AI qualitative analysis of WHiSE 2.0 explains how researchers and program teams can use AI-enabled methods to extract themes, frequency counts, and cross-segment patterns from the WHiSE 2.0 cohort profile. Evidano is an AI-powered qualitative data analysis platform that helps researchers analyze interviews, open-ended surveys, and documents. According to the PLOS One cohort profile (Verma et al., 2026), WHiSE 2.0 enrolled 356 Indigenous participants across Thunder Bay, Sault Ste. Marie, and Sudbury and published baseline results on July 24, 2026 (PLOS One). This post gives a short method playbook and concrete examples (using the WHiSE 2.0 numbers reported in PLOS One) to show how AI-enabled qualitative research can speed synthesis while respecting Indigenous data governance.
Key Takeaways
The PLOS One cohort profile reports that WHiSE 2.0 enrolled 356 Indigenous participants and highlights urgent, locally specific harm reduction gaps across three northern Ontario cities (PLOS One).
AI-enabled qualitative research can rapidly surface culturally grounded themes from WHiSE 2.0's 119-question, interviewer-administered survey and support targeted program design.
- According to PLOS One (Verma et al., 2026), 356 participants were enrolled with site counts reported as Thunder Bay n = 173, Sault Ste. Marie n = 101, and Sudbury n = 82 (published July 24, 2026).
- According to PLOS One (Verma et al., 2026), 84.0% of participants reported knowing what harm reduction is, and 94.8% knew where to access harm reduction services (data reported in the article's baseline tables).
- According to PLOS One (Verma et al., 2026), 43.3% of participants reported injecting drugs in the past three months and 92.1% reported smoking drugs (city-stratified rates appear in the Results tables).
- According to PLOS One (Verma et al., 2026), 71.3% of participants engaged in First Nations, Inuit, or Métis ceremonies, and 76.4% identified as intergenerational residential school survivors, underscoring the need to incorporate cultural practice into analysis and services.
What happened and how WHiSE 2.0 measured it
According to the PLOS One cohort profile (Verma et al., 2026), WHiSE 2.0 used a community-based, explanatory sequential mixed-methods design with an interviewer-administered 119-question quantitative questionnaire as the first round of data collection.
According to PLOS One (Verma et al., 2026), data collection began in November 2022, questionnaires were administered in Thunder Bay from February 2023 to June 2024, in Sault Ste. Marie from May to November 2024, and in Sudbury from May to December 2024 using REDCap for secure capture and offline data entry.
According to PLOS One (Verma et al., 2026), eligibility required participants to self-identify as Indigenous, be aged 16 or older, live in one of the study regions, and have used non-cannabis substances within the past three months; recruitment used purposive and snowball sampling through local harm reduction partners.
Findings snapshot (selected baseline metrics from PLOS One)
| Date / Reported | Metric (source) | Value | Implication (researcher view) |
|---|---|---|---|
| Published July 24, 2026 | Total enrolled (PLOS One) | 356 participants (Thunder Bay 173; Sault Ste. Marie 101; Sudbury 82) | Local stratification supports city-level thematic and cross-segment analysis to guide tailored services |
| Baseline (data collection through Dec 2024) | Percent who know what harm reduction is (PLOS One) | 84.0% | Qualitative probes should focus on barriers to practice despite high awareness |
| Baseline (data collection through Dec 2024) | Mode of use in past 3 months (PLOS One) | 43.3% injected; 92.1% smoked; 31.5% ingested; 38.5% snorted/inhaled | Codebook should include consumption mode, context (alone/with others), and supplies access |
| Baseline (data collection through Dec 2024) | HIV and HCV testing history (PLOS One) | 82.3% ever tested for HIV; 84.3% ever tested for HCV | Link qualitative themes about testing attitudes to service access and discrimination codes |
| Baseline (data collection through Dec 2024) | Cultural connection (PLOS One) | 71.3% participated in ceremonies; 76.4% intergenerational residential school survivors | Prioritize culturally informed coding frameworks and Indigenous data governance in analysis |
Implications for qualitative researchers and program teams
AI-assisted qualitative analysis should prioritize culturally safe coding and community governance because the WHiSE 2.0 team built OCAP and CARE principles into data governance, as described in PLOS One (Verma et al., 2026).
