Using immunisation caregiver journey interviews to understand and optimise vaccination uptake: lessons from Sierra Leone

Quantitative and qualitative assessments have revealed diverse factors that influence the uptake of childhood immunisation services and shed light on reasons for vaccination delays and refusals. UNICEF and partner organisations developed the Immunisation Caregiver Journey Framework as a novel way to understand caregiver experiences in accessing and receiving immunisation services for children. This framework aims to help immunisation programmes identify vaccination barriers and opportunities to improve vaccination uptake by enhancing the overall caregiver journey in a systems-focused manner, using human-centred design principles. In this paper, we adapt the framework into a flexible qualitative inquiry approach with theoretical guidance from interpretative phenomenology. We draw from the implementation experiences in Sierra Leone to inform methodological guidance on how to design and implement the Immunisation Caregiver Journey Interviews (ICJI) to understand the lived experiences of caregivers as they navigate immunisation services for their children. Practical guidance is provided on sampling techniques, conducting interviews, data management, data analysis and the use of data to inform programmatic actions. When properly implemented, the ICJI approach generates a rich qualitative understanding of how caregivers navigate household and community dynamics, as well as primary healthcare delivery systems. We argue that understanding and improving the caregiver journey will enhance essential immunisation outcomes, such as the completion of the recommended vaccination schedule, timeliness of vaccination visits and reduction in dropouts between vaccine doses.


INTRODUCTION
Vaccines play a critical role in protecting children from life-threatening and debilitating diseases. 1 In the last decade alone, more than 1 billion children were vaccinated worldwide through the Expanded Programme on Immunisation. 2 These efforts have led to the elimination of measles in some geographical regions and near eradication of polio globally. 3 4 The WHO estimated that in 2019, approximately 85% of all infants globally received three doses of diphtheriatetanus-pertussis-containing vaccine (DPT3), a standard measure for the performance of an immunisation programme. 5 Despite high coverage globally, inequities between low-income and middle-income countries Summary box ► The global burden of childhood diseases has decreased substantially since 1990 largely due to improved access and increased uptake of vaccines offered through the Expanded Programme on Immunisation. ► Despite the availability of vaccination services, caregivers may delay or refuse some or all scheduled vaccines for their children for a variety of reasons, including contextual, individual or vaccine-specific and vaccination-specific influences. ► We developed the Immunisation Caregiver Journey Interviews (ICJI) approach based on the associated Caregiver Journey Framework using interpretative phenomenology as the basis to deeply understand caregiver experiences before, during and after accessing immunisation services for their children. ► The caregiver experience at each stage of the immunisation journey may vary and be influenced by disparate or interrelated factors. Immunisation programmes and independent evaluators alike may find the ICJI approach useful to understand the complexities of vaccination uptake within or across caregiver groups at national or subnational levels. ► A rich, qualitative understanding of the caregiver journey can inform context-specific and culturally responsive interventions to improve vaccination uptake.

BMJ Global Health
(LMICs) and high-income countries persist, though their differences-as measured by DPT3 coverage-declined from approximately 30% in the year 2000 to 15% in 2010. 1 Inequities in vaccination coverage also exist across and within LMICs, 6 contributing to undervaccination among those living in conflict-affected places, hard-toreach areas and urban poor settings [7][8][9][10][11] and populations with low socioeconomic status. 12 The proportion of children globally who received DPT3 has stalled at around 85%, despite efforts over the past decade to increase its coverage. 5 As countries grappled with stagnated vaccination coverage, the COVID-19 pandemic further complicated the situation by inadvertently disrupting the delivery of childhood immunisation services in many countries due to restrictions in population movement and other risk mitigation measures. 13 However, even before the COVID-19 pandemic, approximately 20 million children missed one or more scheduled vaccine doses annually. 14 Current evidence suggests that even when vaccination services are available, caregivers may delay or refuse some or all vaccines for their children for a variety of reasons, including contextual, individual or vaccine-specific and vaccination-specific influences. 15 This phenomenon has been termed 'vaccine hesitancy' by the WHO and has been listed among the top 10 threats to health globally. 16 Beyond reasons related to vaccine hesitancy, hard-toreach populations may also be undervaccinated, which contributes to inequities in vaccination coverage. Hardto-reach populations include groups that face 'barriers to vaccination due to geography by distance or terrain, transient or nomadic movement, healthcare provider discrimination, lack of healthcare provider recommendations, inadequate vaccination systems, war and conflict, home births or other home-bound mobility limitations or legal restrictions'. 17 TOWARDS AN EXPERIENCE-DRIVEN APPROACH Expectancy-value theories such as the Health Belief Model have largely informed past assessments of vaccination attitudes, intentions and behaviours. [18][19][20][21][22][23][24][25][26] These expectancy-value theories do not adequately account for the dynamic role of caregivers' repeated experiences in shaping their vaccination behaviours. Recognising these limitations, various international partner organisations, under the leadership of UNICEF, developed the Immunisation Caregiver Journey Framework 27 as a novel way to contextualise and understand how caregivers and families navigate childhood immunisation services, especially in LMIC settings. The framework aims to help identify vaccination barriers and leverage opportunities to improve vaccination uptake in a systems-focused manner using principles of human-centred design. 27 In operationalising the framework in the context of qualitative inquiry, phenomenology provides a suitable methodological basis 28 29 for examining the lived experiences of caregivers as they access childhood immunisation services in diverse settings. Focusing on specific events and the related experiences allows caregivers to provide rich information that thoroughly describes the immunisation journey in the context of both health systems and social conditions. 30 The caregiver experience may vary and be influenced by disparate or interrelated factors at each stage of the immunisation journey. Holistic exploration of the journey helps to provide a nuanced understanding of how caregivers navigate household dynamics; community norms, processes and structures; and healthcare delivery systems. Therefore, rigorous qualitative inquiry of the caregiver journey is needed to explore the range of diverse experiences and events occurring before, during and after accessing immunisation services. Here, we describe the development of a qualitative approach for conducting Immunisation Caregiver Journey Interviews (ICJI), using interpretative phenomenology as the basis to deeply understand caregiver experiences and the potential influences on vaccination outcomes.

