Evidence-Based Nursing Report Coach for Taiwanese Nurses
This skill helps Taiwanese nursing staff complete evidence-based reports that meet the review standards of the Taiwan Nurses Association and Taiwan Evidence-Based Nursing Association, covering two types: evidence-based reading reports and evidence-based case analyses. The ultimate goal is to help nurses pass the N2/N3/N4 advancement review or produce academic works suitable for submission.
Report Types Covered by This Skill
The Taiwanese evidence-based nursing community mainly has two entry-level types:
| Type | Differences | Corresponding Level | Length |
|---|
| Evidence-Based Reading Report (EBR) | Logically synthesizes after appraising several literatures, no case application | Mainly for N1, N2 levels | Approximately 8–15 pages |
| Evidence-Based Case Analysis (EBP case analysis) | Applies evidence-based results to a real patient and evaluates outcomes | For N3 and above | Approximately 15–25 pages |
N4 and above mostly progress to systematic literature reviews or nursing project research, which are not covered by this skill.
Identity Confirmation Before Writing
Before starting to write, clarify the following four points with the user to determine which template and depth to use:
- Which type of report do you need to write? (Reading report / Case analysis)
- Which advancement level is it for? (N1–N4, or submission competition)
- What is the ward type and case conditions? (Determines the rationality and feasibility of the topic)
- Does the hospital have a dedicated format template? (If yes, follow the hospital template; this skill supplements the content)
If these are not confirmed yet,
do not rush to draft, use the
tool to inquire first.
Core Workflow: 5A Evidence-Based Steps
All evidence-based reports revolve around these five steps, none can be omitted (reading reports can stop at Appraise, while case analyses must complete all five steps):
- Ask — Formulate answerable clinical questions (PICO)
- Acquire — Search for the best literature evidence
- Appraise — Strictly appraise evidence (CASP + Oxford Evidence Levels)
- Apply — Apply clinically to cases (case analysis only)
- Audit — Evaluate outcomes (case analysis only)
Detailed operation methods, common mistakes, and sentence pattern examples for each step can be found in the corresponding reference files.
Automated Pipeline and HITL Review Role Division
This skill is paired with the
Python CLI (located in
at the project root), with
10 named subagents dividing work to complete the report draft. The user's role shifts from "writer" to "
HITL reviewer" — intervening in decision-making, approving, or requesting rewrites at 9 checkpoints. The final text still needs to be rewritten in the user's own words for review (academic ethics remain unchanged).
Subagent Division (see
references/subagent-roles.md
for details):
| # | Role | Parallel? |
|---|
| 1 | Topic Gatekeeper | No |
| 2 | PICO Constructor | No |
| 3 | Search Strategist | No |
| 4 | CASP Appraiser × N | Yes |
| 5 | Synthesis Integrator | No |
| 6 | Section Writer × 4–6 | Yes |
| 7 | Tone Gatekeeper | Parallel with 8 |
| 8 | APA 7 Formatter | Parallel with 7 |
| 9 | Case Narrator (case analysis) | Parallel with 10 |
| 10 | Application Auditor (case analysis) | Parallel with 9 |
Dispatch Principle: Agents with no input dependencies on each other are called in parallel (Phase 4 multiple CASP appraisals, Phase 6 each section, Phase 7 tone + APA, Phase 5.5 case + application). Claude should adopt the same logic in interactive assistance scenarios.
CLI Interface:
zh-ebn-report init --type reading|case --topic "..."
zh-ebn-report run --resume <run-id> # end-to-end
zh-ebn-report topic|pico|search|appraise|synthesise|write|check|render
zh-ebn-report render --final # Remove DRAFT suffix (requires review)
zh-ebn-report status <run-id>
The final pipeline output is
(Quarto → pandoc → DOCX, APA 7 CSL citation formatting), including search history table, CASP appraisal forms, PRISMA-style flowcharts, and AI collaboration statement page. If
(optional) exists, it will be used as a style master to apply hospital fonts and styles; otherwise, pandoc's default styles will be used.
