Fast, actionable interpretation of mixed-methods trials is the bottleneck for scaling school-based programmes. This post distills the June 30, 2026 PRO-HIIT process evaluation (n=369, 12 weeks) and shows how to run a rigorous process evaluation of HIIT interventions using AI-enabled qualitative research. Key published metrics: mean sessions delivered ≈26.5 vs planned 60, average HR 146 bpm (71% HRmax) and peak HR 175 bpm (85% HRmax), session RPE mean = 5, retention T2 = 100% and T3 = 96% (see original paper: PLOS ONE and trial registration: ClinicalTrials.gov).
Key Takeaways
This post summarises the PRO-HIIT process evaluation (published 30 June 2026) and explains why process evaluation is necessary to interpret school-based HIIT outcomes.
- PRO-HIIT was published 30 June 2026 and is registered as NCT06374732 (registered 18/04/2024).
- The trial planned 60 sessions per class but delivered a mean of ≈26.5 sessions (class counts 26, 25, 24, 31).
- Physiological intensity: average HR 146 bpm (71% HRmax) and mean peak HR 175 bpm (85% HRmax); session RPE mean = 5.
- A school-wide sports event caused cancellations from week 5 and session length adapted from 10 to 6–8 minutes to preserve effort and curriculum fit.
Fast take, why this matters for evaluators
This mixed-methods process evaluation (PRO-HIIT) explained implementation fidelity, adaptations, barriers and facilitators alongside outcome data.
This mixed approach means evaluators and program teams cannot rely on effect sizes alone, they must link qualitative logs, focus-group transcripts, and session-level telemetry to interpret what worked, for whom, and why.
- Study snapshot: Published 30 June 2026; cluster RCT protocol registered NCT06374732 (registered 18/04/2024).
- Core qualitative inputs: teacher training logbooks plus three focus groups (9 students, 2–3 teachers).
- Practical problem: planned dose (60 sessions) vs delivered (mean ≈26.5), contextual interruptions drove the dose shortfall.
Findings snapshot (key numbers)
| Metric | Value | Source / Note |
|---|---|---|
| Publication date | 30 June 2026 | PLOS ONE |
| Sample (intervention) | 369 students (age ~13), 8 classes, 3 PE teachers | Methods section |
| Intervention length | 12 weeks; planned 5 sessions/week | Protocol & methods |
| Planned sessions per class | 60 | Study protocol |
| Mean sessions delivered | ≈26.5 (class counts 26, 25, 24, 31) | Teacher logbooks |
| Average HR | 146 bpm (71% HRmax) | 127 individuals across 13 sessions |
| Mean peak HR | 175 bpm (85% HRmax) | 127 individuals |
| Session RPE | Mean 5 (range 3–8) | 225 individuals across 5 sessions |
| Retention | T2 = 100%; T3 = 96% | Follow-up data |
| Adverse events | No intervention-related injuries | Training logbooks |
What happened: methods & pragmatic deviations
This section summarises PRO-HIIT methods and the pragmatic deviations observed during implementation.
PRO-HIIT embedded 6–10 minute HIIT into PE/PA lessons (March–June 2024) and implementation was evaluated against the MRC process-evaluation framework using 12 measures across three domains: implementation, mechanisms of impact, and context.
- Primary quantitative inputs: training logbooks, HR monitors (Polar Verity Sense for 10 participants/session), OMNI RPE scale.
- Primary qualitative inputs: teacher and student focus groups, transcripts transcribed in Chinese and translated to English for analysis.
- Main deviations: PA lesson sessions were largely cancelled from week 5 due to a school-wide sports event; session length adapted (10 → 6–8 min) to preserve effort and curriculum fit.
So what for researchers, program managers and evaluators
Researchers
Researchers should triangulate session-level dose, objective intensity, subjective intensity and focus-group themes to explain effect sizes.
Triangulation matters: link session-level dose (logbooks), objective intensity (HR), subjective intensity (RPE) and focus-group themes to explain effect sizes.
Design process evaluations up front: register measures (reach, dose delivered, fidelity, adaptations) and plan sampling for intensity monitoring to avoid partial fidelity data.
Program / school teams
Program and school teams should build flexibility into curricula to sustain motivation and reduce teacher workload.
Build flexibility into curricula: allow shorter sessions (6–8 min) and richer movement variety to sustain motivation, especially for girls.
Leverage peer coaches and teacher participation, both were key facilitators in PRO-HIIT and reduced teacher workload.
Evaluation teams & funders
Evaluation teams and funders should budget for measurement capacity and plan for external events.
Budget for measurement capacity: HR monitors and on-site observers are expensive but crucial if you want reliable fidelity metrics across lesson types.
Expect external events (sports days, weather); pre-specify fallback plans and document adaptations in structured logbooks.
Do more, faster with Evidano (map to PRO-HIIT use case)
Ingest and unify mixed inputs
Evidano is an AI-powered qualitative data analysis platform that ingests teacher logbooks (CSV), HR/time-series exports, and focus-group audio/transcripts into a single project.
