AI-assisted Qualitative Analysis: Delphi Study on PT Navigator

Evidano is an AI-powered qualitative data analysis platform that helps researchers analyze interviews, open-ended surveys, and documents. The primary keyword for this post is "qualitative analysis of Delphi study" and this article explains how AI-enabled qualitative research can extract consensus-ready insights from a modified Delphi about a Physiotherapist (PT) Navigator role. According to PLOS One (Edward et al., published August 7, 2026), the study used two online rounds and a consensus meeting to define role domains, producing concrete counts and consensus metrics that are ideal inputs for AI-assisted thematic synthesis.

Key Takeaways

According to PLOS One, a two-round modified Delphi (Round 1 n = 34; Round 2 n = 18) and a June 2025 consensus meeting established a consensus PT Navigator role for acute cancer care in Canada, providing both qualitative themes and quantitative thresholds that AI can rapidly structure and visualize.

34 clinicians completed Round 1 between January 8 and March 19, 2025, and 18 completed Round 2 between April 2 and April 16, 2025, according to PLOS One (published August 7, 2026).
Participants rated perceived usefulness of the proposed role as mean 9.19/10 (SD 2.12) in Round 1, according to PLOS One (Edward et al., 2026).
Consensus in Round 2 was defined as ≥75% agreement (mean ≥5.25/7) and every role item met that threshold, according to PLOS One (Edward et al., 2026).
Qualitative feedback prioritized more objective testing and longer appointment times, changes that were incorporated before the consensus meeting in June 2025, according to PLOS One (Edward et al., 2026).

What happened and how the Delphi was run

A two-round, modified reactive Delphi engaged Canadian clinicians to rate and refine a PT Navigator role, according to PLOS One (Edward et al., published August 7, 2026).

According to PLOS One, Round 1 (open Jan 8 to Mar 19, 2025) collected qualitative open-text feedback from 34 respondents and two utility/clarity ratings; Round 2 (Apr 2 to Apr 16, 2025) invited the 19 consenting Round 1 participants and 18 completed the round, providing quantitative ratings on a 7-point Likert scale.

According to PLOS One, the team used duplicate content analysis of open responses to extract themes, descriptive statistics (means, SDs) to evaluate agreement, and a June 2025 consensus meeting (n = 7 attendees) to finalize the role before publishing on August 7, 2026.

According to Edward et al. in PLOS One (2026), key qualitative additions included more objective testing (suggested by 29.4% of Round 1 respondents) and process changes such as longer initial appointment times and patient self-referral options.

Findings snapshot table

Date	Metric	Value	Implication
Jan 8 – Mar 19, 2025	Round 1 participants	n = 34	Generated qualitative themes and two rating questions (usefulness, figure clarity)
Apr 2 – Apr 16, 2025	Round 2 participants	n = 18 (of 19 eligible; 94.7% completion)	Quantified agreement on role items; all items reached consensus (≥75%)
June 2025	Consensus meeting attendees	n = 7 (including 6 PTs, 1 oncologist, 1 patient advocate noted)	Finalized role wording and emphasized self-management over hands-on treatment
Aug 7, 2026	Publication	PLOS One (Edward et al., 2026)	Peer-reviewed record for the proposed PT Navigator role and methods

Implications for researchers and implementation teams

Researchers should treat Delphi open-text feedback as structured qualitative data that benefits from rapid thematic coding plus cross-tabulation by profession and region, according to PLOS One (Edward et al., 2026).

Implementation teams should prioritize the three changes endorsed in the study: longer initial appointment times, increased objective testing, and patient self-referral pathways, according to PLOS One (Edward et al., 2026).

Design trials with mixed outcomes: patient-reported outcomes, objective physical tests, and service metrics as recommended in PLOS One (Edward et al., 2026).
Plan for variable follow-up timings across treatment modalities (surgery vs chemotherapy), a concern raised in the June 2025 consensus meeting, according to PLOS One.

How Evidano Helps: mapping Delphi pain points to AI features

Problem: slow manual coding of open-text Delphi feedback → Solution: automated thematic synthesis

Evidano is an AI-powered qualitative data analysis platform that helps researchers analyze interviews, open-ended surveys, and documents.

Researchers running Delphi studies like the one reported in PLOS One (Edward et al., 2026) can upload Round 1 open-text responses and get duplicate-checked thematic codes, frequency counts, and co-occurrence networks to prioritize items that meet a 10% suggestion threshold or other a priori rules.

