Writing up, publication, authorship and demonstrating impact

• Structure clear reports for QI, audit, evaluation and research using appropriate reporting guidance.
• Apply transparent authorship and contributorship principles.
• Recognise conflicts of interest, research misconduct and selective reporting risks.
• Describe AI-assisted work responsibly where relevant to publication rules.
• Demonstrate project impact for governance, learning, appraisal and dissemination without overstating causality.

A QI report, clinical audit report, service evaluation and research paper have overlapping elements but different purposes and governance histories. State the project type accurately and explain the setting, question, methods, findings, limitations and implications.

For healthcare QI, SQUIRE 2.0 remains a major reporting framework. It asks authors to describe context, rationale, interventions, study of the intervention, measures, analysis, ethical considerations, results, interpretation and limitations. The framework is a guide to transparent reporting, not a licence to force every local project into a journal-style manuscript.

Clinical audit reports should make the standard, denominator, action and re-audit visible. Evaluation reports should remain linked to the service decision. Research manuscripts should follow design-specific reporting guidance from EQUATOR where available.

Good reporting includes what did not work. Document adaptations, abandoned tests, missing data, unintended consequences and contextual factors. Selective reporting produces a misleading account and prevents others from learning.

Separate chronology from causal inference. A change followed by improvement may be plausible evidence, especially with time-series patterns and repeated testing, but do not claim causation beyond the design.

Show the baseline and the trajectory, not just the best result. If performance deteriorated later, include it and explain the sustainability response.

The EQUATOR Network curates reporting guidelines for many study designs. SQUIRE 2.0 is specific to healthcare improvement. CONSORT, STROBE, PRISMA, COREQ and others apply to different designs.

A checklist should be used early, ideally during protocol development, because it exposes information that will need to be collected. Applying it only during final writing may reveal irretrievable gaps.

Do not treat checklist completion as evidence of methodological quality. Reporting guidelines improve transparency; they cannot correct a weak design.

ICMJE recommends four criteria for authorship: substantial contribution to conception/design or acquisition/analysis/interpretation; drafting or critical intellectual revision; final approval; and accountability for the work. All four criteria should be met.

Authorship should be discussed early and revisited as contributions evolve. Routine supervision, funding acquisition or administrative support alone do not automatically justify authorship, although such contributions can be acknowledged.

Avoid gift authorship, ghost authorship and coercive authorship. Contributor statements can make roles transparent and reduce later disputes.

A conflict of interest does not necessarily imply misconduct, but undisclosed relevant relationships can undermine trust. Consider financial, professional, intellectual and organisational interests that could influence design, interpretation or presentation. Follow the target journal or organisation's disclosure requirements.

In local improvement work, a leader who designed the service may also evaluate it. That is not automatically unacceptable, but independent review, transparent methods and explicit acknowledgement of the dual role can strengthen credibility.

Do not allow reputational concerns to suppress legitimate adverse findings.

Fabrication is making up data; falsification is manipulating research materials, processes or data so the record is inaccurate; plagiarism is presenting another person's work or ideas without appropriate attribution. Other questionable practices include selective outcome reporting, inappropriate image manipulation, duplicate publication and undisclosed protocol changes.

Keep source data, analysis decisions and version history according to governance and retention requirements. Correct genuine errors promptly. Serious concerns should follow institutional research-integrity procedures.

A culture that punishes inconvenient findings increases the risk of distorted reporting. Integrity requires that the evidence, not organisational preference, drives the account.

Current ICMJE recommendations state that AI tools should not be listed as authors because they cannot take responsibility or accountability. Authors remain responsible for accuracy, originality, attribution and integrity. Use of AI-assisted technologies should be disclosed as required by the journal, including how they were used.

AI-generated text or analysis can be wrong, incomplete or biased. Human verification is essential. Do not enter confidential patient or research data into systems that are not approved for such use.

For local educational or governance reports, apply the organisation's policy even if no journal submission is planned.

Impact can include patient outcomes, safety, experience, access, equity, efficiency, staff experience, learning, policy change, sustained adoption and spread. Distinguish outputs from outcomes: delivering a teaching session is an output; sustained improvement in a clinical process may be an outcome.

Build an impact chain linking the problem, intervention, evidence, change and consequence. Retain run charts, audit cycles, stakeholder feedback, revised SOPs, meeting decisions, dissemination and evidence of sustainability.

For appraisal or portfolios, reflection should explain the individual's contribution, learning and effect on practice rather than simply attaching a certificate or poster.

A pharmacy-led project reduces omitted doses using iterative PDSA cycles. The publication report describes the baseline process, change theory, sequential tests, run-chart data, balancing measure, context, failed early test, later implementation and 6-month sustainability. Authors are agreed by contribution using ICMJE criteria; a manager who enabled access but did not meet authorship criteria is acknowledged.

The team uses SQUIRE 2.0 as a reporting checklist and explicitly states the local governance route and limitations. The paper claims improvement in the studied setting, not universal effectiveness.

Choose the audiences who need the findings: patients, frontline team, governance committee, executive group, professional network, conference or journal. Decide what each audience needs and what format is proportionate.

Complete the Stage 4 Writing, Authorship and Impact worksheet: project type, reporting guideline, provisional contributor roles, conflict declarations, key figures/tables, limitations, impact evidence and dissemination routes. Review authorship before submission and record any AI-assisted processes according to policy.

• Turning a mixed-result project into a success narrative.
• Using a reporting checklist only after data collection.
• Adding senior names as authors without qualifying contributions.
• Treating AI output as authoritative or listing AI as an author.
• Describing activity or dissemination as impact without evidence of change.

1. EQUATOR Network: SQUIRE 2.0 reporting guideline for healthcare quality improvement. https://www.equator-network.org/reporting-guidelines/squire/
2. International Committee of Medical Journal Editors (ICMJE): Recommendations: authorship, contributorship, conflicts and responsible use of AI. https://www.icmje.org/recommendations/

Educational reference. Verify current local governance requirements before starting a project.