This study was guided by the Health Literacy Skills (HLS) framework (Figure 1),¹⁷ which examines how individuals access, process and apply health information to make informed decisions. The framework highlights four main components: (1) factors influencing the development and use of HLS (e.g. demographics, prior knowledge, resources); (2) health-related stimuli (e.g. sources and channels of health information); (3) the HLS required to comprehend and apply information; and (4) mediators linking health literacy to health outcomes.

In this study, the first three components were adapted to examine how rural Kenyan women access CC information and how these factors relate to screening uptake. Health literacy was conceptualised as a potential mediator between information exposure and screening behaviour. The HLS framework also informed our use of a mixed methods design to achieve a comprehensive understanding of both the measurable associations and the subjective experiences shaping women’s health behaviours.

This was a cross-sectional study using a convergent mixed methods design to comprehensively assess CC information access and screening uptake among rural Kenyan women. Quantitative data were collected from 174 participants through interviewer-administered surveys. Qualitative data were obtained from a purposive sub-sample (n = 21) of survey participants through semi-structured interviews. Quantitative and qualitative findings were integrated during interpretation to enhance contextual understanding.

The study was conducted in Bomet and Kericho Counties. Both counties are predominantly rural, with 96.8% of Bomet’s population and 89.6% of Kericho’s population residing in rural areas.¹⁸ These areas are characterised by smallholder farming, with most residents engaged in subsistence agriculture, tea farming and livestock rearing as primary sources of livelihood. Health facilities in both counties consist mainly of dispensaries and health clinics, with referral hospitals.¹⁹ However, access to specialised health services, including cancer screening and treatment, remains limited, requiring residents to travel to larger urban centres such as Nairobi.²⁰ The population is predominantly from the Kipsigis subgroup of the Kalenjin ethnic community, with culturally embedded practices and beliefs influencing health-seeking behaviours. Kipsigis is the primary language spoken. Similar to other women in rural Kenya, women in these counties face challenges such as illiteracy or low literacy rates, limited access to health information, and inadequate healthcare infrastructure.^21,22 The setting’s rural nature, coupled with socio-economic and health system constraints, provides a critical context for understanding access to CC information, health literacy and screening behaviours among women in these counties.

Eligible participants were women who: had completed Grade 8 or lower, self-identified as female, were eligible for CC screening (aged 18–65 years with no history of total hysterectomy), and were proficient in the Kipsigis language. Participants were excluded if they had an acute or terminal illness or cognitive impairments that would limit their ability to participate meaningfully.

Trained local research staff recruited participants from homes, churches and health centres, with support from village leaders and trusted community members, including church leaders. Convenience sampling was used for survey enrolment. For qualitative interviews to capture diverse perspectives, purposive maximum variation sampling was employed based on participants’ age, CC awareness, screening status and access to CC information.

Survey instruments and interview guides were developed in English using validated tools and prior studies and interview guide.^23,24,25,26 Instruments were translated into Kipsigis and back-translated into English to ensure accuracy. Bilingual registered nurses and a local gynaecologist assessed the tools for face validity and cultural appropriateness. Pilot testing with 24 participants led to minor revisions for clarity and cultural sensitivity. Quantitative data were collected using the REDCap Mobile Application. Midway through data collection, a protocol amendment was made to address knowledge gaps identified among participants. An information sheet summarising key facts about CC was introduced, and research assistants reviewed this information with each participant after administering the survey, especially for those with limited literacy.

Semi-structured interviews were conducted in Kipsigis, audio-recorded, and focused on barriers and facilitators to CC screening, trusted information sources, and preferred communication methods. Interviews lasted approximately 45 min to 60 min. Interviewers fostered open discussions to elicit rich, reflective responses. Participants received an incentive of 330.00 Kenyan shillings (KSh) for completing the survey, and an additional KSh 330.00 (approximately $2.31) for participating in interviews, in appreciation of their time and contributions.

Participants who had heard about CC were asked to identify their information sources, including health workers (nurses, community health workers, or doctors), news media (radio or television), social networks (family, friends or neighbours), religious leaders, teachers, printed materials, or others. Multiple responses were allowed. Screening status was assessed with a binary (yes/no) question: ‘Have you ever been screened for cervical cancer in your lifetime?’

Health literacy was measured using the 12-item Health Literacy Test for Limited Literacy (HELT-LL) populations.²⁷ Sample questions include: ‘if you are worried about a health problem, do you usually ask your friends and neighbours first for information and advice before going to the clinic?’ and ‘how often do you need to have someone help you when you read instructions, pamphlets, or other written material from your doctor/nurse/pharmacist?’

