Health researchers often need to analyze interviews, photovoice discussions, field notes, and open-ended responses while preserving context, emotion, and participant voice. AI qualitative data analysis can accelerate that work, but the study design, coding framework, translation process, and ethical safeguards still determine the quality of the findings. The November 2, 2026 study in PLOS Neglected Tropical Diseases offers a useful example: researchers combined in-depth interviews and photovoice to examine how participatory storytelling influenced stigma and empowerment among people affected by leprosy in Karachi, Pakistan. This article explains what the study found, how its qualitative method worked, and how AI can support similar health research without replacing researcher judgment.
Key Takeaways
According to the November 2, 2026 study Owning the narrative: Exploring the impact of a creative storytelling intervention on stigma and empowerment among persons affected by leprosy in Pakistan, participatory storytelling was associated with reduced self-stigma, greater confidence, and stronger individual empowerment, although wider community stigma remained.
- The 2026 PLOS Neglected Tropical Diseases study included 17 people affected by leprosy, with 16 participating in photovoice and 14 completing in-depth interviews.
- Data collection took place during a four-week period in June 2025 after recruitment between May 30 and June 20, 2025.
- The researchers found the strongest reported changes at the individual level, including movement from shame and concealment toward self-acceptance, openness, and confidence.
- The study reported that gender norms, visible disability, community misconceptions, travel costs, and economic constraints limited broader social change.
- AI qualitative data analysis can help researchers organize transcripts, compare themes across participant segments, and identify recurring language, but human researchers must interpret sensitive narratives and protect participant dignity.
How the Karachi Storytelling Study Used Qualitative Analysis
The Karachi study used thematic analysis to examine how storytelling influenced stigma and empowerment across individual, interpersonal, community, institutional, and structural levels.
According to the 2026 PLOS study, the intervention was piloted in 2024 through the Marie Adelaide Leprosy Centre in Karachi and combined participatory storytelling, peer-support groups, health education, skills-building, advocacy, and contact-based activities.
According to the study methods, researchers conducted 60 to 90 minute interviews in English through live Urdu interpretation, retained relevant Urdu expressions where appropriate, and complemented interview data with participant photographs, captions, and group discussions.
According to the study, the research team used a preliminary deductive codebook based on stigma, empowerment, and the Social Ecological Model, then expanded the codebook through inductive coding. A senior researcher reviewed part of the coding process, and disagreements were resolved through discussion and consensus.
One participant described the emotional effect of disclosure by saying, "I have no fear, no shame, not feeling any guilt. I’m feeling like... I feel lighter and so relaxed." The PLOS study attributes this quotation to participant P06, a male government worker in his 40s.
The study authors caution that the findings describe perceived impact and possible mechanisms of change rather than definitive causal effects because the research had no baseline or comparison group.
Findings Snapshot for AI Qualitative Data Analysis
| Date or study detail | Metric | Value | Implication for qualitative researchers |
|---|---|---|---|
| November 2, 2026, according to the PLOS study | Total qualitative study participants | 17 | Small samples can generate detailed insight, but findings should not be treated as broadly representative without further testing. |
| June 2025, according to the PLOS study | Photovoice participants | 16 | Visual data can add participant-led context that may not appear in interview transcripts alone. |
| June 2025, according to the PLOS study | Completed in-depth interviews | 14 | Interview coding can be triangulated with photovoice and group-discussion data. |
| June 2 to June 25, 2025, according to the PLOS study | Interview duration | 60 to 90 minutes | Long interviews produce rich material that benefits from structured transcript review and thematic organization. |
| November 2, 2026, according to the PLOS study | Women represented in the participant sample | 4 of 17 | Segment comparisons must account for uneven representation and limited gender-specific evidence. |
| 2024, according to the PLOS study | Pakistan Grade 2 Disability among newly diagnosed patients | 20.7%, compared with a global average of 5.5% | The health context shows why stigma, disability, delayed diagnosis, and participant experience need to be analyzed together. |
Implications for Health and Policy Researchers
Health and policy researchers should analyze both personal change and the social conditions that enable or restrict that change.
According to the 2026 PLOS study, the storytelling intervention produced its clearest reported benefits in intrapersonal empowerment, including confidence, self-acceptance, and reduced anticipated stigma. The study also identified interpersonal gains such as peer solidarity and improved communication skills.
According to the same study, community-level effects were more limited because contact activities were episodic, institution-based, and directed toward a relatively narrow audience. The finding suggests that qualitative analysis should distinguish between changes in how participants see themselves and changes in how communities treat them.
AI-assisted coding can help teams compare themes by gender, disability status, employment context, time since diagnosis, location, or intervention exposure. Researchers should use those comparisons as prompts for review rather than as automatic conclusions, especially when one segment contains only a few participants.
