Qualitative Analysis in Public Health, Faster

According to the PLOS One study on the ELLY intervention, qualitative analysis in public health can reveal how community interventions are experienced and how participants perceive pathways from activities to wellbeing outcomes. The ELLY researchers interviewed 20 participants after a 12-week intervention delivered from September 2023 to November 2023, then used framework analysis to develop a conceptual logic model.

• The ELLY study recruited 75 residents in two disadvantaged Forth Valley communities between September and November 2023, according to the November 5, 2026 PLOS One publication.
• The ELLY qualitative sample included 20 interview participants, 10 from each community, and all 20 agreed to participate, according to the November 5, 2026 PLOS One publication.
• The ELLY intervention combined free soup twice weekly, community-asset engagement, a £25 shopping card after at least 9 activities in 12 weeks, goal setting, and self-monitoring, according to the study authors.
• The ELLY interviews identified perceived links between social connection, confidence, healthier routines, dietary changes, new skills, and wellbeing, while the authors stated that qualitative data alone could not establish definitive causation.
• AI-enabled qualitative analysis can help public health teams organize large transcript sets, compare communities, surface recurring themes, and connect coded evidence to an evaluative logic model.

The ELLY study examined how residents experienced a co-designed, incentive-based healthy weight and wellbeing intervention in two disadvantaged Scottish communities. According to the study published in PLOS One on November 5, 2026, the intervention ran for 12 weeks, from September 2023 to November 2023, and included free soup, a loyalty card, community activities, goal setting, information resources, and optional self-monitoring.

The study recruited 35 adults in one community and 40 adults in the other, producing a total feasibility sample of 75 residents, according to the study authors. Researchers invited 20 participants to qualitative interviews at 12 weeks, with 10 participants from each community, and all 20 agreed to take part.

The researchers conducted approximately 30-minute, face-to-face, semi-structured interviews in participants' homes or community settings, according to the PLOS One methods section. Two researchers collected the interviews, used an encrypted audio recorder, took field notes, and explored participants' experiences of the intervention, community activities, goals, health, wellbeing, and behaviour.

The research team developed an initial coding framework after reviewing five interviews, refined the framework through team discussion, and independently cross-checked 10% of the data, according to the study authors. The team then used interview evidence to build a conceptual logic model informed by the COM-B model, distinguishing intervention mechanisms from reported outcomes.

Participant accounts described several pathways rather than a single outcome. One participant said, "It’s actually got me going out more, meeting other people, having a laugh, which is important… It’s been good, instead of sitting here vegetating in front of the telly" (C1, P36), as quoted in the PLOS One study. Another participant said, "I feel the ELLY Project was very helpful for me…I did a lot of activities, achieved my goal, which was my weight… Now I walk every night, " (C2, P34), as quoted in the same study.

Public health researchers should analyze intervention interviews as connected pathways, not as isolated themes. The ELLY authors reported that social connection, confidence, community participation, diet, physical activity, and wellbeing overlapped, and they used a logic model to represent those relationships.

Researchers should separate participant perception from causal proof. The ELLY authors wrote that "this qualitative data alone is insufficient to establish definitive causation, " because the interviews captured perceived pathways during a 12-week period rather than proving long-term effectiveness.

Evaluation teams should preserve negative and uneven experiences alongside positive accounts. The ELLY study reported that some participants found groups unwelcoming, irregular, difficult to approach, or poorly communicated, showing why qualitative coding should retain barriers, exceptions, and non-participation rather than only summarizing successful engagement.

AI-enabled qualitative analysis is most useful when it makes evidence traceable. A defensible workflow links each theme to transcript excerpts, compares findings across communities or participant characteristics, records the coding structure, and leaves final interpretation with the research team.

• Use thematic analysis to identify recurring experiences such as social connectedness, healthier routines, and confidence.
• Use cross-segment analysis to compare communities, engagement levels, demographic groups, or intervention components.
• Use frequency analysis as a descriptive aid, not as a substitute for contextual interpretation.
• Use co-occurrence views to examine relationships between codes such as community participation, wellbeing, and behaviour change.
• Retain direct quotations and negative cases so reports show both reach and limitations.

The ELLY study demonstrates why qualitative analysis in public health must connect participant voices to intervention components, contextual barriers, and plausible pathways to outcomes. The study's 20 interviews generated detailed evidence about social connection, confidence, behaviour, diet, skills, and wellbeing, while also showing the limits of causal claims from qualitative data alone.

AI-enabled analysis can make this work more manageable by helping researchers structure transcripts, compare segments, explore code relationships, and retrieve supporting quotations. Evidano adds these capabilities while keeping human interpretation and research governance central.

Public health researchers and evaluators can try Evidano for free to analyze interviews, open-ended responses, and supporting documents in a more organized and traceable workflow.