The PLOS One study shows that UK trial staff face a persistent "tightrope" balancing over- and under-reporting of safety events, and that targeted qualitative analysis can convert staff experience into actionable fixes. This post explains how to apply AI-enabled qualitative analysis of safety reporting to the study by Thompson et al., PLOS One (2026), so clinical trials units and operational researchers can prioritise training, clarity, and process change. The primary payoff is faster synthesis of focus-group and transcript data to produce clear, evidence-based decisions for safety reporting policy and SOPs, using reproducible thematic outputs and searchable evidence for audits and regulator dialogue.
Key Takeaways
Thompson et al., PLOS One (2026) found that UK academic trials unit staff described an overarching theme of "Walking on a tightrope: Making justifiable decisions" when enacting risk-proportionate safety reporting, and they recommended clearer guidance, shared learning and role‑specific training. Read the full study in PLOS One.
- Thompson et al., PLOS One (2026) ran four online focus groups between 01-July-2024 and 04-March-2025 with 23 participants from 10 of 52 UK registered CTUs.
- Thompson et al., PLOS One (2026) report a mean participant experience of 13.3 years (SD 9.5) and that phase II (N = 10) and phase III (N = 11) trials were most common among respondents.
- Thompson et al., PLOS One (2026) note regulatory reform came into force on 28-Apr-2026, which the authors expect will reduce duplicative reporting for SUSARs and annual safety reports.
- Key actionable solutions from Thompson et al., PLOS One (2026) were: clearer regulator guidance, CTU-level mentoring and shared templates, and concise, role-specific training materials.
What happened and how the study measured it
Thompson et al., PLOS One (2026) used reflexive thematic analysis of verbatim focus-group transcripts to identify barriers and solutions to safety reporting efficiency in UK academic CTUs.
Thompson et al., PLOS One (2026) recruited 23 participants across four online focus groups conducted between 01-July-2024 and 04-March-2025 and analysed transcripts with NVivo v11.7, as described in their Methods section.
Thompson et al., PLOS One (2026) generated one overarching theme and five subthemes: uncertainty about consequences, bureaucratic red-tape, competing demands, lack of clarity/transparency, and variable knowledge/experience.
The study captured specific quotes that illustrate the decision tension, for example Participant 16 stating, "Generally, you’ve got to balance what you need to collect in terms of safety reporting and in terms of what’s needed to monitor the safety of the drug" (Thompson et al., PLOS One, 2026).
Findings snapshot
| Date | Metric | Value | Implication |
|---|---|---|---|
| 01-July-2024 → 04-March-2025 | Focus groups conducted | 4 online focus groups | Primary data source for thematic analysis (Thompson et al., PLOS One, 2026) |
| Published 30-Jul-2026 | Publication | PLOS One article (Thompson et al., 2026) | Peer-reviewed open access evidence of CTU staff experience |
| Sample dates reported 2024–2025 | Participants | 23 CTU staff across 10 of 52 UK CTUs | Findings reflect academic CTU operational experience (Thompson et al., PLOS One, 2026) |
| Reported participant careers | Mean experience | 13.3 years (SD 9.5) | Responses reflect experienced trial professionals |
| 28-Apr-2026 | Regulatory change | UK Clinical Trials Regulations in force | Reduces duplicative reporting to REC; context for implementation challenges |
Implications for CTU researchers and trial managers
CTU researchers should prioritise building concise, role-specific evidence packages and shared templates because Thompson et al., PLOS One (2026) identify lack of clarity and inconsistent SOPs as a primary barrier to risk-proportionate reporting.
CTU operational leads should adopt mentoring and knowledge‑sharing platforms because Thompson et al., PLOS One (2026) report that participants recommended CTU-level mentoring and shared case studies to increase confidence and reduce over-reporting.
CTU trial managers should align early with Sponsors and sites to agree reporting parameters because Thompson et al., PLOS One (2026) found that competing stakeholder requirements drive duplication and inefficiency.