Cronbach’s alpha was 0.60 in the original study. In our sample, four items with Cronbach’s alpha < 0.49 were removed (Online Appendix 1), resulting in an 8-item scale with a Cronbach’s alpha of 0.57 within the acceptable range.²⁸ The remaining items assessed numeracy, print, critical and communicative literacy. Total scores ranged from 0 to 8. Participants scoring below the mean (4.2) were categorised as having inadequate health literacy.

The sample size estimation assumed that 50% of women had no exposure to CC information.²⁹ Based on an anticipated screening rate of 1.6% among the unexposed and 14% among the exposed,^11,30 a sample of 140 participants was required to achieve 80% power at α = 0.05.^31,32 To account for potential non-response, we aimed to recruit at least 150 participants.

The final sample included 150 survey respondents and a sub-sample of 24 participants from the survey pilot (total N = 174). Descriptive statistics (means, standard deviation [s.d.], medians, ranges, frequencies and percentages) were used to summarise participant characteristics. Bivariate logistic regression was used to examine associations with the outcome. No covariates were significantly associated with CC screening in bivariate analysis; thus, multivariable regression was not performed. Analyses were conducted using STATA/BE (version 17).

For qualitative data, interviews were transcribed verbatim in Kipsigis and coded by four bilingual coders using Dedoose. A combination of inductive and deductive thematic analysis was employed.³³ Coding was conducted in Kipsigis to preserve cultural meaning, with each transcript independently coded by two coders. Themes were discussed iteratively. Translated English summaries were produced for reporting. Quantitative and qualitative results were integrated using joint display techniques.³⁴

Ethical clearance to conduct this study was obtained from the Johns Hopkins University, National Commission for Science, Technology & Innovation (NACOSTI/P/23/23289, Ref: 774947), the Johns Hopkins Medicine Human Subjects Research Institutional Review Board (IRB00357410) and the Amref Ethical and Scientific Review Committee in Kenya (REF: AMREF-ESRC P1394/2023). Verbal informed consent was obtained from all participants prior to data collection. Interviews were conducted in Kipsigis, with participants given the opportunity to ask questions before consenting. To maintain confidentiality, no identifying information was collected. All data collection devices were password-protected, and audio files were securely uploaded to a cloud-based database and deleted from recording devices after transcription.

Table 2 summarises the participants’ access to CC information, needs and preferences for CC information dissemination. The majority of participants had heard of CC (82.2%), mainly from the news media (36.8%), followed by health workers (24.1%) and social networks such as family or friends (21.3%). Few participants cited teachers or religious leaders (1.7%) as sources. Among those who had heard of CC (n = 143), only 39.9% demonstrated good knowledge. Most women (93.7%) reported never having been screened for CC, and 92% felt uninformed about the disease. Reasons for not screening are summarised in Appendix 1 – Table 2-A1.

Less than half of the participants (48.3%) had actively searched for CC-related information. When asked about their preferred and trusted sources for receiving CC information, the most frequently endorsed sources were nurses and doctors, followed by the radio. The television, mobile applications (WhatsApp or Facebook) and short message services (SMS) were least preferred. Verbal communication was considered the most effective method for delivering CC information (93.7%), followed by videos (12.6%), and pictures (11.5%) were the least preferred. Preferred locations for receiving CC information were health facilities (40.2%), followed by community settings (20.7%).

Bivariate analysis indicated that among participants who had heard about CC (n = 143), the odds of ever being screened were 4.66 times higher for each additional source of CC information accessed (OR = 4.66, 95% confidence interval [CI]: 1.19–18.25). Health literacy was not significantly associated with lifetime CC screening among this subgroup (OR = 1.39, 95% CI: 0.95–2.02). Therefore, mediation analysis to test health literacy as a mediator between information sources and screening uptake was not conducted.

The qualitative findings underscore the pivotal role of accessible and trusted CC information in promoting screening uptake among rural women. Participants frequently cited health workers, radio broadcasts and community leaders as the primary sources that motivated their decision to screen – often marking their first meaningful exposure to the severity of CC and the importance of early detection. Many also voiced a strong need for comprehensive CC education, covering causes, symptoms, prognosis, treatment options and where to access services. Preferences for how, by whom, and where this information should be communicated were clear, highlighting a deep appreciation for verbal, face-to-face messaging delivered by trusted figures in accessible and private settings. These themes are elaborated below.

