Evidano is an AI-powered qualitative data analysis platform that helps import, transcribe, translate, code, and visualize multilingual qualitative datasets. This post unpacks a new PLOS qualitative study (30 interviews; data collected Apr–Sep 2025; published 20 Jul 2026) and shows how researchers and UX/policy teams can run a rigorous qualitative analysis of recovery in Thailand using AI tools. Readers will get a clear methods summary, a short 7-step workflow, and practical notes on handling Thai dialects, translation, and participant confidentiality. Use Evidano by visiting Evidano to import transcripts, apply coded frameworks, and produce thematic and cross-segment analyses in hours instead of weeks.
Key Takeaways
The PLOS study in Chiang Mai shows recovery narratives shift from clinical and functional milestones toward personal recovery shaped by family, community, and Buddhist-influenced meanings.
- Sample and timing: n=30 (10 service users, 5 carers, 15 healthcare professionals), interviews conducted Apr–Sep 2025, paper published 20 Jul 2026.
- Methodology: Experience-Based Co-Design experience-gathering combined with reflexive thematic analysis using NVivo v15.
- Core themes: four themes identified, conceptions of recovery, attitudes, characteristics of successful recovery journeys, and barriers to recovery.
- Practical implication: bilingual checks, relational codes (family, religion, stigma), and EBCD touchpoint mapping are essential for culturally valid qualitative work.
Fast take: what the PLOS study found
The PLOS study found that recovery in Chiang Mai moves from clinical and functional improvement toward personal recovery shaped by family, community, and Buddhist-influenced meanings.
The study by Inta, Grealish & Leamy interviewed 30 participants (10 service users, 5 carers, 15 healthcare professionals) in Chiang Mai between April and September 2025 and identified four themes: conceptions of recovery, attitudes, characteristics of successful recovery journeys, and barriers to recovery.
Read the full paper at PLOS ONE.
Findings snapshot
| Date | Sample | Method | Key themes | Source |
|---|---|---|---|---|
| Interviews Apr–Sep 2025; Published 20 Jul 2026 | n=30: 10 service users, 5 carers, 15 HCPs | Semi-structured interviews; reflexive thematic analysis; 2 feedback workshops | 1) Conceptions of recovery 2) Attitudes 3) Recovery journey features 4) Barriers | PLOS ONE |
What happened: methods & rigor (short)
The methods section reports that the lead researcher conducted 30 semi-structured interviews which were audio-recorded, transcribed verbatim, and translated from Thai/Northern Thai into English.
The lead researcher (NI) conducted 30 semi-structured interviews (23–70 minutes), audio-recorded, transcribed verbatim and translated from Thai/Northern Thai into English.
Reflexive thematic analysis followed Braun and Clarke’s six-phase approach, and NVivo v15 supported coding and touchpoint identification; two feedback workshops validated preliminary findings with participants.
- Translation and transcription safeguards included a glossary of study-specific terms, translator/transcriber briefings, and transcript checks by the lead researcher.
- Ethics note: the study was approved by King’s College London (HR/DP-24/25–45539) and the Chiang Mai Provincial Public Health Office, and the analysis is research-focused and non-diagnostic.
- Limitations: purposive sample, a small carer subgroup, limited medical staff participation, which are typical trade-offs for deep qualitative work.
Implications for qualitative researchers, UX and policy teams
For qualitative researchers
Qualitative researchers should prioritize bilingual checks, coding in English for team access but cross-validating codes in original-language transcripts to preserve cultural nuance.
Capture relational and contextual codes such as family support, religious practice, and stigma, and treat these as analytic variables for cross-segment comparisons (for example, urban versus rural, substance use versus mood disorders).
For UX / service designers
UX and service designers should design interventions that surface family and community touchpoints because participant quotes link recovery to family harmony, meaningful work, nature, and religious practice.
Use touchpoint mapping, as in Experience-Based Co-Design, to turn emotionally significant moments into actionable product or service changes and training content.
For policy & health system teams
Policy and health system teams should translate the Recovery Support Model into measurable service KPIs, for example family education uptake, peer-support referrals, and follow-up contact rates, to move beyond medication-only indicators.
Invest in workforce training, protected time for relationship-based care, and multi-sector pathways that reduce overreliance on families for long-term support.