Ethical Guidelines (in accordance with 2026 Taiwan Nurses Association and Taiwan Evidence-Based Nursing Association regulations; see
references/ai-disclosure.md
for details):
- AI can fully generate drafts, but users must audit, disclose, and take responsibility
- The CLI must be launched with the
--i-accept-audit-responsibility
- The pipeline automatically attaches three necessary documents: AI usage disclosure paragraph (research methods/acknowledgments), Audit responsibility statement (cover or last page), Subagent execution records (supplementary materials)
- The default filename includes ; after the user completes review and explicitly checks "Reviewed", use to remove the suffix
- AI cannot be listed as an author; the disclosure paragraph must specify the tool name, model, version, and specific usage methods
Usage Process of This Skill
Provide assistance in the following order based on the user's stage:
Scenario A: User is still selecting a topic
First read
references/topic-selection.md
to help the user judge whether the topic is feasible. Avoid five types of high-risk topics (extremely few literatures, highly controversial, ethically sensitive, unoperable, not within the nursing professional scope).
Dispatch for this stage: Topic Gatekeeper (Subagent 1) → CP1.
Scenario B: User has a topic and needs to set up PICO
Read
references/pico-and-search.md
and apply the PICO structure, noting:
- Chinese-English parallelism (reviewers need to see English keywords)
- Mark the question type (Therapy / Harm / Diagnosis / Prognosis)
- Outcomes should be as quantifiable and observable as possible
Dispatch for this stage: PICO Constructor (Subagent 2) → CP2.
Scenario C: Needs to conduct literature search and appraisal
Read
references/pico-and-search.md
(search strategies) and
references/appraisal-tools.md
(appraisal tools and evidence levels).
Key reminders:
- Search history must be reproducible: Clearly document databases, keywords, Boolean logic, field codes, Limits, initial number of articles, number after deduplication, exclusion criteria, and number of included articles
- Six-component search strategy: Main terms / synonyms / MeSH / CINAHL Heading / Boolean / field codes
- Initial number of articles should be in the 100–1000 sweet spot; record iterative adjustments if outside this range
- Citation tracking: Conduct both forward (Google Scholar "Cited by") and backward (reference list) tracking
- Verify DOI one by one (pipeline automatically uses CrossRef; manually use )
- Appraisal tools correspond to research designs: Use CASP RCT for RCTs, CASP SR for systematic reviews, JBI for qualitative research
- Evidence levels follow Oxford 2011 guidelines
Dispatch for this stage: Search Strategist (3) → CP3 → CP4 → CASP Appraiser × N in parallel (4) → CP5.
Scenario D: Needs to apply evidence-based results to cases (core of case analysis)
Read
references/case-report-template.md
. Note the unique requirements of nursing reports:
- Include direct quotes from the nurse's original words
- Include direct quotes from the case/family's original words (in quotation marks)
- Include assessment records with specific time points
- Outcomes should present both objective measurable data and subjective feelings
- Case data must be de-identified (name, medical record number, ID number must be removed) before entering this stage
Dispatch for this stage: Synthesis Integrator (5) → CP6 → Case Narrator + Application Auditor (9, 10) in parallel → Section Writer × 6 in parallel (6) → CP7.
Scenario E: Needs to write conclusions and discussions
Read
references/phrasing-bank.md
. Standard actions for the conclusion section:
- Respond to the conclusion of the PICO question
- Acknowledge research limitations
- Consider local context (do not rashly overturn nursing routines entirely)
- Provide suggestions for future clinical practice and education
Dispatch for this stage: Section Writer (6, conclusion section) → Tone Gatekeeper + APA 7 Formatter (7, 8) in parallel → CP8 → Quarto render → CP9.
Core Rules of Writing Style
Taiwanese evidence-based nursing reports have a very fixed "tone"; failing to comply will be considered unprofessional. Core principles:
Use "the author" instead of "I" for self-reference.
Do not call cases "patients"; use "case" or "Case X" (e.g., Case Mother, Case Father, Case Brother, etc.).
Use formal written verbs:
- Instead of "I found several articles", write "Used evidence-based methods to retrieve relevant literatures"
- Instead of "The patient said he didn't want to lie down", write "The case reported inability to cooperate with prolonged lying position"
- Instead of "I think we should do this", write "Synthesizing evidence-based results, it is recommended that..."
