We employed a convergent parallel mixed-methods design, conducting quantitative and qualitative components simultaneously within the same research phase. Both methods were given equal weight and analyzed independently.

The study objectives and variables were premised on the WHO's Behavioral and Social Drivers (BeSD) framework. This comprehensive framework consists of four main domains: (i) the thinking and feeling domain, which encompasses individual risk and benefit perceptions; (ii) social processes, including social norms; (iii) motivation; and (iv) practical issues, such as vaccine availability, access factors, quality of service, and the reception by healthcare workers. The BeSD framework is designed to enable health programs to collect, analyze, and utilize data effectively to understand the factors influencing vaccine uptake, thereby informing planning and implementation strategies (16).

Our study incorporated elements from the BeSD framework by leveraging its standardized tools and guidance during the development of data collection instruments, including both questionnaires and qualitative interview guides. This approach facilitates comparisons with other COVID-19 vaccine uptake studies conducted globally under the same framework, promoting consistency and cross-contextual learning.

In this study, the BeSD framework was adapted to assess key variables likely to influence COVID-19 vaccine uptake among persons with disabilities. These variables were categorized into two main groups: socio-demographic factors and behavioral factors, which were considered as independent variables. The dependent variable was defined as the uptake of the COVID-19 vaccination.

This theoretical grounding provided a structured way to analyze how individual, social, and practical factors interact to influence the decision-making process regarding COVID-19 vaccination among PWDs. The framework's application enabled a comprehensive examination that aligns with global studies on vaccine uptake.

The study was conducted in four counties in Kenya, selected based on their disability prevalence and representing the main regions of the country to attain a representative rural-urban mix. The counties were Embu in Eastern Kenya, with the highest disability prevalence of 4.4%, Siaya (4.1%) in the Western side, Mombasa (1.4%) in the Coastal region, and Nairobi (1.1%) in the Central region.

The study participants were PWDs in the hearing, communication, self-care, cognition, mobility, visual and albinism domains, as per the Washington Group on Disability Statistics (WG). The study participants were chosen from the National Council of People with Disabilities (NCPWDs) registration databases of Mombasa, Siaya, Embu and Nairobi Counties. Albinism was included as it is classified as a disability in Kenya because of the associated visual impairment and other vulnerabilities in society, including mistreatment and exclusion (17). All participants were 18 years and above. Critically ill PWDs and those who declined participation in the study were excluded.

The study was conducted in March 2023, about 2 years after rolling out Kenya's national COVID-19 vaccination program. Both quantitative and qualitative data collection took place within a 5-day period from 27^th to 31^st March 2023. The data collection was structured for efficiency and accuracy, involving a team of 29 trained research assistants distributed across four counties based on sample size needs for each county. Each research assistant received in-depth training, and fieldwork was closely supervised by Amref staff who conducted spot checks on collected questionnaires and performed call-backs to verify respondents' participation. PWDs were systematically chosen using a K^th of four to reach a sample size of 792 from the National Council for PWDs Registration database within Mombasa, Siaya, Embu and Nairobi counties. The respondents were contacted prior to field activity, while those not found were replaced by choosing the next participant based on the skipping pattern of four. Qualitative data collection, including 7 FGDs and 4 KIIs, was conducted concurrently by an additional team. This qualitative phase was completed within 3 days and aligned with data saturation principles. Purposive sampling was used to select key informants for in-depth interviews at both the county and national levels for qualitative data. The same technique was used to select 6 to 12 persons with disabilities who participated in the focus group discussions and chosen to represent a diverse cross-section of persons with disabilities (PWDs) across domains such as self-care, mobility, hearing, and cognition.

A structured questionnaire embedded in a mobile collection tool (KOBO) was used to collect quantitative data from PWDs in the selected counties. Internal checks were done within the tool to ensure accuracy and completeness. Moreover, Focus Group Discussions (FGDs) were conducted among PWDs and their caregivers, while Key Informant Interviews (KIIs) were done among disability service officers, social development officers, COVID-19 vaccination in-charges and NCPWD representatives in the target counties. The WHO's framework on Behavioral and Social Drivers (BeSD) of vaccination, displayed in Figure 1 informed the data collection tools. Research assistants were trained on data collection methods, including consenting and interviewing processes, and on the various disability domains and related disability etiquette. Interviews were done in languages understandable to the respondents, with sign-language interpreters engaged to administer interviews to the hearing-impaired.

In this study, qualitative data collection was carried out using semi-structured Focus Group Discussion (FGD) and Key Informant Interview (KII) guides. These guides were developed to capture a comprehensive understanding of the factors influencing COVID-19 vaccine uptake among persons with disabilities (PWDs), aligning with the WHO's Behavioral and Social Drivers (BeSD) framework. The FGD guide comprised approximately 11 key questions, focusing on areas such as perceived COVID-19 risk, vaccine confidence, social norms, and practical issues related to vaccine access. Each main question was supported by probes to facilitate a deeper exploration of participants' experiences and opinions. The KII guide, tailored for officials and community representatives, contained about 11 main questions covering topics like community perspectives, healthcare recommendations, and strategies to enhance vaccine access for PWDs. Probing questions were included to delve into aspects such as vaccine safety perceptions and challenges faced by PWDs during the pandemic. The semi-structured nature of these guides ensured consistency across interviews while allowing flexibility for participants to share detailed insights, enriching the data collected for thematic analysis.

