From Stories to Insights: AI Qualitative Research Analysis

According to the 2026 PLOS Neglected Tropical Diseases study, “Owning the narrative”, participatory storytelling in Karachi was associated with reduced shame and concealment, greater self-acceptance, and stronger individual empowerment among people affected by leprosy.

AI qualitative research analysis can help researchers compare interview themes, participant groups, and levels of change, but the Pakistan study shows why human interpretation, ethical safeguards, and attention to structural context remain essential.

• The study recruited 17 participants for the qualitative evaluation conducted in June 2025, including 16 people who took part in photovoice and 14 who completed in-depth interviews.
• All participants reported increased confidence, while the study found that the strongest effects occurred at the individual level through reduced internal and anticipated stigma.
• The 2026 study found that community misconceptions, gender norms, visible disability, transport costs, and economic constraints continued to limit broader social change.
• One participant described the intervention as making them feel “very light in my heart” after sharing their story, according to the study’s interview findings.
• The researchers concluded that sustained community contact, peer leadership, inclusive outreach, and links to socioeconomic opportunities are needed to extend impact beyond individual empowerment.

The 2026 PLOS study used interviews and photovoice to understand how a 2024 participatory storytelling pilot affected stigma and empowerment among people affected by leprosy in Karachi.

The researchers collected data during a four-week period in June 2025, after recruiting participants between May 30 and June 20, 2025. The researchers used purposive and maximum-variation sampling to include differences in gender, age, disability status, time since diagnosis, and social circumstances.

The study included 14 in-depth interviews, 13 of which were completed by people who had also participated in photovoice. Interviews lasted 60 to 90 minutes and were conducted in English through live Urdu interpretation, according to the study methods.

The researchers analyzed the data thematically in NVivo 14. The analysis combined deductive codes based on stigma, empowerment, and the Social Ecological Model with inductive codes that emerged from participant accounts.

The Social Ecological Model enabled the researchers to examine change at five levels: individual, interpersonal, community, institutional, and structural. That layered approach matters for AI-assisted analysis because a single keyword, such as confidence or disclosure, can represent different mechanisms at different social levels.

The study also used participant photographs, captions, and group discussions to triangulate interview findings. One participant summarized the shift in identity by saying, “We are different from others, not less, ” according to the photovoice discussion reported in the study.

Public health researchers should treat AI qualitative research analysis as a way to expose patterns across narratives while retaining participant voice and social context.

The Pakistan study found that storytelling was most strongly associated with intrapersonal empowerment, including confidence, self-acceptance, reduced fear, and a stronger sense of purpose. The study did not establish a causal effect because it lacked baseline and comparator data, so researchers should describe these findings as perceived impacts and possible mechanisms rather than definitive treatment effects.

The study also found that individual gains did not automatically produce community-level change. Participants reported persistent misconceptions and discrimination, especially for people with visible impairments, while women faced constraints related to permission, mobility, and mixed-gender participation.

For research teams, these findings create a practical coding framework:

• Code identity and emotional change separately from community attitudes, institutional support, and structural barriers.
• Compare themes by gender, disability status, employment, age, and participation format where the sample supports responsible comparison.
• Distinguish internal stigma, anticipated stigma, and enacted stigma instead of combining them into a single stigma code.
• Use participant quotations and visual narratives as evidence for interpretation, not as interchangeable substitutes for structured outcomes.
• Record limitations such as translation, social desirability, small samples, and the absence of comparator data alongside the findings.

The 2026 Pakistan storytelling study shows that qualitative research can reveal how narrative ownership supports confidence and self-acceptance while also exposing the structural barriers that limit wider change. AI qualitative research analysis can make this kind of evidence easier to organize across interviews, translations, participant groups, and analytic frameworks, but researchers must preserve context and maintain ethical oversight. Evidano helps research teams move from raw qualitative data to thematic and cross-segment insight with reviewable analysis workflows. Try Evidano for free and turn complex qualitative data into clearer research evidence.