INTENDED USE AND AUDIENCE
The ICJI approach provides a practical methodology to qualitatively explore and critically examine caregiver experiences when accessing childhood immunisation services, with a contextual focus on LMIC settings. Adaptions of the ICJI may also be appropriate in high-income country settings. The approach is intended to be used to conduct face-to-face in-depth interviews [31][32][33] with caregivers of children who are eligible for immunisation services (or catch-up doses), and the child has received one or more vaccines through a fixed-post strategy at a vaccination site (eg, health facility). It may also be adapted for use in focus group discussions. Given that the approach explores caregiver experiences in accessing and receiving immunisation services, its current format is not readily appropriate for use with caregivers who have never accessed immunisation services for their children or whose children only received vaccines via door-to-door campaigns (eg, polio vaccination campaigns).
Depending on the context and needs, the ICJI approach can be used as a stand-alone assessment or incorporated into other planned data collection efforts. In table 1, we have summarised the strengths, challenges and considerations when integrating the ICJI approach into existing data collection opportunities within immunisation programmes such as the Review of the Expanded Programme on Immunisation (EPI Reviews), 34 Tailoring Immunisation Programmes, 35 New Vaccine Post-Introduction Evaluation 36 and Reaching Every District. 37 Standalone ICJI assessments may not always be feasible for various reasons-for instance, when resources are not available for a standalone assessment. Even when resources are available, integration may be more advantageous where the emphasis is on ensuring that the ICJI data can be triangulated and used alongside other immunisation systems assessment data such as during EPI Reviews. Moreover, the ICJI approach may be adapted BMJ Global Health Table 1 Summary of opportunities for integrating the Immunisation Caregiver Journey Interviews (ICJI) into existing data collection efforts in immunisation programmes

Why integrate?
What are the key strengths?
What are the potential challenges?
What are some key considerations for success?
Review of the Expanded Programme on Immunisation (EPI Reviews): comprehensive assessments of national and subnational immunisation programmes 34 ICJI may inform issues and barriers related to service delivery and vaccination uptake from the caregiver perspective ► Ability to get information systematically every 3-5 years across multiple countries ► Strong participation from diverse stakeholders across the immunisation programme

BMJ Global Health
for use in efforts aiming to understand the impact of the COVID-19 pandemic (or other future health emergencies) on childhood immunisation.

