What ELLY Shows: AI Qualitative Analysis for Public Health

According to the PLOS One study on the ELLY intervention, qualitative interviews suggested that social connection and wellbeing were more motivating than weight loss for participants in two disadvantaged Scottish communities.

• The ELLY intervention ran for 12 weeks, from September 2023 to November 2023, and involved 75 residents across two communities, according to Cowie and colleagues in November 2026.
• The qualitative component included 20 interviews, with 10 participants from each community, according to the November 2026 PLOS One study.
• The study used semi-structured, approximately 30-minute interviews and framework analysis, with 10% of coded data independently cross-checked by a second researcher, according to the study methods published on November 5, 2026.
• Participants described reported pathways involving stronger social connectedness, healthier dietary habits, positive behavior changes, improved confidence, and new skills, according to the study results published on November 5, 2026.
• The authors stated that qualitative data alone was "insufficient to establish definitive causation, " making the findings useful for understanding perceived mechanisms rather than proving intervention effectiveness.

The ELLY researchers used qualitative interviews to examine how participants believed a co-designed community intervention influenced health, wellbeing, and behavior. According to the November 2026 PLOS One study, the intervention combined free soup twice weekly, a loyalty card, a £25 shopping card for attending at least 9 community activities over 12 weeks, optional goal setting, information resources, and self-monitoring.

According to the study methods, researchers recruited 35 participants in one community and 40 in the other between June 20 and August 31, 2023. Researchers purposively selected 20 interview participants to reflect variation in age, sex, employment, household composition, BMI, intervention engagement, and goal setting.

The researchers developed an initial coding framework after reviewing five interviews, refined the framework through team discussion, and divided themes into subthemes. According to the study published on November 5, 2026, one researcher led the coding, a second researcher independently checked 10% of the data, and a third researcher helped resolve disagreements.

The analysis identified mechanisms such as snowball recruitment, access to community assets, goal setting, and social interaction. The research team then connected those mechanisms to reported outcomes in a conceptual logic model informed by the COM-B behavior change model.

Participant language illustrates why qualitative analysis adds context to outcome measures. One participant said, "It’s actually got me going out more, meeting other people, having a laugh, which is important, " while another said, "It introduced me to other places here." The PLOS One authors attributed these comments to increased social interaction and awareness of local community assets.

The ELLY study suggests that public health teams should analyze intervention mechanisms and participant experiences alongside headline outcomes. According to the November 2026 PLOS One findings, participants often connected healthier routines to social interaction, confidence, accessible activities, and a sense of purpose rather than to weight goals alone.

The ELLY findings also show why segment-level analysis matters. According to the study authors, most participants described positive social connection, but some participants found groups unwelcoming, difficult to approach, irregular, or poorly communicated. A single overall theme such as "community engagement" could hide these contrasting experiences.

Researchers should therefore compare themes by community, engagement level, demographic characteristics, and intervention component. The ELLY study reported different community contexts, including a rural community of approximately 1, 500 people and an urban community of approximately 9, 000 people, so cross-segment analysis can help distinguish transferable mechanisms from local conditions.

The study also provides an important evidence boundary. According to the authors, the interviews captured participant perceptions during a 12-week intervention, did not include the 9 participants who dropped out, and could not establish definitive causation. Public health reporting should preserve those limitations rather than present qualitative perceptions as clinical proof.

For health research, qualitative analysis should remain non-diagnostic and research-focused. Participant accounts can explain experience, acceptability, barriers, and perceived pathways, but they do not independently establish medical effectiveness.

The ELLY study shows how qualitative analysis can explain the pathways behind a community health intervention, including social connection, confidence, healthier routines, and local participation. The study also shows why researchers must compare communities and preserve negative or exclusionary experiences instead of reporting only positive themes.

AI-enabled qualitative research can reduce the manual burden of organizing interviews and comparing segments, but credible findings still depend on transparent methods, source verification, ethical data handling, and human interpretation. Evidano gives public health and UX researchers a practical environment for thematic, cross-segment, and document-based analysis.

To analyze interview transcripts and public health documents with AI, try Evidano for free.