How to Turn a QI Dashboard Into a Publishable Research Paper

A dashboard is not a paper. That’s the first problem.

I’ve seen plenty of students, residents, and early faculty walk into a meeting with a slick quality dashboard and say, “Can we publish this?” Usually the honest answer is: not yet. A dashboard tells you what happened. A publishable paper explains what you did, why you did it, how you measured it, what changed over time, and why anyone outside your building should care.

That gap is fixable.

If you already have QI data, you’re not starting from zero. You’re starting with raw material. The trick is to stop thinking like an operations team showing a monthly scorecard and start thinking like an author building a reproducible, credible, reviewer-proof story.

Here’s how to do it.

Start With the Right Question: Is Your Dashboard a Study or a Report?

Most QI dashboards are internal reports. Useful. Necessary. Not automatically scholarship.

Here’s the difference:

• Internal QI output answers: “How are we doing right now?”
• QI scholarship answers: “What intervention was implemented, how was it evaluated, what changed over time, and what can others learn from it?”

That distinction matters because journals don’t publish dashboards. They publish questions, methods, results, and interpretation.

Step 1: Write the study question in one sentence

Use a PICO- or SMART-style sentence. Keep it tight.

Examples:

• PICO-style: “Among adult emergency department patients with suspected sepsis, did implementation of a nurse-driven lactate protocol increase lactate collection within 60 minutes compared with the pre-intervention period?”
• SMART-style: “We aimed to increase discharge summary completion within 24 hours from 48% to 85% on the general medicine service over 6 months.”

If you can’t write the question in one sentence, your project is still mushy. Fix that before you write anything else.

Step 2: Name the primary aim and primary outcome

Pick one main aim. One. Not five.

Bad:

• Improve throughput
• Reduce length of stay
• Increase patient satisfaction
• Improve staff morale
• Reduce costs

That’s not a paper. That’s wishful thinking.

Better:

• Primary aim: Reduce median time from antibiotic order to administration in febrile neutropenia.
• Primary outcome: Median minutes from order to administration.

Then add secondary measures if they matter:

• Process measure
• Balancing measure
• Sustainability measure

Step 3: Decide why this matters beyond your institution

Reviewers are asking one hard question: Why should anyone else care?

Your answer might be:

• Common clinical problem
• Widely used workflow
• Low-cost intervention
• Easy adaptation at other sites
• Gap in existing literature

If the only reason it matters is “our chair wanted to track it,” that’s operations, not scholarship.

Step 4: Identify the right readership

Know who you’re writing for:

• Clinical audience: wants patient-centered outcomes and clinical relevance
• Operational audience: wants workflow, implementation detail, feasibility
• Academic/QI audience: wants methodology, measurement, rigor, and lessons for spread

This choice shapes your abstract, discussion, and target journal. Get it wrong and the paper feels homeless.

Audit the Dashboard Before You Write: Data, Definitions, and Missing Pieces

This is where many promising projects die. Not because the intervention failed, but because the data are sloppy.

A pretty dashboard can hide ugly methods.

Before you draft a manuscript, do a formal audit of the dataset.

Your dashboard audit checklist

1. Verify the data source

Ask:

• EHR extraction?
• Manual chart review?
• Administrative claims?
• Registry data?
• Dashboard vendor feed?

Write down exactly where the data came from. “Hospital dashboard” is not a data source. That’s a screen.

2. Lock the denominator and inclusion criteria

You need precise definitions:

• Which patients were included?
• Which units?
• What age range?
• Which diagnoses or orders triggered inclusion?
• What exclusions were used?

This is where reviewers smell trouble. If your denominator changes every month because the team kept redefining “eligible patient,” your trend line is junk.

3. Check for missingness and instability

Look for:

• Missing time stamps
• Missing discharge data
• Months with tiny sample sizes
• Denominators swinging wildly
• Different reporting intervals before and after intervention

A common failure: pre period reported monthly, post period reported weekly. Don’t do that. Align intervals.

4. Classify your measures

Every decent QI paper should tell readers whether the dashboard tracks:

• Outcome measures: what ultimately happened
  Example: CLABSI rate, readmission rate, mortality
• Process measures: whether the intended action happened
  Example: % of patients receiving med reconciliation
• Balancing measures: whether you created a new problem
  Example: faster discharges but higher bouncebacks

If you only have a process metric, say so. Don’t dress it up as a patient outcome.

5. Document confounders

This part gets ignored too often. Big mistake.

