Effectiveness of a Dementia Anti-Stigma Intervention in Rural Kenya

The overall aim of this research is to refine and evaluate a dementia stigma reduction intervention tailored to meet the needs of people in Kenya, in preparation for future implementation and adoption within existing policies.

This will be achieved through the following three specific aims. i) AIM 1: To adapt dementia-related knowledge and attitudes tools and develop a representative dementia stigma profile of the general public in Kenya.

ii) AIM 2: To refine an existing anti-stigma intervention to better target culturally specific misconceptions and negative beliefs about dementia.

iii) AIM 3: To determine the effectiveness of the anti-stigma intervention through a stepped wedge cluster randomized trial (SWT) among members of the general public.

Study site is in Makueni County in rural Kenya AIM 1: Adapting a dementia-related knowledge and attitudes tools and developing a representative dementia stigma profile of the general public in Kenya A survey of 600 adults (18 years and older) will be conducted in Makueni County, a rural setting in Kenya. Individuals in rural communities have higher rates of dementia, experience more health disparities and additional challenges with cultural belief systems, and utilization of care compared to their urban counterparts. Participants will be recruited from three geographically defined areas served by six Community Health Workers (CHWs), and ensure a diverse demographic profile based on age, sex and socio-economic status. The three geographically selected rural areas are representative of other rural regions in Kenya due to their similar settings. Participants will complete measures of knowledge (Dementia Knowledge Assessment Scale (DKAS)) and attitudes (Dementia Attitudes Scale (DAS)). DKAS is a reliable and valid tool for assessing dementia knowledge across a range of participants including members of the general public. It has been used among older adults in Africa. DAS is based on the tripartite model of attitude, has been validated for members of the general public and has a high Cronbach's alpha coefficient ranging from 0.83 to 0.85. To ensure that the measures are suitable for use within Kenya, they will undergo cross-cultural adaptation (forward and backward translations will be conducted, inconsistencies checked, and then pilot tested (n=24 general public members). Up to twenty individual interviews will also be held with a convenient sample of members of the general public who completed the survey to better explore what stigmatizing beliefs are held in the communities and why they form.

Recruitment and approach: Members of the general public (n=24 for piloting the study tools and procedures) and (n=600 participants who will be recruited for the main study), aged 18 years and above will be recruited by the CHWs. Households will be approached (i.e., door knocking) upon a convenience basis in the first instance, within geographic regions that the CHWs serve. CHWs will ensure that there is equal representation based on age, sex and socioeconomic status. CHWs will be familiar with the demographics of participants that reside in the local community. All participants will provide either written/thumb print informed consent. The research team will liaise with the CHWs on a weekly basis to reflect on the recruitment progress, but also to identify underrepresented groups. Door knocking benefits from minimizing sources of bias that might occur through other channels (e.g., social media), who reflect certain demographics. The team will however be vigilant that door knocking during the day may lead to an under-representation of working population (e.g., younger, males). In instances within a household where an underrepresented group resides but not present during the initial visit, CHWs will reorganize the visit for a time and date convenient for the potential participant. Guided by the need to achieve thematic saturation, up to 20 members of the general public who completed the survey will also be identified by CHWs through convenience sampling based on their availability, to better explore what stigmatizing beliefs are held in the communities and why they form.

Analysis plan and sample size: Descriptive data will be reported to identify misconceptions and common beliefs within the sample. Regression models will be developed to identify what factors are associated with stigma. To detect a small effect size (f2=0.04) and enter 20 predictors in the model (Power =0.8, alpha = 0.05), accounting for 10% missing data, a sample size of 600 is required. Inductive thematic analysis will be used for the qualitative interviews. Participants will be recruited until thematic saturation is achieved (i.e., until no new themes are inductively identified). Thematic saturation typically occurs between 9 and 17 participants. Triangulation techniques will be used to integrate quantitative and qualitative data.

AIM 2: Refining an existing anti-stigma intervention to better target culturally-specific misconceptions and negative beliefs about dementia. A consortium of the 20 key stakeholders (people with dementia, carers, clinicians, policy makers, academicians and researchers) will be gathered to review the summary report comprising of AIM 1 findings, which will help to identify stigma drivers and target common culturally specific misconceptions and negative beliefs about dementia. Their distribution will be purposive, based on the relevance of stakeholders to the research objectives. This will be discussed alongside the existing intervention to identify any gaps, or whether further refinement of the content is required. The intervention includes four sessions that promote: (i) understanding of dementia, (ii) demystifying myths and misconceptions, (iii) empathetic connections using videos of people with dementia and carers (virtual social contact) and (iv) social inclusion through a case vignette and discussions.

Recruitment: From the list of 20 stakeholders, persons with lived experience will be recruited from Alzheimer's Dementia Organization, Kenya. The investigators will draw on their established networks of collaborators to identify the other stakeholders for the refinement process.

AIM 3: Determining the effectiveness of the anti-stigma intervention through a stepped wedge cluster randomized trial (SWT) among members of the general public This will involve sequential delivery of the intervention to six community-level groups from one site at a time, until all the three sites are exposed to the intervention (n=184). Using participants surveyed in AIM 1, six randomly selected community- level groups consisting of 10-12 members from each of the three geographically defined areas (sites) will be randomly allocated to start on one of three dates (three months apart). Randomization to receive the intervention will be conducted by a blinded independent researcher. Participants will be blinded to allocation while the research assistant will be blinded to the schedule of intervention delivery. Following the listing of the 600 assessed participants, sites will be block randomized into six community-level groups (clusters), then randomly ordered to receive the intervention. Cluster randomization will reduce contamination before the intervention phase, expected to be seen at an individual level. Although all groups will start with a three-month "control condition", six groups from each site will receive the intervention after every three months. With the three-month interval, the rollout program is expected to take about 12 months. The SWT design ensures that all clusters receive the intervention rather than entirely denying benefits to control groups, since there is evidence that the proposed intervention has a short-term positive effect in a context that lacks implementation of evidence-based dementia anti-stigma interventions. Stigma-related outcome data will therefore be collected at baseline (three months before being exposed to the four-session group awareness program), switch points (rollout period during which the clusters will cross over from the control phase), 6- and 12-months post two-week intervention.

Recruitment and retention: Members of the general public (n=184) will be selected from the participants identified in AIM 1. As such, all participants will be aged 18 years and older, able to speak Kamba and have capacity to consent. All participants will have been given the option to participate in the anti-stigma intervention in AIM 1. If there are more than 184 participants (from AIM 1) who have consented to be contacted about the anti-stigma intervention, all potential participants will be randomly listed within each village site. Then from this list, six sets of 10-12 members will be randomly selected and randomly allocated to start on one of three dates (three months apart).