AI-assisted thematic analysis: breast cancer in Ghana

According to PLOS ONE (published August 5, 2026), a qualitative study of 14 non-formally trained traditional medicine practitioners in Kumasi, Ghana, found that practitioners can act as both community gatekeepers for breast symptom recognition and as sources of misinformation that may delay biomedical care.

• The study interviewed 14 TMPs (n = 14) between 10 February and 30 April 2023, with interviews lasting 35–60 minutes and averaging 45 minutes, according to PLOS ONE (published August 5, 2026).
• PLOS ONE reported that in Ghana about 80% of breast cancer cases were diagnosed at advanced stages, a statistic cited in the article’s introduction and discussed as a motivation for the study (PLOS ONE, August 5, 2026).
• The PLOS ONE analysis generated four themes, sources of knowledge, multi-dimensional causation beliefs, cardinal signs used for recognition, and experiential diagnostic practices, indicating both an openness to biomedical collaboration and persistent misconceptions (PLOS ONE, August 5, 2026).
• Direct quote from a participant in the study: "I received the medicine from my grandmother; it was a gift. She healed cancer." (TMP004, quoted in PLOS ONE, August 5, 2026).

The PLOS ONE study (published August 5, 2026) conducted an exploratory descriptive qualitative analysis using semi-structured interviews with 14 non-formally trained TMPs in Kumasi to document beliefs, recognition practices, and diagnostic logic.

According to PLOS ONE (August 5, 2026), participants were recruited via the Traditional Medicine Practice Council, interviews were audio-recorded with consent, transcribed from Twi into English, and analyzed inductively in NVivo 12.0 Plus using thematic analysis.

The PLOS ONE paper (received February 4, 2026; accepted July 20, 2026) reports that analyst triangulation and member checking were used to support credibility, and the authors shared the dataset via Figshare (10.6084/m9.figshare.30589079), enabling reproducibility and secondary analysis.

AI-enabled thematic analysis can convert the PLOS ONE study’s insights into reproducible codebooks and measurable indicators within days, not months.

The PLOS ONE study (August 5, 2026) identifies four themes that map directly onto common qualitative codes: knowledge sources, explanatory models, symptom recognition, and diagnostic practices; researchers can seed AI-assisted coding with these themes to accelerate cross-case synthesis.

Researchers designing follow-up studies should note that PLOS ONE used purposive sampling and achieved saturation by interview 12, with two confirmatory interviews (n = 14), which indicates sample-size expectations for similar exploratory qualitative work.

For implementation researchers and health program designers, PLOS ONE’s finding that TMPs attend occasional workshops and consult biomedical staff implies a feasible entry point for training and referral pilots that target the 80% late-stage diagnosis problem the article highlights.

The PLOS ONE study (published August 5, 2026) shows that traditional medicine practitioners in Kumasi are trusted community actors who both recognize suspicious breast signs and hold misconceptions that could delay biomedical care.

AI-assisted thematic analysis can transform the PLOS ONE qualitative findings (n = 14 interviews, data collected Feb–Apr 2023) into reproducible codebooks, measurable indicators, and training materials faster than manual methods.

If you want to operationalize the PLOS ONE findings and build evaluation-ready datasets, consider combining the study’s codebook with an AI-driven workflow that preserves quotes and provenance.

Get started by exploring Evidano features and Try Evidano for free.