For audit and regulator dialogue, teams should produce extractable thematic summaries and exemplar decisions because Thompson et al., PLOS One (2026) show that perceived regulatory ambiguity increases risk‑averse behaviour.
How Evidano helps translate qualitative findings into action
What is Evidano?
Evidano is an AI-powered qualitative data analysis platform that helps researchers analyze interviews, open-ended surveys, and documents.
Evidano extracts themes, frequencies and cross-segment patterns from transcripts so CTUs can turn discussion into policy-ready evidence quickly.
Problem: Slow synthesis of focus groups → Solution: Automated thematic analysis
Thompson et al., PLOS One (2026) used Reflexive Thematic Analysis on transcripts, a process that can be time-consuming; Evidano accelerates this by producing reproducible thematic outputs and hierarchical codes from transcript corpora.
Evidano features such as features powered thematic clustering let CTU teams move from raw transcript to actionable recommendations faster while preserving audit trails.
Problem: Training gaps and inconsistent examples → Solution: Searchable case libraries and AI chat
Thompson et al., PLOS One (2026) recommended sharable case studies and mentoring to raise confidence; Evidano supports indexed case libraries and an AI chat over your documents so staff can query past decisions and examples on demand.
Evidano’s AI chat returns verbatim quotes and source attributions to support regulatory discussions and SOP drafting.
Problem: Transcription and redaction overhead → Solution: Secure speech-to-text with PII controls
Thompson et al., PLOS One (2026) recorded and auto-transcribed focus groups before cleaning; Evidano’s speech-to-text supports intelligent verbatim transcription and PII redaction so teams can produce analysis-ready transcripts with less manual work.
Evidano encrypts data and documents provenance for inspections and sponsor reviews; see data security for details.
FAQ: qualitative analysis of safety reporting
What barriers to efficient safety reporting does the PLOS One study identify?
Answer: Thompson et al., PLOS One (2026) identify five subthemes that impede efficient safety reporting: uncertainty about consequences, bureaucratic red-tape, competing demands, lack of clarity/transparency, and variable knowledge and experience.
Thompson et al., PLOS One (2026) support this with data from 23 participants and verbatim quotes illustrating how these barriers push teams toward over-reporting or under-reporting.
Do the April 2026 UK regulatory changes solve these problems?
Answer: The April 28, 2026 UK regulations reduce duplication by requiring SUSARs and annual reports only to the MHRA, but Thompson et al., PLOS One (2026) caution that implementation challenges remain around clarity and training.
Thompson et al., PLOS One (2026) therefore recommend practical resources and CTU-level networks to realise the regulatory intent.
How can CTUs document justifiable decisions for audits and regulators?
Answer: CTUs should capture decision rationales, exemplar case notes and theme-based summaries so decisions are auditable and reproducible, as recommended by Thompson et al., PLOS One (2026).
Using structured templates and searchable transcripts makes it easier to show why an event was judged reportable or not during inspections and sponsor queries.
Can AI help maintain patient-safety while reducing reporting noise?
Answer: Yes, when AI is applied to qualitative and safety data under expert supervision it can flag emerging patterns and surface relevant low-grade toxicities while reducing manual noise triage, an approach aligned with the study’s call for better tools and training (Thompson et al., PLOS One, 2026).
Thompson et al., PLOS One (2026) emphasise that any tooling must be paired with clear guidance and human oversight to avoid under-reporting.
Conclusion & Next Steps
Thompson et al., PLOS One (2026) show that experienced CTU staff balance a difficult trade-off between over- and under-reporting and that clarity, training, and shared examples are required to implement risk-proportionate reporting.
AI-enabled qualitative analysis can convert focus-group transcripts and SOPs into reproducible themes, extractable quotes and audit-ready rationales to support those changes.
If your CTU needs faster synthesis, searchable case libraries and secure transcription to implement the study’s recommendations, Try Evidano for free to pilot thematic analysis and AI chat over your safety reporting documents.