Evidano is an AI-powered qualitative data analysis platform that ingests recordings, transcribes verbatim with custom dictionaries, supports PII redaction, offers AI-assisted coding, links themes to analytics, and exports stakeholder-ready deliverables. Researchers and UX teams evaluating diet apps can learn from the MyIBDDiet pilot protocol published July 2, 2026, which describes a 60-day, single-centre randomized trial (n=40) of an anti-inflammatory diet app for people with IBD, using mixed-methods usability assessment. Read the full protocol in the PLOS One protocol.
Key Takeaways
This post summarizes how to run a rapid, mixed-methods qualitative analysis of an mHealth diet app usability pilot (MyIBDDiet, July 2, 2026; n=40) and how Evidano supports transcription, AI-assisted coding, and linking themes to analytics to accelerate reproducible synthesis.
- MyIBDDiet is a 60-day, single-centre pilot randomized trial (published July 2, 2026) enrolling 40 participants that uses MAUQ, TFA, semi-structured interviews, and app analytics to assess usability.
- Pair validated acceptability scales (MAUQ, TFA) with interview themes and app analytics to distinguish usability problems from contextual engagement issues.
- Evidano ingests audio, performs verbatim transcription with custom dictionaries, supports PII redaction, provides AI-assisted auto-tagging of CFIR/TFA codebooks, and links themes to backend metrics for cross-segment analysis.
- A 7-step Evidano workflow can move a pilot from raw interviews to stakeholder-ready themes in roughly 10–14 days at pilot scale.
Fast take: what the pilot did and why it matters
The MyIBDDiet protocol described a 60-day, single-centre pilot randomized trial (published July 2, 2026) of an anti-inflammatory diet app for people with IBD, enrolling 40 participants and using mixed-methods usability assessment.
Full protocol: PLOS One protocol.
- Primary outcome: usability (quantitative MAUQ plus qualitative interviews).
- Why this matters: co-design and qualitative interview data determine acceptability, implementation barriers, and refinement priorities before a larger RCT.
Findings snapshot (protocol details)
| Metric | Value | Note / Source |
|---|---|---|
| Publication date | July 2, 2026 | PLOS One protocol |
| Design | 60‑day pilot randomized crossover (single centre) | University of Alberta |
| Sample size | n = 40 (20 per arm) | Pilot; no formal power calc |
| Primary outcome | Usability & acceptability (MAUQ, TFA, interviews) | Mixed‑methods |
| Key data types | Interview transcripts, app analytics, ASA24 dietary recalls, biomarkers | Qualitative + quantitative |
| Trial registration | NCT06683105 (registered 8 Nov 2024) | clinicaltrials.gov |
| Recruitment window | Mar 2026 – Mar 2028 | Protocol timeline |
Qualitative analysis of mHealth app usability: Methods in the protocol
The protocol specified a mixed-methods approach: validated usability scales (MAUQ), the Theoretical Framework of Acceptability (TFA), and semi-structured interviews framed by CFIR, with verbatim transcription and analysis until thematic saturation.
- Interview framing: CFIR captured barriers and facilitators to adoption and implementation.
- Coding: two researchers independently coded transcripts with a third reviewer resolving conflicts, using thematic analysis per Braun and Clarke (2006).
- Triangulation: themes were compared with app analytics (usage patterns) and quantitative acceptability scores to prioritise features for iteration.
So what, implications for researchers, UX teams, and clinicians
For trialists / qualitative researchers
Mixed-methods plans are essential: pair MAUQ and TFA scores with interview themes and app analytics to separate usability issues from contextual low engagement.
Pre-define codebook anchors tied to implementation constructs such as burden and perceived effectiveness to speed inter-rater agreement and support variance estimation for later power calculations.
For UX/product teams
Link qualitative themes to measurable product changes and track short pilots to validate fixes like video loading or barcode lookup.
Segment qualitative insights by engagement level (high versus low users) to identify power users and early friction points.
For clinicians and policy teams
Qualitative data reveals acceptability and equity concerns that usage metrics alone do not show, use interview themes to decide whether to integrate apps into usual care.
Objective biomarkers in the protocol (urine sodium/chloride, CRP, fecal calprotectin) provide cross-validation for self-reported dietary changes.
Do more, faster with Evidano (mapped to this protocol)
Problem: Transcripts plus PII → Solution: Secure ingestion and transcription
Evidano ingests recorded interviews, performs verbatim transcription with custom dictionaries, and supports PII redaction before analysts access data, matching the protocol’s privacy needs.
Problem: Manual coding bottleneck → Solution: AI-assisted thematic coding
Evidano lets teams import an initial CFIR/TFA codebook, auto-tag candidate excerpts with AI, and preserve human oversight during adjudication to reduce first-pass coding time.
Problem: Cross-referencing themes with analytics → Solution: Cross-segment analysis
Evidano links themes to app-backend metrics such as engagement and feature use for cross-segment comparisons (for example, high versus low MAUQ scorers).
Problem: Stakeholder reporting → Solution: Clickable quotes and visuals
Evidano generates exportable deliverables including theme hierarchies, quote libraries, co-occurrence networks, and word clouds tailored to clinicians, product teams, or funders.
Data security note
Evidano encrypts data end-to-end and does not use customer data to train third-party models, suitable for clinical research pipelines.
Checklist: 7-step rapid workflow to reproduce the study’s qualitative synthesis in Evidano
Follow these seven steps to move from raw interviews to stakeholder-ready themes in 10–14 days at pilot scale.
- 1) Import audio and metadata into Evidano; enable PII redaction and load a custom diet/IBD dictionary.
- 2) Auto-transcribe; review and correct high-priority segments.
- 3) Upload MAUQ and TFA survey CSVs for cross-linking to participant IDs.
- 4) Import the initial CFIR/TFA codebook and run AI-assisted auto-tagging.
- 5) Human review: adjudicate tags, merge duplicates, and finalise hierarchical themes.
- 6) Run cross-segment analyses (for example, by engagement quartile or disease activity) and generate co-occurrence networks.
- 7) Export a one-page decision brief and a quote library for product sprint planning.
FAQ: qualitative analysis of mHealth app usability
How many interviews are enough?
The protocol answers that interviews are planned until thematic saturation, with pilots often enrolling about 20 per arm (the MyIBDDiet pilot enrolled 40 total).
The MyIBDDiet protocol plans interviews until thematic saturation; Evidano’s coding frequency visualisations help identify saturation points.
Can I combine MAUQ scores with themes?
Yes, you can map numeric MAUQ and TFA scores to participants and run cross-segment thematic frequency and co-occurrence analyses to prioritise fixes.
Map MAUQ/TFA scores to participant IDs and use cross-segment analysis to prioritise usability changes that will have the largest user impact.
Is AI-assisted coding reliable for clinical research?
AI-assisted coding is reliable as a first-pass time-saver but requires human verification for validity; the MyIBDDiet protocol combines both approaches.
Evidano preserves audit trails for reproducibility and supports human adjudication to maintain research validity.
Conclusion, next steps and a quick invite
Adopt a mixed-methods pipeline that links interview themes to usage data and validated acceptability scales for pilots like MyIBDDiet (published July 2, 2026; n=40; 60-day pilot).
Ready to compress weeks of analysis into days? See how Evidano integrates transcription, codebook import, AI-assisted thematic analysis, and cross-segment visualisations to produce reproducible deliverables, or Try Evidano for free.
Ethics note: this guidance is research-focused and non-diagnostic; follow local ethics and consent procedures when handling participant data.