Do more, faster with Evidano (mapped to this study)
Problem: multilingual transcripts + cultural nuance
Evidano provides automatic transcription with custom dictionaries and PII redaction plus translation workflows that preserve glossary terms to ensure Thai and Northern Thai terms and religious or spiritual terms map consistently.
Use custom dictionaries and translator briefings inside Evidano to maintain consistent term mapping and preserve cultural nuance during translation.
Problem: time-consuming coding and inconsistent codebooks
Evidano supports importing an initial codebook and running AI-assisted coding across transcripts, then refining hierarchical codes into subcodes to enable reproducible reflexive coding while keeping the analyst in control.
Import CHIME-inspired codes or the study’s preliminary codebook into Evidano, run AI-assisted initial coding, and iteratively refine codes with team review.
Problem: comparing segments (users, carers, HCPs) at scale
Evidano enables cross-segment frequency analysis and co-occurrence networks to highlight which themes cluster by role or region, and it provides clickable quotes for fast export into stakeholder-ready evidence packs.
Use cross-segment filters and co-occurrence tools in Evidano to compare users, carers, and HCPs and surface differential theme patterns.
Problem: stakeholder buy-in and reporting
Evidano generates visualizations such as word clouds, co-occurrence networks, and hierarchical theme maps and offers an AI chat over uploaded documents so non-technical stakeholders can query results directly.
All data in Evidano is encrypted and not used to train external models, and further information is available at Evidano.
7-step checklist: reproduce this analysis in two weeks
The 7-step checklist explains how to go from raw audio to stakeholder-ready insights using AI-enabled tooling.
- 1) Collect and catalog audio and video files and consent forms (input).
- 2) Auto-transcribe in Evidano with a Thai/Northern Thai custom dictionary and enable PII redaction (output: cleaned transcripts).
- 3) Import transcripts and the study’s preliminary codebook (CHIME plus inductive codes) into Evidano.
- 4) Run AI-assisted initial coding and review and refine hierarchical codes into subcodes (output: coded corpus).
- 5) Run thematic frequency and cross-segment analyses (users versus carers versus HCPs) and generate co-occurrence networks.
- 6) Extract verified participant quotes and assemble an evidence pack for feedback workshops (EBCD touchpoints).
- 7) Produce visuals and an executive brief and use Evidano’s AI chat to answer stakeholder questions and iterate.
FAQ: qualitative analysis of recovery
What did the PLOS study in Chiang Mai find about recovery?
The PLOS study found that recovery narratives in Chiang Mai shift from clinical and functional milestones toward personal recovery shaped by family, community, and Buddhist-influenced meanings.
The study identified four themes: conceptions of recovery, attitudes, characteristics of successful recovery journeys, and barriers to recovery.
How were interviews conducted and analyzed?
The research team conducted 30 semi-structured interviews, transcribed and translated them, and used reflexive thematic analysis with NVivo v15 and two feedback workshops to validate findings.
Interview durations ranged from 23 to 70 minutes, and translation safeguards included glossaries and transcript checks by the lead researcher.
How should researchers handle Thai and Northern Thai translations?
Researchers should prioritize bilingual checks, code in English for team access but cross-validate codes in original-language transcripts to preserve cultural nuance.
The study used a glossary of study-specific terms and translator briefings to ensure consistent term mapping across translations.
How quickly can a team reproduce a similar analysis with AI tools?
A team can reproduce a similar analysis in about two weeks by following a structured 7-step workflow that includes transcription, codebook import, AI-assisted coding, cross-segment analysis, and visualization.
The checklist in this post describes specific outputs at each step, from cleaned transcripts to stakeholder-ready evidence packs.
Conclusion, next moves
The PLOS study (published 20 Jul 2026) demonstrates that recovery is culturally situated, with clinical and functional milestones often preceding personal recovery while family, community, and spiritual practices play central roles.
Researchers and teams that need to scale reliable, transparent qualitative analysis of recovery narratives, especially across languages and stakeholder groups, should use an AI research platform that preserves context and privacy.
- Ready to try this on your transcripts? Start a pilot on Evidano at Try Evidano for free and import a small set (5–10 interviews) to test transcription, custom-dictionary translation, coding, and cross-segment visuals.
- Read the original paper at PLOS ONE.