The study also demonstrates why translation and context matter. Interviews were conducted in Urdu with live interpretation, and the researchers used repeated listening, meaning checks, translator discussions, and reflexive analysis to reduce semantic loss. AI translation or transcription should therefore be reviewed against recordings and culturally specific language.
How Evidano Helps with AI Qualitative Data Analysis
Problem: Rich interview datasets are slow to organize
Evidano is an AI-powered qualitative data analysis platform that helps researchers analyze interviews, open-ended surveys, and documents.
Evidano can support thematic, content, frequency, and cross-segment analyses across research materials, helping teams move from raw transcripts to a structured view of recurring themes and differences. Researchers can review the underlying evidence before incorporating findings into a report.
Researchers can learn more about Evidano’s research workflows on the Evidano features page.
Problem: Multilingual interviews can introduce meaning loss
Evidano provides transcription and translation workflows with custom dictionaries, which can help research teams preserve domain-specific terms and review multilingual qualitative data more consistently.
The 2026 PLOS study shows why human review remains necessary: the researchers used live Urdu interpretation, repeated meaning checks, and translator discussions to address ambiguity and cultural nuance.
Problem: Researchers need to compare lived experiences across segments
Evidano’s cross-segment analysis can help teams compare themes across groups such as gender, disability status, geography, or intervention participation when the dataset contains suitable segment information.
The PLOS study’s sample included 4 women among 17 participants, so any AI-generated segment comparison would require cautious interpretation and explicit reporting of sample limitations.
Problem: Sensitive health narratives require careful data governance
Evidano supports encrypted data handling and states that customer data is never used to train third-party models, a relevant consideration for research involving stigma, health status, photographs, or identifiable stories.
AI tools should support ethical qualitative research rather than replace consent processes, anonymization, participant checking, reflexivity, or researcher accountability.
FAQ: AI Qualitative Data Analysis in Health Research
What did the 2026 leprosy storytelling study find?
+
The 2026 PLOS study found that participatory storytelling was associated with reduced self-stigma, increased self-acceptance, stronger confidence, and greater individual empowerment among participants in Karachi.
The study also found that community misconceptions, gender norms, visible disability, and structural barriers limited wider social change.
How many people participated in the Karachi qualitative study?
+
The study included 17 participants in total, with 16 participating in photovoice and 14 completing in-depth interviews.
According to the study, 13 of the 14 interview participants had also participated in photovoice.
How can AI support thematic analysis of health interviews?
+
AI can support thematic analysis by organizing transcripts, suggesting codes, grouping related passages, counting recurring concepts, and comparing themes across participant segments.
Researchers must validate AI-assisted outputs against source data because sensitive health narratives, translated speech, metaphors, and small samples require contextual interpretation.
Can AI replace a qualitative researcher?
+
AI cannot replace a qualitative researcher’s responsibility for study design, ethical interpretation, reflexivity, contextual judgment, and final claims.
The PLOS study used a codebook, iterative analysis, senior review, consensus discussions, and an intersectional lens, demonstrating that rigor depends on a human-led analytic process.
What are the main limitations of the Karachi study?
+
The main limitations were the small sample, the inclusion of only intervention completers, the representation of women as 4 of 17 participants, the lack of baseline and comparison data, and possible translation and social desirability bias.
According to the study authors, future research should use larger and more diverse samples, longer follow-up, and controlled or matched comparison designs where appropriate.
Conclusion and Next Steps for AI Qualitative Data Analysis
The Karachi study shows that qualitative research can reveal how storytelling changes identity, confidence, peer relationships, and public participation while also exposing the structural barriers that limit impact.
AI qualitative data analysis can make large or multilingual collections easier to organize, compare, and query, but credible health research still requires careful sampling, ethical safeguards, transparent coding, and human interpretation.
Researchers working with interviews, open-ended surveys, or documents can use Evidano to accelerate thematic and cross-segment analysis while keeping participant context central. Try Evidano for free to begin analyzing qualitative data with an AI-assisted workflow.
Topics
- AI qualitative data analysis
- AI analysis of health interviews
- thematic analysis in health research
- qualitative stigma research
- AI-powered qualitative research
Keep reading
- Commentary on NewsStudent Voices: AI Qualitative Analysis in BeninLearn how AI qualitative analysis can reveal student perspectives on school food environments, using the 2026 Northern Benin study as a practical research guide.
- Commentary on NewsAI Thematic Analysis for Healthcare ResearchSee how AI thematic analysis helps healthcare researchers synthesize interviews, surveys, and mixed-methods evidence for better care decisions. Try Evidano.
- Commentary on NewsAI Thematic Analysis for Healthcare ResearchSee how AI thematic analysis supports mixed-methods healthcare research, from interview coding to cross-source synthesis. Explore the SYNC-RF protocol.