According to PLOS One (Verma et al., 2026), the survey included open-ended responses and a qualitative strand is planned; researchers should therefore combine AI-driven thematic extraction with human review to honor local meanings and avoid misinterpretation.
According to PLOS One (Verma et al., 2026), regional differences (for example, Sudbury shows higher injection rates and Sault Ste. Marie showed very high smoking rates) mean that thematic summaries must be stratified by city and by mode of use to produce actionable service recommendations.
How Evidano helps: mapping WHiSE 2.0 needs to AI features
Problem: Large interviewer-administered survey with 119 items → Slow manual synthesis
Solution: Evidano automates thematic and frequency analysis across 119-question instruments, producing code co-occurrence matrices and segment-level theme counts.
Evidano integrates securely with REDCap exports and supports PII redaction so teams can import WHiSE 2.0 style datasets without exposing identifiers; learn more at Evidano features.
Problem: Need culturally grounded codes and community governance
Solution: Evidano supports custom codebooks, hierarchical codes and subcodes, and analysts can lock and export codebooks for review by Indigenous advisory committees to respect OCAP and CARE principles described in PLOS One (Verma et al., 2026).
Evidano also stores audit trails and role-based access controls to align with community data governance expectations described in the WHiSE 2.0 methods.
Problem: Cross-site comparisons and small cells risk re-identification
Solution: Evidano provides count suppression rules and automated cell-size masking, and outputs can be limited to aggregated summaries to reduce disclosure risk while keeping analytic value.
Evidano’s workflows let teams generate city-stratified thematic dashboards so providers can see that, for example, Sudbury reports higher injecting rates while Sault Ste. Marie reports higher smoking rates, matching the PLOS One findings (Verma et al., 2026).
FAQ: AI qualitative analysis of WHiSE 2.0
How can AI help analyze WHiSE 2.0's interviewer-administered questionnaire?
AI can rapidly produce an initial thematic codebook and frequency counts and surface co-occurrence patterns for WHiSE 2.0-style open responses.
Because WHiSE 2.0 used an interviewer-administered 119-question instrument and collected open-ended responses (PLOS One, Verma et al., 2026), AI speeds the first-pass synthesis and flags quotes for community review.
Does AI replace community-led interpretation for Indigenous data?
No. AI augments human interpretation but does not replace community-led analysis.
PLOS One (Verma et al., 2026) documents OCAP and CARE governance for WHiSE 2.0; teams should use AI outputs as draft summaries that community advisory groups validate and contextualize.
How should analysts handle small cell sizes and re-identification risk in WHiSE 2.0-style analyses?
Answer: Suppress or aggregate small cells and report ranges rather than exact counts when n < 5.
PLOS One (Verma et al., 2026) suppressed cell sizes under 5; AI workflows should replicate that rule and produce masked tables automatically before dissemination.
Can AI detect culturally specific themes like ceremony or intergenerational trauma in WHiSE 2.0 data?
Yes, when the training and codebook include culturally specific terms and local phrasing.
Because WHiSE 2.0 reports that 71.3% of participants engaged in ceremonies and 76.4% were intergenerational residential school survivors (PLOS One, Verma et al., 2026), adding those phrases as priority codes yields higher fidelity extraction.
Conclusion & Next Steps
The PLOS One WHiSE 2.0 cohort profile (Verma et al., 2026) provides concrete baseline statistics (356 participants, high rates of smoking (92.1%), injection (43.3%), and cultural engagement (71.3%)) that AI-enabled qualitative research can analyze at scale while honoring community governance.
Researchers should use AI to produce reproducible thematic drafts and frequency tables, then complete interpretation with Indigenous advisory partners as the WHiSE 2.0 team modeled in its data governance section of PLOS One (Verma et al., 2026).
To test AI-assisted synthesis on WHiSE-style interviewer data, Try Evidano for free.