Relevant feature: use Evidano features to produce thematic maps and export codebooks for Round 2 survey design.

Problem: inconsistent transcripts and PII risk → Solution: accurate transcription and redaction

Evidano is an AI-powered qualitative data analysis platform that helps researchers analyze interviews, open-ended surveys, and documents.

Teams using audio or video for consensus meetings can use Evidano Speech-to-Text with custom dictionaries and PII redaction to create reproducible transcripts for duplicate coding, matching the duplicate content analysis method used in the PLOS One study.

Problem: integrating qualitative themes with numeric consensus → Solution: cross-segment and frequency analyses

Evidano is an AI-powered qualitative data analysis platform that helps researchers analyze interviews, open-ended surveys, and documents.

Evidano can produce frequency tables, cross-segment comparisons (for example PTs vs nurses), and export publication-ready tables so researchers can report exact counts (for example n = 34, n = 18) and statistics exactly as PLOS One did.

FAQ: qualitative analysis of Delphi study

How can AI speed up qualitative analysis in a Delphi study?

AI can accelerate coding, synthesis, and cross-segment comparisons in Delphi open-text data by automating initial theme generation and frequency counts.

According to PLOS One (Edward et al., 2026), the study team used duplicate human coding for Round 1 open responses; AI-assisted coding can reproduce that duplicate check quickly and highlight items suggested by ≥10% of respondents as candidate additions for Round 2.

What numeric thresholds should I plan for when reporting consensus?

Define consensus a priori, for example the PLOS One study used ≥75% agreement and set mean ≥5.25/7 on a 7-point Likert scale as consensus.

According to PLOS One (Edward et al., 2026), using explicit thresholds (75%, 10% for new item inclusion) lets AI flag which items meet or fail consensus automatically.

Can AI help combine objective tests and patient-reported outcomes in a role definition?

Yes, AI can link qualitative themes to quantitative measures by tagging mentions of objective tests and extracting counts for co-occurrence with outcome terms.

According to PLOS One (Edward et al., 2026), participants requested "more objective testing" (suggested by 29.4% in Round 1), and AI tools can summarize which specific tests are named and how often they co-occur with triage decisions.

How should I preserve trust and ethics while using AI on clinical Delphi data?

Use encrypted platforms that do not share data with external model trainers and apply PII redaction before automated analysis.

Evidano keeps data encrypted and does not use customer data to train third-party models, and teams should also obtain ethics approval and participant consent similar to the PLOS One study's Hamilton Integrated Research Ethics Board approval (ID: 18218).

Conclusion & Next Steps

The PLOS One modified Delphi (published August 7, 2026) produced a consensus PT Navigator role with clear counts and qualitative priorities that are ideal for AI-enabled synthesis.

Using AI-assisted thematic analysis can shorten the Round 1 → Round 2 cycle, produce reproducible codebooks, and generate the frequency and cross-segment tables necessary for pilot RCT planning, consistent with the recommendations in PLOS One (Edward et al., 2026).

If you run Delphi studies, combine structured thresholds (for example 10% addition rule, 75% consensus) with AI pipelines to speed iteration and preserve audit trails.

Ready to operationalize Delphi qualitative data at scale? Try Evidano for free.

Topics

qualitative analysis of Delphi study
Delphi study qualitative analysis
PT Navigator qualitative analysis
AI-enabled qualitative research

Keep reading

Browse all articles

AI-assisted Qualitative Analysis: Delphi Study on PT Navigator

Key Takeaways

What happened and how the Delphi was run

Findings snapshot table

Implications for researchers and implementation teams

How Evidano Helps: mapping Delphi pain points to AI features

Problem: slow manual coding of open-text Delphi feedback → Solution: automated thematic synthesis

Problem: inconsistent transcripts and PII risk → Solution: accurate transcription and redaction

Problem: integrating qualitative themes with numeric consensus → Solution: cross-segment and frequency analyses

FAQ: qualitative analysis of Delphi study

How can AI speed up qualitative analysis in a Delphi study?

What numeric thresholds should I plan for when reporting consensus?

Can AI help combine objective tests and patient-reported outcomes in a role definition?

How should I preserve trust and ethics while using AI on clinical Delphi data?

Conclusion & Next Steps

Topics

Keep reading

Company

About

Preview Evidano

Account

Newsletter