Qualitative Analysis of Leprosy Perceptions

According to the November 2026 PLOS Neglected Tropical Diseases study, “Kokobe” and “gwasi”: Perceptions of leprosy in French Guiana, culturally specific terms and explanatory models shaped how people understood leprosy, sought care, disclosed illness, and experienced stigma.

The study shows that AI-enabled qualitative analysis can help health researchers manage complex interview data, but culturally sensitive interpretation, participant confidentiality, and researcher reflexivity remain essential.

• Between April and September 2025, the researchers interviewed 20 patients and 37 community members in French Guiana across 37 interviews.
• The study found that Maroon and Haitian participants often emphasized spiritual explanations such as taboo transgression or curses, while bacterial causation was rarely mentioned.
• The terms “kokobe” and “gwasi” referred to locally recognized manifestations of disease, with “kokobe” associated mainly with deformities and “gwasi” associated mainly with skin lesions.
• The researchers reported that local plant-based remedies were widely used, but the remedies did not appear to reduce adherence to multidrug therapy in the study sample.
• The study concluded that culturally sensitive communication should clarify person-to-person transmission while taking local terminology and practices into account.

The November 2026 PLOS study found that leprosy perceptions in French Guiana were shaped by multiple cultural, linguistic, historical, and biomedical models rather than by one shared explanation.

The researchers used semi-structured interviews because French Guiana has high linguistic diversity, different literacy levels, and overlapping local and biomedical representations of illness. Patient interviews explored lived experience, explanatory models, stigma, socioeconomic effects, and care pathways, while community interviews examined collective memories and representations of leprosy.

The researchers combined deductive and inductive thematic analysis. The initial coding framework drew on the interview guide and a leprosy perception toolkit, while inductive coding identified culturally specific meanings that were not predetermined.

The study reported that only two patients and one community member referred to bacteria as a cause of leprosy. A participant described the disease as “a disease in the blood, caused by bacteria difficult to treat, ” illustrating how biomedical knowledge could coexist with other explanations.

The researchers also documented diagnostic confusion, delayed care, transport barriers, stigma, and the use of plant-based or spiritual practices alongside biomedical treatment. One community member summarized the challenge by saying, “We don't know how it is acquired.”

The study authors cautioned that the findings illustrate a diversity of perspectives rather than generalizable characteristics of every cultural group in French Guiana. The small number of participants within each cultural group limited group-specific interpretation.

AI-enabled qualitative research can make culturally complex health studies easier to organize, compare, and audit, provided that researchers retain responsibility for interpretation and ethics.

The French Guiana study demonstrates why researchers should preserve participant language and local terminology during analysis. Translating “kokobe” or “gwasi” into a generic biomedical code too early could erase differences between visible symptoms, disability, stigma, and local disease categories.

• Researchers should use a code hierarchy that separates biomedical concepts, local terms, explanatory models, care practices, stigma, and access barriers.
• Researchers should compare themes by participant group, language, geography, and interview setting rather than treating all interviews as one homogeneous dataset.
• Researchers should retain quotations with participant identifiers and contextual metadata while limiting access to sensitive transcripts.
• Researchers should use AI-generated themes as an analytic aid, then review code definitions, contradictory cases, translations, and culturally specific meanings manually.
• Researchers should distinguish frequency from significance because an infrequently mentioned belief may still be important for clinical communication or trust.

The study’s own data governance approach reinforces the need for caution. The authors stated that the transcripts could not be publicly shared because detailed narratives, the rarity of the disease, and the small size of some communities could make participants identifiable even after pseudonymization.

For medical and public health research, AI-assisted analysis is non-diagnostic and should support research interpretation, not replace clinical judgment or informed community engagement.

The French Guiana study shows that qualitative analysis of leprosy perceptions can reveal how local language, spiritual beliefs, stigma, treatment experiences, and community support influence health journeys.

AI-enabled analysis can help researchers manage multilingual interviews, build transparent code systems, compare participant segments, and retrieve evidence efficiently, but human cultural interpretation and ethical safeguards remain central.

Evidano gives health and social researchers one workspace for transcript analysis, translation support, thematic coding, cross-segment comparisons, and evidence-focused questions over research documents. Try Evidano for free to organize and analyze your qualitative research data.