Import teacher logbooks (CSV), HR/time-series exports, and focus-group audio/transcripts into one project.
Evidano supports transcription (custom dictionaries, PII redaction) and translation, useful when transcripts are in Chinese and you need bilingual coding.
Reproduce the MRC process-eval pipeline at scale
Evidano lets teams reproduce the MRC process-eval pipeline at scale.
Create a codebook for the 12 process measures and import it into Evidano. Use AI-assisted coding to apply themes (implementation, mechanisms, context) and then refine with human review.
Run thematic and frequency analyses and cross-segment comparisons (by sex, class, or week) to detect who received the dose and when fidelity dropped.
Link qualitative themes to quantitative signals
Evidano merges coded qualitative themes with session metadata to quantify the impact of contextual factors.
Merge session metadata (date, delivered minutes, HR averages) with themes like 'weather', 'competing priorities', or 'peer coach' to quantify impact of contextual factors.
Generate co-occurrence networks (for example, 'PA lesson cancellations' co-occurring with 'school event') and export visualizations for stakeholders.
Secure, reproducible, and audit-ready
Evidano keeps data encrypted and produces audit-ready, reproducible reports suitable for ethics submissions and funder audits.
All data is encrypted and never used to train third-party models; Evidano produces reproducible reports you can attach to ethics submissions or funder audits.
Quick workflow: run a PRO-HIIT style process evaluation in Evidano (7 steps)
This seven-step workflow describes how to run a PRO-HIIT style process evaluation in Evidano.
Step 1: Centralize inputs, upload logbooks (CSV), HR exports, and focus-group audio to one project.
Step 2: Auto-transcribe and translate (if needed) using custom dictionaries for study terms and acronyms.
Step 3: Import a predefined MRC-aligned codebook and run AI-assisted coding across transcripts and free-text logs.
Step 4: Map session metadata to coded themes and compute cross-segment frequencies (by class, sex, week).
Step 5: Visualize: dose-delivered timelines, SAAFE compliance heatmaps, co-occurrence networks for barriers and facilitators.
Step 6: Interpret with AI chat over your project, ask targeted questions like 'Which barriers predict missed sessions in week 5? ' and get evidence-linked answers with quotes and timestamps.
Step 7: Export an audit-ready mixed-methods report with linked quotes, charts, and a reproducible analysis log for publication or stakeholder briefings.
FAQ: process evaluation of HIIT interventions
How do I compare fidelity across classes when HR data are sparse?
Use mixed indicators: combine objective HR from the subsample with sessional RPE, attendance, and teacher-rated session quality.
Use mixed indicators: combine objective HR from the subsample with sessional RPE, attendance, and teacher-rated session quality. In Evidano you can weight and visualize combined fidelity indices.
Can AI help identify subtle contextual drivers (e.g., schedule conflicts)?
Yes, AI methods like topic clustering and co-occurrence analysis surface recurring contextual factors that correlate with dose drops.
Topic clustering and co-occurrence analysis on logbook notes and focus-group transcripts surfaces recurring contextual factors like 'sports day' or 'rain' that correlate with dose drops.
Is this approach secure for student data?
Evidano supports PII redaction, encryption, and a no-third-party-model-training policy to meet research ethics expectations.
Evidano supports PII redaction in transcripts, encryption at rest and in transit, and a no-third-party-model-training policy to meet research ethics expectations.
Conclusion, your next two moves
PRO-HIIT (published 30 June 2026) shows how process evaluation explains implementation gaps (dose vs planned), fidelity signals, and practical adaptations that shaped outcomes.
- Try a pilot: import one class of logbooks plus two focus-group transcripts into Evidano and run the 7-step workflow above.
- Try Evidano for free.
Ethics note: this post is research-focused guidance, not clinical advice. For ethics and consent procedures refer to the PRO-HIIT paper and your institutional review board.
Read the original study: PLOS ONE and trial registration: ClinicalTrials.gov.
Topics
- process evaluation HIIT
- PRO-HIIT
- school-based HIIT evaluation
- mixed-methods process evaluation
- implementation fidelity
Keep reading
- Commentary on NewsBetter LLM Coding: Codebook Prompting for LLMsHow codebook prompting for LLMs achieves 82–88% human agreement in PLOS One (Aug 4, 2026), and how researchers can operationalize that pipeline with Evidano.
- Commentary on NewsCommunity Photos to Policy: Photovoice Greenspace AnalysisHow photovoice captures community health in greenspaces and how AI-enabled qualitative analysis speeds synthesis. Learn photovoice greenspace analysis methods and next steps.
- Commentary on NewsFaster Photovoice Qualitative Analysis with AIScale thematic synthesis of photovoice qualitative analysis on greenspaces using AI tools, informed by a PLOS One scoping review protocol (published Aug 4, 2026).