List anything else that changed during the study period:

• New staffing model
• New EHR build
• Department move
• Policy change
• Parallel initiative
• Seasonal surge
• National guideline shift

I’ve seen teams claim victory on a throughput intervention that overlapped exactly with expanded attending coverage. That’s not causality. That’s a confounded mess.

Convert Operational Metrics Into a Research Story

Now you’ve got cleaner data. Good. Next job: turn it into a story that makes sense.

A publishable QI manuscript needs a clean arc:

1. There was a meaningful problem.
2. You identified a local gap.
3. You implemented a specific intervention.
4. You measured what changed over time.
5. You explain what others can learn from it.

That’s the story. Not “we tracked a metric and it improved.”

Build the background the right way

Keep the introduction focused:

• What is the clinical or operational problem?
• What is already known?
• What local gap existed at your site?
• Why did you choose this intervention?

Example: “Discharge summaries delayed beyond 24 hours impair care transitions and frustrate outpatient clinicians. Our institution’s completion rate lagged well below internal targets. We implemented a resident-facing reminder bundle with attending feedback to improve timely completion.”

That works because it gives:

• Problem
• Consequence
• Local gap
• Intervention rationale

Translate trends into an intervention-outcome narrative

Your dashboard probably shows a line moving up or down. Fine. But a manuscript needs structure.

Use this format:

• Baseline state: What was happening before?
• Intervention design: What exactly changed?
• Implementation phases: When did each piece roll out?
• Observed effect: What happened after each phase?
• Interpretation: Why do you think the change occurred?

If you used multiple PDSA cycles, say what changed in each cycle. Don’t hide the iteration. That’s the point of QI.

Keep the Methods section lean but complete

A strong QI methods section usually covers five things:

1. Setting

   • Academic center or community hospital
   • Unit, clinic, department
   • Relevant staffing/workflow context

2. Sample

   • Who was included
   • Time period
   • Inclusion/exclusion criteria

3. Measures

   • Primary outcome
   • Process measures
   • Balancing measures
   • Operational definitions

4. Intervention

   • What was done
   • By whom
   • When
   • How adherence was supported

5. Analysis

   • Run chart
   • Statistical process control
   • Interrupted time series
   • Pre/post comparison
   • Simple descriptive stats if that’s all the design supports

Don’t overcomplicate weak data with fancy stats. Reviewers can tell when a fragile project is wearing a tuxedo two sizes too big.

Use real QI methodology

This is where a lot of “dashboard papers” become credible.

Include methods such as:

• PDSA cycles
• Run charts
• Control charts
• Annotated timeline of interventions
• Process maps
• Fishbone or root cause analysis, if relevant

These tools show that your project wasn’t random operational drift. It was structured improvement work.

A practical move I recommend: create a one-page “story spine” before writing the manuscript.

Use these headings:

• Problem
• Local context
• Intervention
• Measures
• Timeline
• Main findings
• Limits
• Why it matters

If your team can’t agree on that one page, you’re not ready to draft.

Write the Results So Reviewers Trust Them

Reviewers trust specifics. They do not trust adjectives.

So stop writing things like:

• “Marked improvement”
• “Significant gains”
• “Substantial reduction”

Show the numbers.

Present results in sequence

Use a logical order:

1. Baseline period
2. Intervention phase(s)
3. Post-intervention outcomes
4. Balancing measures
5. Sustainability, if available

That sequence makes the paper easy to follow and harder to misread.

Report absolute numbers and denominators

Always include:

• Numerator
• Denominator
• Rate or percentage
• Time period
• Effect size if applicable

Example: “Timely discharge summary completion increased from 91 of 190 discharges (47.9%) at baseline to 168 of 205 discharges (82.0%) after full implementation.”

That sentence is doing real work. It’s concrete.

Show trends over time, not just pre/post

A single before-and-after snapshot is weak. Time-series visuals are stronger because they show:

• Baseline stability
• Timing of intervention
• Immediate shift vs gradual change
• Sustainability
• Noise

If the metric was already improving before your intervention, that matters. If it fell back to baseline two months later, that also matters.

Be direct about limitations

A weak discussion dodges flaws. A good discussion names them early and plainly.

Common limitations to acknowledge:

• Single-site design
• Small sample size
• Missing data
• Nonrandomized implementation
• Secular trends
• Documentation bias
• Simultaneous workflow changes
• Limited follow-up

Don’t panic about this. Every QI project has limitations. Reviewers aren’t offended by imperfection. They’re offended by denial.