Similarly, the survey questionnaire for this study was developed based on the BeSD framework. This framework provided a structured approach to capture key behavioral and social drivers across four main domains: thinking and feeling about vaccines, social processes influencing vaccination, motivation or hesitancy toward vaccination, and practical issues related to accessing vaccines. The questionnaire was organized into three main sections. Section A captured demographic information, including participant age, gender, marital status, education level, employment status, and disability status. Section B collected socio-economic data, with questions focusing on household composition, living arrangements, caregiving responsibilities, and the use of assistive devices. Section C addressed COVID-19 vaccine uptake, including questions on perceived COVID-19 risk, vaccination history, motivation for vaccination, sources of information, and practical challenges related to vaccine access. Each section was designed with detailed items allowing for both quantitative and qualitative responses, enabling a comprehensive assessment of beliefs, social influences, and access-related issues. The questionnaire was administered by trained interviewers who followed standardized instructions, including directives for optional questions and emphasis on priority items. This structured approach ensured consistent data collection, capturing distinct insights into vaccine-related attitudes and behaviors across diverse respondent demographics.

All data collected through KOBO were first transferred to Excel for cleaning and later to STATA (version 14) for analysis. Descriptive statistics, including mean, standard deviation and frequencies, were generated. Tables and bar graphs were used to present the frequencies. Additionally, χ² tests were used to test associations between behavioral and social predictors with the uptake of the COVID-19 vaccine. Cross-tabulation was used to describe data and to explain the relationship between COVID-19 uptake and the above independent variables. Multivariate regression was used to investigate how COVID-19 uptake was related to demographic, behavioral and social predictors, among other independent variables. The statistical significance level was set at p < 0.05 (18). For qualitative data, all interviews were transcribed and coded for analysis using NVivo software. Themes were generated using WHO's framework on the Behavioral and Social Drivers (BeSD) of vaccination, and deductive content analysis was used for analysis.

We derived a total of 23 unique codes from the qualitative data. Each code represented a specific concept or observation captured from participant responses, such as “Risk to self,” “Confidence in vaccine,” and “Family support/permission to vaccinate.” These codes were organized into broader categories to help identify patterns and connections within the data, with categories such as “Risk Perception,” “Social Influences,” and “Access and Practical Issues.” Finally, we synthesized these categories into overarching themes that provided insights into key findings. Qualitative findings revealed unique motivations for PWDs, for instance, getting vaccinated to safeguard oneself against additional exposure from assistive devices, e.g., white canes and reading braille. The influence of the political class on vaccine uptake was also cited in several accounts by PWDs on getting the vaccine upon seeing their political leaders, notably the president, getting it.

Perceived vaccine effects were also conspicuously reported among respondents in the qualitative findings, with a majority reporting having heard of the effects from external sources and hardly any from personal experience. Other notable reasons attributed to low uptake were distance and associated cost implications. One of the respondents recounted, “The vaccination place was far from our residential area. Most of the people didn't get vaccinated because they didn't have transport money to move from the residential area to where the vaccination was taking place.”—Visually Impaired FGD Respondent from Siakago, Embu County.

The qualitative findings similarly revealed that health workers were considered as credible sources of information, as recounted for instance: “If I hear this information from a medical person, I would believe [it] because they have knowledge in the field. I would trust them [healthcare providers], but if I get it [information] from people around here I would not trust the vaccine since people have so many things going on. If it [information] comes from a religious leader, I will accept but not believe since that is not their area of specialty.”—–A Physically Impaired FGD Respondent from Mbeere North, Embu County. Health workers were however required to provide clear information to influence vaccine uptake, as recounted “If safety is guaranteed and the healthcare providers give clear [information] and sensitize, there will be an increase the uptake.”—The Regional In-charge of the National Council for PWDs for Meru, Tharaka Nithi, Isiolo, and Marsabit counties. Key recommendations made were on working with PWD organizations and their social networks to reach people in all disability domains, intensifying sensitization of PWDs for informed decisions, giving special consideration to PWDs, equipping health workers to handle PWDs, and making transportation arrangements to vaccination sites for PWDs.

This study highlights several key areas for improving vaccine uptake among persons with disabilities (PWDs). Tailored communication strategies, including accessible formats and clear messaging, are essential to address vaccine hesitancy. Engaging healthcare providers, caregivers, and family members is also crucial, as they serve as trusted sources of information. Policies should prioritize decentralized vaccination services, transport support, and ensuring PWDs are prioritized in vaccine distribution. Community-based initiatives, such as mobile clinics, and fixed-time appointments can further enhance access.

Clinically, healthcare providers should be trained in disability-inclusive care, focusing on clear communication and addressing vaccine concerns. Involving family members and caregivers in the vaccination decision-making process is also vital. Workplace vaccination policies, such as paid leave and transport assistance, could increase uptake among employed PWDs, and extending these policies to rural and smaller employers may further improve access.

From a policy perspective, disability-inclusive health strategies should be integrated into national vaccination campaigns. Collaborating with disability advocacy groups to design accessible and supportive interventions will help ensure that PWDs are included in future vaccination efforts.

While this study provides valuable insights into vaccine uptake among PWDs, a few limitations should be considered. First, the study's cross-sectional design limits the ability to infer causal relationships between the identified factors and vaccine uptake. Additionally, the sample was limited to specific regions and may not be fully representative of the broader PWD population in Kenya, particularly those in rural or hard-to-reach areas. Although efforts were made to ensure diverse participation, there may be regional variations in attitudes toward vaccination that were not fully captured in this study. The reliance on self-reported data also introduces the possibility of response bias, which is common in survey-based research, as participants may have over-reported vaccination status or may have been influenced by social desirability bias. Finally, while we focused on understanding the perspectives of PWDs, further research that includes the views of healthcare providers and policymakers would provide a more comprehensive understanding of the barriers and enablers to vaccine uptake from a broader, systemic viewpoint.