Access to timely and relatable CC information significantly influenced screening behaviour. Most women who underwent screening attributed their decision to persuasive messages from health professionals, media or local leaders. A radio report on the death of a prominent leader because of CC, for example, sparked fear and awareness in one participant, prompting her to seek screening:

Across the sample, women expressed a desire for detailed, CC-literacy-focused education. They sought to understand what causes CC, how it manifests, whether it is treatable and where services are available. This educational gap was articulated by a participant in her early 30s:

Participants articulated specific preferences regarding the source, format and setting of CC information:

Trusted Messengers: Women preferred information from familiar and authoritative figures, such as health workers, community leaders and religious leaders. Face-to-face delivery enhanced credibility and allowed for immediate clarification. In contrast, SMS messages were viewed with scepticism, often ignored or not taken seriously:

Preferred Formats: Verbal communication was favoured, especially among those with limited literacy or without smartphones. Videos were seen as useful complements, while illustrations helped with anatomical understanding:

Accessible Settings: Preferred venues ranged from hospitals and community gatherings to individual homes. However, the presence of a trusted health worker often mattered more than the setting itself:

Four actionable strategies emerged to guide effective and inclusive CC information dissemination in rural contexts:

Leverage Trusted Messengers and Channels: Health workers, CHWs and local leaders (e.g. chiefs, clergy) were viewed as the most credible messengers. Trusted channels included radio and in-person communication, while SMS and official alerts were often dismissed:

Health-Worker-Led Education Coupled with Screening: Participants favoured integrated approaches where health workers provided both education and immediate access to screening – enhancing trust and convenience:

Multi-Channel and Community-Based Outreach: To ensure wide coverage, participants suggested blending traditional (e.g. radio) and digital platforms (e.g. WhatsApp, Facebook) with community outreach, especially for women relying on traditional medicine or facing mobility barriers:

Tailored Delivery to Balance Reach and Privacy: While group education in churches or schools allowed for broad reach, many women valued private, home-based visits to discuss sensitivities:

This highlights the need for flexible strategies that respect both social dynamics and individual comfort levels.

The integrated findings from this mixed methods study highlight the critical role of CC information in promoting screening uptake among rural women. Quantitative data showed that each additional increase in access to CC information was associated with over fourfold higher odds of being screened. Qualitative findings reinforced this association. Women commonly described being motivated to seek screening after receiving information from trusted health workers, radio broadcasts or community leaders. Despite this exposure, however, more than 90% still felt inadequately informed about CC – especially regarding its causes, symptoms, treatment options and prognosis.

These knowledge gaps point to the need for ongoing, CC-literacy-focused education. Participants favoured clear, verbal communication delivered by trusted messengers, particularly health workers. They also preferred receiving information in familiar and accessible settings such as health facilities, homes and community gatherings. Additionally, the integration of findings emphasised the importance of using multiple communication channels. A combination of traditional media (like radio and community meetings) and digital platforms (such as WhatsApp and Facebook) can help extend the reach of CC information while accommodating different preferences and access levels. Together, these findings support the implementation of community-engaged, health-worker-led interventions. Such approaches should prioritise trust-building, bridge persistent knowledge gaps, and adapt delivery methods to local contexts to ensure both effectiveness and sustainability. Table 3 shows a summary of integrated qualitative and quantitative results.

This study has some limitations. Firstly, its cross-sectional design precludes causal inference. Secondly, reliance on self-reported data introduces the potential for recall bias. Thirdly, the Health Literacy Test for Limited Literacy (HELT-LL) demonstrated moderate internal consistency reliability (α = 0.57) even after item reduction; however, no better-validated alternative existed for this context, and the tool was retained for comparability with prior studies. Fourthly, the actual screening uptake (6.3%) was lower than expected based on prior estimates (14%), which may have reduced the power to detect associations involving health literacy. Fifthly, qualitative interviews did not capture participants’ preferences for specific radio programmes, broadcasters or social media channels, limiting insights into message delivery optimisation. Lastly, the use of a convenience sampling strategy and recruitment through community gatekeepers (i.e. church and village leaders) may have introduced selection bias, limiting the representativeness of the sample and the generalisability of findings to broader rural populations, especially those less engaged with formal or community structures.

Despite these limitations, this study has notable strengths. It is among a few studies in SSA to quantitatively assess health literacy in a low-education rural sample and examine its relationship with CC information exposure and screening behaviours. The mixed methods design enriched understanding by integrating subjective experiences with objective associations, enhancing the credibility and applicability of the findings. Methodological rigour, including culturally appropriate tool adaptation, bilingual data collection, and robust thematic analysis, further strengthens the validity of the results.