A good line sounds like this: “Because the intervention was introduced during a broader staffing redesign, we cannot attribute the observed improvement solely to the intervention bundle.”

That honesty increases trust. Overclaiming kills papers fast.

Package It for Submission: Journal Fit, Authorship, and Revision Strategy

A decent paper can still get rejected if you send it to the wrong place or organize the team badly.

This part is logistics. Boring, maybe. Still crucial.

Choose the right journal lane

Match your manuscript to what it actually is:

• QI journals: best for practical improvement projects with strong implementation detail
• Implementation science journals: better if the paper focuses on adoption, fidelity, context, or scale-up
• Clinical operations journals: useful for workflow, safety, throughput, and systems redesign
• Specialty journals: best if the project addresses a specialty-specific problem readers know well

Don’t send a narrow internal throughput paper to a top-tier general medical journal and act shocked when it gets bounced in 48 hours. That’s not ambition. That’s poor targeting.

Use SQUIRE 2.0

If your paper is QI, use SQUIRE 2.0 as your reporting backbone.

It helps you cover:

• Rationale
• Context
• Intervention
• Measures
• Analysis
• Ethical considerations
• Results
• Interpretation

You don’t need to worship the checklist. But you do need to use it. It keeps you from forgetting the exact details reviewers expect.

Set authorship early

Do this before the paper is drafted, not after acceptance when everyone suddenly remembers they “helped.”

Clarify:

• First author
• Senior author
• Statistician/methods support
• Data extraction contributor
• Operational leader
• Chart reviewer
• Writing roles

Also confirm:

• Institutional approvals
• QI vs IRB determination
• Conflict disclosures
• Contributor roles

Nothing derails momentum like an authorship fight in version 7.

Plan for reviewer objections before they arrive

Expect questions about:

• Generalizability
• Statistical approach
• Sustainability
• Fidelity to the intervention
• Confounding
• Whether the outcome is clinically meaningful

Write your discussion to answer those concerns in advance.

One practical trick: before submission, ask a colleague outside the project to review the paper and answer three questions:

1. What exactly was the intervention?
2. What is the strongest evidence that it worked?
3. What would make you doubt the conclusion?

If they can’t answer #1 or #2 clearly, your manuscript still has holes.

Common Failure Points and How to Fix Them Fast

Here are the usual reasons a dashboard never becomes a publishable paper. I’ve seen every one of these.

Failure point 1: The aim is vague

Problem: “We wanted to improve care coordination.”

Fix: Rewrite as a measurable aim with a defined population, metric, and timeframe.

Failure point 2: The denominator is unclear

Problem: Nobody knows who counted as eligible.

Fix: Create a one-line operational definition and apply it consistently across all periods.

Failure point 3: The paper overclaims causality

Problem: “The intervention caused the reduction.”

Fix: Use honest language:

• “was associated with”
• “coincided with”
• “followed implementation of”

Unless your design truly supports causal inference, don’t pretend.

Failure point 4: The discussion is thin

Problem: Results are dumped on the page with no context.

Fix: Add three things:

• How findings compare with prior work
• What parts are transferable to other settings
• What the next cycle or next study should test

Failure point 5: The visual display is weak

Problem: One pre/post table and no timeline.

Fix: Add:

• Run chart or control chart
• Annotated intervention timeline
• Clear denominators in figure legends

Pre-submission checklist

Before you hit submit, confirm that you have:

• A one-sentence study question
• One primary aim and one primary outcome
• Clear inclusion criteria and denominator
• Defined intervention timeline
• Outcome, process, and balancing measures if available
• Time-based visual display of results
• Honest limitation statements
• SQUIRE-aligned structure
• Journal fit confirmed
• Authorship and approvals finalized

Summary

Here’s the bottom line: a QI dashboard becomes publishable when you stop treating it like a scorecard and start treating it like a study.

The fastest route is straightforward:

1. Ask a real question.
2. Audit the data brutally.
3. Define the intervention and measures clearly.
4. Show trends over time with honest numbers.
5. Write limitations like an adult.
6. Submit to a journal that actually wants this kind of work.

Clean definitions beat flashy graphics. Honest interpretation beats overclaiming. Strong visuals, reproducible methods, and the right framing do most of the heavy lifting.

If your dashboard is messy, fix the dataset first. If your story is vague, fix the aim. If the paper feels thin, you probably haven’t explained context, denominators, or confounders well enough.

That’s the work. Do it well, and a local QI project can absolutely become a paper worth publishing.