Publishing Quality Improvement Projects (SQUIRE)

Improvement teams change real systems every week. Turning that work into a paper means showing what changed, why, and how you know the change caused the result.

Guide · 11 min read · Directive Publications Editorial Team

Why improvement work rarely reaches a journal

Hospitals and clinics run quality improvement (QI) projects constantly: a sepsis bundle, a discharge checklist, a new handover routine. Much of that work ends as a slide deck or a poster. When teams do write it up, the draft often reads as a project log ("first we met, then we audited, then we made a poster") and reviewers cannot tell what changed, in what setting, or whether the change caused the result.

Publishing a quality improvement project is worth the effort because other teams can only learn from work they can find and judge. But a QI report is not a standard research paper. The intervention usually changed during the project, the setting matters as much as the intervention, and the evidence is often a run of data over time rather than a comparison of two groups. The reporting guideline built for exactly this is SQUIRE.

This guide is written for the people who lead or support that work: QI leads, clinical governance teams, residency programme directors and research offices.

What SQUIRE 2.0 covers

SQUIRE stands for Standards for QUality Improvement Reporting Excellence. The original version appeared in 2008 with 19 items. The current version, SQUIRE 2.0, was developed through interviews, focus groups, consensus meetings, pilot testing with authors and public comment, and was published in 2015 in several journals at once. It has 18 items grouped by the questions a reader asks:

The authors describe it as intended for reports of system-level work to improve the quality, safety and value of health care, and say it can be adapted to the whole range of improvement methods, from iterative Plan-Do-Study-Act (PDSA) cycles in one unit to multisite randomised trials.

Three practical points matter before you start writing:

• It is not a template. The EQUATOR Network's SQUIRE resources note that the SQUIRE items do not dictate how to order the information. You can reorder items to suit the journal.
• Not every item will apply. Consider each one, and where an item does not apply, say why in the text or in your checklist.
• Extensions exist. Use SQUIRE-EDU (Academic Medicine, 2019) for educational improvement work, such as a change to a residency curriculum, and SQUIRE-SIM (Simulation in Healthcare, published online 2024) for simulation-based improvement projects.

If the improvement was tested in a randomised trial, the trial's reporting guideline applies as well, and ICMJE-following journals will expect the trial to be registered. Our guide to choosing between CONSORT, PRISMA, STROBE and other guidelines covers the other designs. SQUIRE 2.0 is now about a decade old and the SQUIRE team has announced work on an update, so check the SQUIRE and EQUATOR pages for the current version before you submit.

Settle the ethics question before you collect data

This is a common sticking point for QI publications, and one you cannot fix at the writing stage.

Whether a QI project needs ethics review depends on your country, your institution and whether the project counts as research. In the United States, the Office for Human Research Protections states in its guidance on quality improvement activities that the intent to publish is not, on its own, enough to make a QI project research, and that a QI project can involve research even with no plan to publish. Other countries and institutions draw the line differently.

Journals will still ask. SQUIRE item 12 asks how the ethical aspects of implementing and studying the intervention were addressed, "including, but not limited to, formal ethics review and potential conflict(s) of interest." In practice that means you need one of two things on paper:

• an approval from the committee that reviewed the project, or
• a written determination from your institution that formal review was not required, with the reason.

Ask for that determination at the planning stage. A retrospective request after the data are collected is harder to obtain, and a journal may not accept a report without one. Our guide to writing ethics approval statements shows how to word either version.

Two other ethical details reviewers look for in QI work: whether patients or staff were exposed to any risk by the change, and whether anyone on the team had an interest in the outcome, such as a role in buying the equipment or software being tested.

Seven steps to a SQUIRE-ready manuscript

1. Describe the local problem and the rationale

Items 3 to 6 ask for the nature and significance of the local problem, what is already known, your rationale, and your aims. The local problem should be stated with your own baseline data, not only with national figures. SQUIRE's authors note that improvement work has not consistently been grounded in theory, so this item is easy to under-report. SQUIRE asks for the frameworks, models or theories you used and the reasons you expected the change to work. One honest paragraph on why you believed a checklist would reduce missed doses is more useful than a citation to a general improvement model.

Separate the aim of the project from the aim of the report. They are often different: the project aimed to cut delays, while the report may aim to describe why the change worked on one ward and not another.

2. Describe the context and the team

Item 7 asks for the contextual elements that mattered at the outset; item 8b asks for specifics of the team. Readers use these to judge whether your result would transfer to their setting. Describe the type and size of unit, staffing, existing systems, competing initiatives running at the same time, and the roles on the team (for example, a ward nurse, a pharmacist and a data analyst).

