries conducted "Multi-Stakeholder Dialogues" (MSD), which aim to convene relevant stakeholders to discuss country-level barriers to immunization and align future objectives and actions. MSD reports summarizing key findings from these sessions were included as data points for that year. JAs and MSDs were obtained from the Gavi website and the Gavi Secretariat.

A total of 126 JAs (*n* = 106) and MSDs (*n* = 20) from 48 additional countries were selected for the data extraction. Up to three most recent data points were chosen for each country, when available. Additionally, given that Syria only became eligible for Gavi support in 2019 and there were no JAs or MSDs available, a report on Syria's National Immunization Strategy was obtained as an alternative data source.

#### 2.2.2. Data Extraction and Synthesis

Strategies were identified via a targeted keyword search, informed by the phase I conceptual framework. The data extraction method was initially calibrated by manually extracting strategies from randomly selected JAs used in phase I. After careful review of the initial mapping and JA reports, additional keywords were generated for the thematic areas to improve the identification of strategies and increase the scope of the extraction.

Data extraction was facilitated by the "text retrieval" feature of the freely available qualitative analysis software QDA Miner Lite to accommodate a large dataset and increase the reliability of results. For efficiency and convenience, keywords were searched across all JA and MSD reports simultaneously. Paragraphs where keywords appeared were analyzed for their relevance and presence of strategies corresponding to the thematic areas. An Excel extraction matrix, including countries, thematic areas and keywords, was populated with strategies.

Data extraction was performed by the first author, with supplemental cross-check for 10% of the data by a second reviewer. Independently obtained results between both reviewers were compared at multiple occasions. Differences in data extraction and strategy capture were discussed among reviewers and co-authors until a consensus was reached and appropriate adjustments or calibrations were made. Additionally, all keywords were translated in French to enable data extraction from reports available only in French. The final set of keywords used for extraction can be found in Appendix B.

Data from phase I and II were collated, yielding a total of 607 unique strategies, extracted from 174 reports and 61 countries. The thematic synthesis of pro-equity strategies was informed by the analysis led by Dadari and colleagues in Phase I, UNICEF's Journey to Health and Immunization (JTHI) framework, and the Global Routine Immunization Strategies and Practices (GRISP) framework [13,15,16]. The JTHI framework is a tool used to identify the factors influencing different points of the immunization service delivery, including before, during, and after immunization. The framework also offers an opportunity for a more targeted identification of barriers and solutions. Table 1 outlines key dimensions related to each step of the JTHI framework.


**Table 1.** Key caregiver and health worker dimensions of the JTHI framework [15].

The thematic synthesis of strategies was also informed by the Global Routine Immunization Strategies and Practices (GRISP) framework [16]. The GRISP comprehensive

framework of strategies and practices for routine immunization introduces key areas of action to strengthen immunization systems and improve coverage. The framework also describes a systemic approach to address barriers by tackling four categories of actions: maximizing reach, managing the program, mobilizing people and monitoring progress [16]. Among these actions, GRISP highlights nine transformative investments, aimed to guide governments to transform immunization programs and achieve better outcomes [16].

We thematically synthesized the collated pro-equity strategies, building on the themes identified in phase I, JTHI steps and GRISP dimensions. When relevant, GRISP approaches were included in the themes. Examples of GRISP themes include the integration of immunization with other routine services, strategies to address vaccine hesitancy and misinformation and practices to build the capacity of healthcare workers. Given that some strategies were found to be relevant to more than one theme and JTHI step, we opted to include them in all relevant dimensions and themes, leading to a total of *n* = 740 data points used for the analysis. To support knowledge use, a learning tool and searchable database of strategies was created by UNICEF. The tool allows filtering of results by country, JTHI steps, health system element, relevance to key populations and the application of a gender lens [17].