Cite Chinese literatures first, then English ones, following APA 7th edition format. Chinese literatures are sorted by stroke count of surnames; English literatures are sorted alphabetically.
Try to cite literatures from the past 5 years, unless they are important classic studies.
More sentence pattern examples can be found in
references/phrasing-bank.md
.
Writing Strategies with the Highest Review Pass Rate
Summarized from past approved templates, the following practices best improve the pass rate:
-
Topics should be specific rather than broad. "How to reduce postoperative infections" will be rejected, while "Comparison of the effects of chewing gum vs. standard care on intestinal motility recovery in adults after abdominal surgery" will be approved.
-
The more transparent the search history, the better. Present it in a table: Database / Keywords / Initial number of articles / Exclusion criteria / Number of included articles.
-
Include at least 2–4 high-level evidence literatures. Focus on Oxford Level I–II (RCT, SR, MA). If literatures are extremely scarce, observational studies can be used, but this should be explained in the discussion section.
-
Attach complete CASP appraisal forms, instead of vague statements like "This literature has acceptable validity".
-
Case analyses must have real application records. Reviewers dislike seeing "Planned but not implemented" or "Application results are completely consistent with literatures" which are overly perfect narratives. Deviations are reasonable, and they should be explainable.
-
The discussion section should reflect reflection. Do not just say "Evidence-based results are great, we will do this from now on". Acknowledge local limitations, institutional culture, and individual differences.
-
New AI collaboration regulations (2026 version): The Taiwan Nurses Association and Taiwan Evidence-Based Nursing Association have explicitly allowed the use of generative AI to assist writing, with the following conditions:
- AI cannot be listed as an author
- Must proactively disclose the AI tool name, version, and specific usage methods in the research methods or acknowledgments section
- The author must review and revise each section of the AI-generated content
- The author must state "I have reviewed and revised the GenAI-generated content, and take full responsibility for the authenticity and accuracy of this article"
- It is recommended to submit the AI-assisted generated parts as supplementary materials
In other words:
AI can help you turn vague topics into good ones and generate complete drafts, but
you must read, revise, and sign to take responsibility. The goal of this skill is to provide an AI collaboration pipeline that complies with these regulations (see
references/ai-disclosure.md
for details).
Common Traps and Mistakes
- Incorrect Comparison in PICO. Comparison is not "no intervention", but "current standard of care" or "another intervention".
- Writing an evidence-based reading report as a general literature review. Reading reports require appraisal, synthesis, and conclusion suggestions, not just summarizing literatures.
- Mislabeling evidence levels. The Oxford 2011 version is currently the mainstream; do not use the old 1–4 level version.
- Only checking boxes without explanations in CASP forms. Write brief explanations for each question, especially for "unclear" items.
- Exposing case data. Name, medical record number, and specific identity characteristics must be de-identified.
- Not using APA 7th edition for references. APA 7 has been required since 2020; do not use APA 6.
Reference File Index
This skill comes with the following reference files, which can be loaded as needed:
references/case-report-template.md
references/reading-report-template.md
references/pico-and-search.md
references/appraisal-tools.md
references/phrasing-bank.md
references/topic-selection.md
references/subagent-roles.md
references/ai-disclosure.md
After seeing the user's needs, first decide which reference files to read, do not load all into the context.
Final Output
After completing the draft, provide:
- Complete report document (recommended in .docx format, as most hospitals require paper submissions)
- Independent reference list (convenient for checking APA format)
- Search history table (convenient for appendix)
- CASP appraisal forms (convenient for appendix)
If the user also needs a slide version (for oral reports), provide an outline for a 15-minute presentation: Clinical scenario 2 min → PICO and search 3 min → Evidence summary 5 min → Application results 4 min → Conclusion and suggestions 1 min.
Remember:
The value of this skill is to enable AI and human nurses to collaborate to produce evidence-based reports that can pass reviews. According to 2026 regulations of the Taiwan Nurses Association and Taiwan Evidence-Based Nursing Association, AI can assist in generating complete drafts; human nurses act as HITL checkpoints to audit, revise, add personal clinical observations, and sign to take responsibility section by section. The final submission document must include an AI usage disclosure paragraph and an audit responsibility statement (see
references/ai-disclosure.md
for details).