Fast Insights: qualitative analysis of Delphi study

Problem: A 2025 Delphi study (published 1 Sept 2025) probed ethical, medical and legal dilemmas around euthanasia based on advance euthanasia directives (AEDs) in dementia. Payoff: this post shows how to run a reproducible qualitative analysis of Delphi study material to surface themes, consensus signals (7/11 statements reached consensus), and role-specific implications, fast. If you want to apply these steps to transcripts, panel arguments, or coded rounds, Evidano can ingest documents, run thematic + cross-segment analyses and produce visual exports (see www.evidano.com) so teams can move from debate to decisions.

In brief: Source & topline

This post refracts: Seeking consensus on dilemmas related to euthanasia in dementia based on an advance directive: a Delphi study from a medical, ethical and legal perspective (published 1 Sept 2025). Read the original at www.jme.bmj.com/content/51/9/593.

Study window: Delphi rounds run May–July 2023; article published 1 Sept 2025.
Panel: 30 invited, 26 (87%) participated in all three rounds; questionnaires run in Castor EDC.
Output: 11 statements scored; 7 reached consensus (predefined ≥80%).
Contextual fact: AED-based euthanasia remains rare in practice (6 cases in the Netherlands, 2022).

Findings snapshot

Metric	Value	Note
Delphi rounds	3 (May–July 2023)	Structured iterative scoring with controlled feedback
Panellists (start)	30 invited; 31 initially recruited	Purposive sampling across medical, ethical, legal expertise
Retention	26 (87%)	Completed all three rounds
Statements scored	11	7 reached consensus (≥80%)
Key consensus signals	Prioritise current wish to live over AED; consider behavioural expressions; AEDs can support decisions; avoid substitute decision-making	Practical guidance emphasis
Published	1 Sept 2025	Original: www.jme.bmj.com/content/51/9/593

What happened, methods in plain English

The authors used a three-round modified Delphi to gather and refine expert argumentation on AED-based euthanasia in dementia. Panellists rated statements on a 4-point Likert scale and supplied free-text arguments. Responses were pseudonymised and fed back between rounds so experts could revise views.

Consensus threshold: ≥80% agreement (agree/agree more than disagree).
Qualitative side: arguments were coded with MaxQDA and double/triple coded for reliability.
Mixed-methods synthesis: quantitative agreement percentages + thematic argument analysis produced actionable statements.

So what for researchers, clinicians and policy teams

For qualitative researchers (primary keyword: qualitative analysis of Delphi study)

Delphi outputs combine short-form numeric signals (agreement %) with long-form argument threads, treat them as a hybrid corpus: run theme extraction on argument text, then map themes back to statements and agreement scores.

Use segment comparisons (medical vs legal vs ethical) to reveal normative divergences; the study shows legal panellists leaned theoretical while clinicians emphasised pragmatics.

For clinicians and ethics committees

Consensus items (e.g., prioritise current wish to live over AED) are not law but operational signals you can use to shape institutional guidance and case reviews.

Keep a documented decision trail: capture AED text, behavioural observations, consult notes and alternative-treatment attempts, the Delphi underlines the need to justify fulfilment of due care criteria.

For policy analysts and regulators

Use Delphi argument maps to identify legal ambiguities (eg, interpretation of 'mutatis mutandis') and prioritize areas for clarifying guidance, training, or statutory amendment.

The low absolute numbers (6 AED-based euthanasia cases in 2022) suggest policy interventions should focus on practitioner support, not only legislative change.

Do more, faster with Evidano (mapped to this use case)

Problem: Hybrid corpus (Likert scores + arguments) → Solution

Evidano imports PDFs, Castor exports, and free-text from rounds, then aligns numeric scores with argument text to produce a single searchable dataset (themes linked to statements).

Problem: Diverse expert language (Dutch legal/medical terms) → Solution

Use Evidano transcription/translation with a custom dictionary to preserve legal/clinical terms and produce accurate multilingual coding for cross-segment comparison.

Problem: Coding fatigue & inconsistent codebooks → Solution

Import your codebook or let Evidano propose hierarchical themes; run AI-assisted coding, review suggested subcodes, and lock a reproducible code tree for reporting.

Problem: Stakeholder alignment → Solution

Generate an evidence pack: consensus table (statement, % agreement), representative quotes, co-occurrence networks and a short decision memo, exportable to slide decks for ethics committees.

Security & compliance

Evidano uses encrypted storage and proprietary LLMs tuned for qualitative research; your data is not used to train third‑party models. Learn more at www.evidano.com.

Checklist: Reproduce a Delphi-themed synthesis in two weeks

Follow these steps to move from raw Delphi exports to stakeholder-ready insight:

1) Gather inputs: round questionnaires, free-text arguments, agreement scores, codebook drafts, and consultation notes.
2) Ingest into Evidano (PDF/CSV/CSV from Castor) and apply a custom dictionary for domain terms (eg, 'AED', 'mutatis mutandis').
3) Run AI-assisted thematic extraction; review and refine the top 10 themes and map them to each statement.
4) Produce cross-segment comparisons (medical vs legal vs ethical) and flag statements with >15% divergence between groups.
5) Generate visual exports: consensus table, quote set, co-occurrence network, and a 1-page decision memo for committees.
6) Validate with a small stakeholder review (n=3–5) and archive reproducible analysis for audit.

Ethics note

This post is research-methods focused and non-diagnostic. When working with sensitive clinical materials (eg, AEDs, behavioural notes) ensure consent, pseudonymisation, and local legal compliance before analysis.

Wrapping up & next steps

What to do now: if you need to convert a Delphi corpus into a decision-ready brief (themes, argued positions, and segment differences), start by running a reproducible ingestion and theme-mapping workflow, the study above shows that argument text matters as much as agreement percentages.

Want a hands-on trial? Upload one round of your Delphi (or a sample of panel arguments) to Evidano for a free pilot and get a consensus + themes export you can use in committee briefings (see www.evidano.com).
Ethics reminder: respect consent and jurisdictional rules when sharing or publishing sensitive data.

Keep reading

Browse all articles