DEVELOPMENT AND REFINEMENT OF THE INTERVIEW GUIDE
We drew from the interpretative phenomenological tradition in qualitative research to inform our development of the interview guide for the ICJI. We initially drafted broad, open-ended questions under three temporal areas related to caregiver experiences when accessing childhood immunisation services-before, during and after accessing immunisation services at a health facility. We developed open-ended questions and probes to understand (1) decision-making and preparation, (2) making the journey to the vaccination site, (3) experiences at the vaccination site, (4) postvaccination experiences and (5) intentions to return for scheduled vaccination visits.
In designing the Sierra Leone ICJI assessment, we held a stakeholder engagement meeting where we received feedback from diverse representatives from the Sierra Leone Ministry of Health, civil society partners and nongovernmental organisations supporting immunisation efforts in Sierra Leone. The feedback we received from the stakeholder engagement informed our initial development of the ICJI interview guide, which was subsequently piloted in Freetown, Sierra Leone, with a convenience sample of four caregivers that were recruited from the immunisation clinic in a paediatric hospital. We used feedback from caregivers to refine and expand the interview guide-for instance, we added questions and probes to explore how immunisation services are promoted in communities. Also, in the revised guide, additional probes were included to get more nuanced information regarding the relationship and interactions between caregivers and health workers. We piloted the revised guide once again with a separate convenience sample of four caregivers who were recruited from urban communities in Freetown, Sierra Leone. This process led to the finalisation of the questions and probes in the ICJI guide (table 2). During implementation, the interviewers had opportunities to do 'on-the-spot' probing for additional information based on the specific experiences of the individual caregivers. In addition to the interview guide, we have provided sample consent statement and interview cover sheet (online supplemental material).

SAMPLING TECHNIQUES AND DATA COLLECTION
When implementing the ICJI assessment in a specific area, we recommend purposive sampling of at least 12 eligible caregivers and continuing until data saturation is reached based on the concept of qualitative information power assessed against the study aim, sample specificity, established theory, quality of dialogue and analysis strategy. 38 For instance, instead of interviewing 12 caregivers, a larger sample of caregivers may be necessary to achieve information power if the assessment is adapted to compare vaccination experiences between multiple population subgroups. According to the concept of information power, sampling burden is generally reduced when the aim is narrow (vs broad), the sample is targeting a specific group (vs multiple groups with varying attributes of interest), established theory is applied to inform the assessment (vs no application of theory), the quality of dialogue is strong (vs weak) and case-based analysis is performed (vs cross-case analysis). These attributes combined should be critically appraised before the initiation of the assessment, periodically during data collection and again after data collection to ascertain information power.
To be eligible, (1) caregivers should have at least one child who is eligible for immunisation services or eligible to receive catch-up doses previously missed, and (2) the child has received one or more vaccines through a fixed-post strategy at a vaccination site. Concerted efforts should be made to obtain a sample of caregivers with diverse backgrounds so that contextualised differences and similarities in caregiver experiences can be captured. To help maximise variation in the sample, quota sampling may be leveraged to ensure having a minimum number of certain groups of caregivers or attributes of interest. We have provided an example of a target sample of caregivers, which may be adapted for different contexts (table 3).
Caregivers may be recruited using snowball sampling whereby four eligible caregivers are identified first, one from each of the four categories of eligible caregivers outlined in table 3; then, ask each of the initial four caregivers to refer you to other similar, eligible caregivers in the community. Alternatively, a review of the health facility immunisation registers may be used to identify children who meet the criteria outlined in table 3. If caregiver selection is based on register review, it is important to avoid the involvement of health workers in the selection because they may bias towards caregivers who they know are likely to provide certain desirable responses. Use of the ICJI approach is also appropriate to embed within a qualitative longitudinal design that recruits one or more cohorts of eligible caregivers to conduct repeat interviews at different time points. 39 One or more data collection teams may be needed depending on the sample design, desired turnaround time to collect the data and other logistical considerations. We recommend that each team be composed of two members-an interviewer and a dedicated notetaker, which is customary in qualitative investigations. The interviewer should primarily be responsible for asking the questions in the guide while the note-taker takes copious notes of the conversation. With the caregiver's permission, the interview may be audio-recorded to allow for a more accurate capturing of the conversation. The interviewer and note-taker should conduct a 20-30 min debriefing session at the end of each interview. 40 In Sierra Leone, our initial strategy was to recruit caregivers through health facilities on immunisation clinic days. In piloting this strategy, we learnt that it was difficult BMJ Global Health

