Methodology

The Senegal Vaccine Delivery study is a collaboration between Emory University, the Institut de Recherche en Santé, de Surveillance Épidémiologique et de Formation (IRESSEF), the Georgia Institute of Technology, the Bill & Melinda Gates Foundation, the Institute for Health Metrics and Evaluation (IHME), and Gates Ventures, as part of the Exemplars in Global Health Program.¹

Exemplar country selection

Filtering criteria were used to identify low-income countries that increased vaccine coverage from 2006 to 2016 (see figure below). From these candidates, countries with diverse locations, cultures, and health system attributes were selected for deeper examination.

Project frameworks

Several frameworks informed this qualitative research.

Conceptual framework for vaccine delivery

An existing framework describing the determinants of success in immunization systems (see figure below)² was adapted to incorporate contextual factors such as governance, financing, and health care system attributes. This framework informed the analysis of delivery system components.

Frameworks for systems change

Three frameworks informed our research concerning systems change.

To explain the implementation of specific interventions, we established a simple “Why-How-What” framework. In this framework, we defined Why as the process of problem identification, How as the mechanisms for change, and What as the solutions that were implemented. This framework was applied to the adaptation of external guidelines and to the development and delivery of internal innovations.

The Consolidated Framework for Implementation Research (CFIR) serves as a guide to understand the contextual factors that influenced the success or failure of a specific intervention. These include the outer context, the inner (organizational) context, the characteristics of the intervention, the implementation approach, and the individual actors responsible for implementation.³

The Context and Implementation of Complex Interventions (CICI) framework aims to advance the understanding of how and why interventions work. ⁴ The CICI framework informed analysis of context and setting for system change.

Desk review

The study team undertook an extensive review of available information, and published data on the state of immunization in Senegal from 2000 to the present. This included coverage data from surveys, the World Health Organization (WHO)/UNICEF Estimates of National Immunization Coverage, and IHME as well as historic and current health policies, economic conditions, the political context, and other factors that may have influenced the success of the immunization program. The desk review informed the interview guides used in the in-depth research visit, data analysis, and synthesis stages.

Primary research

Primary data collection was conducted in two stages. An initial scoping was conducted in November and December 2018. In this stage, key informant interviews (KIIs) were conducted to understand the evolution of immunization programming and the factors thought to contribute to Senegal’s performance in immunization. Altogether, five KIIs were conducted with a ministry official, officials in partner organizations, and a member of the national immunization technical advisory group.

The scoping and desk review informed the second stage of primary research, which consisted of in-depth KIIs and focus group discussions (FGDs). These activities were delayed by the COVID-19 pandemic and were ultimately conducted from December 2020 to April 2021. At this time, COVID-19 infection rates in Senegal were still relatively low but rising.⁵

The in-depth research was conducted at national, regional, health district, health facility, and community levels. The selection of study regions and districts aimed for diversity in geography and coverage, and considered recommendations from MSAS and other stakeholders. Study districts included the national and provincial capitals.

Interview guides were informed by the CFIR domains and adjusted to local context. The interview guides were translated into French and Wolof by local research assistants. All interview guides were piloted before use and adjusted iteratively throughout data collection.

KIIs were conducted at the national, regional, district, and health facility levels to (1) identify the strategies behind current immunization programs, as to inform the current context; (2) identify previous immunization activities occurring during changes in DTP1 and DTP3 coverage from 2000 through 2018; and (3) identify key internal and external actors, stakeholders, and partnerships in the immunization program both past and present. They also explored the effects of the COVID-19 pandemic on Senegal’s immunization system.

At the national level, an initial list of KIIs was developed with the help of MSAS officials. Snowball sampling was then used to identify other potential informants. KIIs were conducted with representatives of MSAS and the Ministry of Education, UN agencies, and other key stakeholders in the immunization sector. Subnational interviews were conducted with health officers at the regional and district levels as well as with heads of health facilities and community leaders. Because data were collected during the COVID-19 pandemic, several potential interviewees were not available, since they were dealing with the pandemic response.

Health facilities were selected in collaboration with local governments, and then KIIs were conducted with the heads of health facilities and community leaders at those locations.

FGDs were conducted with CHWs and caregivers. The goals of the FGDs were 1) to understand the service provided by the health system, including past and present vaccine interventions; 2) to identify the roles and responsibilities of CHWs and volunteers in the health and immunization sector; and 3) to understand historical and current factors behind vaccination coverage in their respective communities.

FGDs were organized in selected health facilities with the support of district health officers. Participants were selected prior to the date of the FGD. CHWs and volunteers had to be currently active and work within the ward; preference was given to CHWs and volunteers who had worked in the community for over five years. Mothers had to be below the age of 45 and have a child who was between 1 and 17 years old. Preference was given to members of women’s and mothers’ groups. Interviews were conducted by local research assistants in Wolof. FGDs had a maximum of eight participants and were conducted in Wolof in private health facility rooms.

Data analysis

We conducted a thematic analysis of KII and FGD transcripts to identify the key drivers of success in immunization programs relative to deductive and inductive themes, applying constructs from the Why- How-What, CFIR, and CICI frameworks.

We considered context and participant roles while uncovering key points, and triangulated data with historical documents and literature review. Emerging themes were identified and grouped with predetermined research questions and objectives, while also accounting for unexpected themes.

Ethical approval

The research protocol was reviewed and approved by the National Ethical Committee for Health Research (CERNS, Comité National d’Ethique pour la Recherche en Santé) in Dakar (00000174). This study was also approved as exempt human research by the Institutional Review Board committee of Emory University, Atlanta, Georgia, US (IRB00111474). This project qualified for exemption from the requirements of federal regulations under 45 CFR 46.104(d)(2). Participation in the KIIs and FGDs was voluntary. Consent forms were read verbatim to participants in their preferred language by a trained interviewer. Participants provided written informed consent in the presence of a witness.

Quantitative decomposition

A quantitative analysis was conducted to assess the relative contributions of each of the immediate drivers in the EGH Vaccine Delivery Framework. This approach linked existing data over time to construct a child-level cohort with estimates of:

• Intent to vaccinate: Caregiver perceptions from the Vaccine Confidence Project Sahel Survey 2016
• Community access: Estimated motor, walking travel time to the closest immunization services using facility locations identified from sources below and the 2019 travel friction surface from the Malaria Atlas Project
• Health facility surveys: DHS SPA (2012-2019)
• Facility lists: COUS, Maina et al 2019, ESRI Kaolack, Health Data Exchange (HDX), Agence Nationale de Statistique et de la Demographie (ANSD)

• Facility readiness: Summary measures of facility readiness based on WHO Service Availability and Readiness (SARA) guidelines for routine immunization, as derived from Service Provision Assessment surveys 2012–2017 and 2019⁶

Tree-based machine learning methods were used to model interacting-nonlinear relationship between drivers, and Shapley additive ExPlanations were used to decompose modeled contributions to likelihood of coverage. In Stage 1, the impact of the three drivers on coverage was estimated. In Stage 2, data on demographic characteristics such as sex, parental education attainment, household wealth, birth cohort, and region were incorporated.