If the journal uses double-anonymous review, describe the setting without naming the hospital in the review version. In small units, take care that the description does not identify individual patients or staff.

3. Describe the intervention and how it changed

Item 8a asks for enough detail that others could reproduce the intervention. Item 13a, in the results, asks for its initial steps and evolution over time, including modifications made during the project, and suggests a timeline diagram, flowchart or table.

This is where QI reports differ most from trials. Changes evolve across PDSA cycles, and that evolution is a finding. A table with one row per cycle (what was tested, what was learned, what was changed next) shows reviewers the logic of the project far better than paragraphs of narrative. Put full materials, such as the checklist or order set, in a supplement.

4. Explain how you studied the intervention

Item 9 is the one that separates a publishable report from a project log. It asks for the approach used to assess the impact of the intervention, and the approach used to establish whether the observed outcomes were due to the intervention.

The SQUIRE authors make a clear distinction: doing improvement aims to make care better locally, while studying it aims to add to knowledge about whether and why an intervention works. Your report needs both. State the design plainly: a before-and-after comparison, a time series with a baseline period, a comparison with a unit that did not receive the change, or a staged rollout. Then say what else could explain the change, such as seasonal variation, staffing changes, another initiative or a change in how data were recorded, and how you addressed it.

5. Define measures and check data quality

Item 10 asks for the measures chosen for processes and outcomes, the rationale, their operational definitions, and their validity and reliability. It also asks how you assessed the completeness and accuracy of the data.

Give each measure a definition precise enough for another team to calculate it: numerator, denominator, data source and sampling. Include at least one balancing measure, which could reveal harm or extra burden elsewhere in the system, because item 13e asks you to report unintended consequences, and you can only report what you measured.

6. Analyse variation over time

Item 11 asks for the qualitative and quantitative methods used to draw inferences, and the methods used to understand variation within the data, "including the effects of time as a variable."

QI reports commonly display data on run charts or statistical process control charts. If you use them, plot the baseline period, annotate each change on the chart, and name the rules you used to decide that a change was a real signal rather than noise, with a citation for those rules. If you used qualitative methods, such as staff interviews, describe how they were analysed. Report missing data, as item 13f requires.

7. Discuss costs, limits and spread

The discussion items ask what readers will use to decide whether to try your change. Item 15 includes the impact on people and systems, reasons why observed and expected outcomes differed, and costs and strategic trade-offs, including opportunity costs. Item 16 asks for limits to generalisability and to internal validity. Item 17 asks about sustainability, potential for spread and suggested next steps. Item 18 asks for funding sources and the funder's role.

Be specific about sustainability. If the improvement faded after the project team stopped auditing, say so; that is useful knowledge for other teams.

What reviewers push back on

What institutions can do to make QI publishable

Most of the barriers above are set before anyone starts writing, so institutions can remove them.

• Offer a clear QI-versus-research determination route with a quick turnaround and a written outcome that authors can cite in a manuscript.
• Make baseline data part of project approval. A project without a pre-change baseline is hard to publish whatever the outcome.
• Agree data access early. Confirm that the team can use and report the data it collects, within your data protection rules.
• Provide analytical support. Access to someone who can build and interpret run charts and control charts is often the difference between a poster and a paper.
• Teach SQUIRE to trainees who complete QI projects as part of their programmes, using SQUIRE-EDU where the project changed education rather than care.
• Count QI publications in the same promotion and appraisal processes that count other scholarly work.

Many QI projects first appear as conference abstracts. That is usually compatible with later journal publication, and our guide to turning a conference abstract into a journal article explains how to disclose it.

Before you submit

1. A written ethics approval or determination, dated before data collection where possible
2. Title says the work is an improvement initiative
3. Baseline data and the local problem are stated with your own numbers
4. Rationale for why the change was expected to work
5. Context and team described without identifying patients or staff
6. Intervention changes shown in a timeline or PDSA table
7. Design and alternative explanations stated (item 9)
8. Operational definitions for every measure, including one balancing measure
9. Charts show the baseline, annotated changes and the rules used
10. Unintended consequences, costs, missing data and funding reported
11. Completed SQUIRE checklist, with page numbers and reasons for any items that do not apply
12. Target journal publishes QI work and its article type and word limit checked

Directive Publications publishes its peer review process so you can see how manuscripts are assessed. When your report is ready, you can submit your manuscript directly.