BMJ Global Health
to locate previously identified caregivers in their communities. This was particularly evident in slum settlements where distinctive address systems are non-existent. We subsequently shifted to a community-based recruitment strategy with on-the-ground support from community health workers (CHWs) who helped us to identify eligible caregivers. After identifying some initial caregivers with the help of CHWs, we used a snowball sampling technique wherein caregivers assisted the data collection team in identifying other eligible caregivers in their communities. An important lesson learnt is that CHWs play a very important role in identifying and locating caregivers and should be leverage in future ICJI assessments other similar settings. In most instances, CHWs had to accompany data collectors to the homes of caregivers, which helped the teams to establish rapport with the caregivers. Moreover, we found it useful to have the flexibility to schedule interviews outside of normal working hours to avoid conflicts with caregivers' livelihood activities. Some caregivers needed the endorsement of their husband/ partner or male head of household to participate in the assessment. At times, a male decision-maker was   33 In implementing the ICJI in Sierra Leone, we conducted both within-case and cross-case analyses. Consistent with interpretative phenomenological analysis (IPA), 41 we dedicated considerable amount of time analysing the transcript from each caregiver and developing individual narrative profiles with thick descriptions, verbatim quotes and iterative interpretations of the documented experiences. We inductively coded all transcripts with the aid of NVivo V.12 software. Additionally, we used a qualitative content analysis approach to facilitate the grouping of the inductive codes into categories using Microsoft Excel. 42 Finally, in developing crosscutting themes, we iteratively interpreted findings from the within-case and cross-case analyses to deeply understand convergence and divergence in the immunisation caregiver journeys. We learnt from this process that although there are theoretic tensions between IPA and qualitative content analysis, the two can be successfully combined wherein content analysis is used in a non-theoretical fashion (ie, only for inductive coding and categorisation) and interpretative phenomenology provides the theoretical basis for exploring, analysing and interpreting the caregiver experiences using within-case and cross-case analyses. For instance, we only identified one caregiver who routinely refused all vaccines for her child. The within-case analysis of this caregiver was used to provide a rich account of why some caregivers may actively refuse all vaccines for their children. In the final thematic write-up, we captured and interpreted both the thick descriptions of individual experiences and their shared experiences to give meaning to the immunisation caregiver journey in low-resource urban communities of Sierra Leone.

REFLEXIVITY AND MAKING SENSE OF THE DATA
Reflexivity is an important attribute of qualitative inquiry. 43 In analysing the ICJI data from Sierra Leone, the two primary analysts (one woman and one man) brought their own lifeworld experiences and perspectives to the inductive analysis and interpretation. Both analysts and their contributors had experience working in global immunisation with diverse experiences conducting social and behavioural assessments to understand childhood vaccination experiences and behaviours in numerous countries in West Africa. On the one hand, having had prior experience in childhood immunisation work within West Africa, including Sierra Leone, allowed the team to draw from these experiences in interpreting the data. On the other hand, such experience had the potential to create blind spots. During the early stage of the inductive coding and analysis of transcripts, the analysts shared three of the initial coded transcripts with a third analyst that did not have any experience working in Sierra Leone or sub-Saharan Africa to 'blindly' code the transcripts so that we get an 'outsider' perspective. For instance, the 'outside' analyst more pronouncedly identified the complex role of gender dynamics on the caregiver journey, which the two analyst also identified but may have been more inclined to take for granted given their preconceived sociocultural knowledge of gender roles and cultural norms in Sierra Leone.
In analysing the pilot data from Sierra Leone, we identified numerous areas in the transcripts that needed additional follow-up probing. For instance, we observed that time was considered as a commodity by most of the caregivers, and many of them consistently complained about the 'waste of time' due to prolonged delays at vaccination sites. Further probing would have provided more nuanced understanding of this notion of time as a commodity. In addition, although we had some anecdotal awareness of the obstacles faced by caregivers when accessing childhood immunisation services, we were surprised by the extent of the compounded practical constraints and logistical challenges that caregivers had to overcome in taking their children for vaccination services. Another surprising observation was the complexity surrounding monetary exchange between caregivers and health workers. Based on our prior experiences in Sierra Leone and LMICs elsewhere and knowledge of the literature, we were aware that some health workers may request informal payments from caregivers in exchange for vaccinating their children, 44 though heavily discouraged by official policies aiming to provide free health services at no charge. 45 After analysing the data, we learnt that BMJ Global Health although many caregivers begrudgingly gave money to health workers, there were instances when they also willingly did so to show their appreciation. Based on what we have learnt from Sierra Leone, future ICJI assessments need to dig deeper into these complexities around the relationship between caregivers and health workers.

INFORMING PROGRAMMATIC ACTIONS
Following the ICJI assessment in Sierra Leone, there were several programmatic actions taken after we shared the results with the Ministry of Health and stakeholders in the immunisation programme including implementing partners and donors. Insights from the assessment informed the design of joint in-service training that covered various topics including interpersonal communication and community outreach for over 400 health staff involved in childhood immunisation in urban catchment areas in Western Area, Sierra Leone. In addition, the findings were used to advocate for and support a more active role of CHWs in the tracking of children who have missed scheduled immunisation in urban communities, including in urban slums.

CONCLUSION
The ICJI approach and the emerging insights from the Sierra Leone assessment provide methodological and programmatic considerations for ongoing and future efforts aiming to understand and improve childhood immunisation services and outcomes. 46 47 Immunisation programmes may be able to improve vaccination outcomes by gaining a deep understanding of the caregiver journey and using those insights to address complex barriers while harnessing opportunities at the individual, community and health systems levels.