QI Paper Workbench

YOUR LEARNING JOURNEY

Your next project starts here.

0 of 5 stations completePractice → feedback → progress

YOUR NEXT MOVE

Pick your paper

Choose a real QI report to work through.

Explore in any order. Points reward practice.

THE WORKBENCH

Five stations. One project blueprint.

Pick your paper Not started

Each one is a real QI report. The squares show how easy the design is to copy. Open the paper in a new tab and keep it beside you.

Sort the paper into SQUIRE Not started

Tap a card, then tap the SQUIRE item it belongs to. On a computer you can also drag. Tap a placed card to send it back.

Pick a paper to start sorting. Tap one here or scroll up to station 1.

Run chart lab Not started

A guided, hands-on lesson in reading a QI run chart, spotting change over time, and explaining what the data can and cannot prove.

YOUR CLINICAL CASE ILLUSTRATIVE DATA

Can we improve oxygen prescription documentation?

A ward team checks the percentage of oxygen orders completed correctly each week. They introduce a standardized prescription prompt during a PDSA cycle. Can we see whether the process changes?

Each dot = 1 week24 weeksHigher compliance is better

FOLLOW THE LESSON

Read it. Change it. Explain it.

Lesson 1 of 6

STEP 01

First, orient yourself to the chart

TRY THIS

LIVE EXPERIMENT

Weekly oxygen prescription compliance

Simulated example. Baseline compared with the period after a change.

Weekly compliance Baseline median PDSA begins Shift points

Select a dot on the graph to inspect an individual week.

Baseline median...middle of pre-change weeks

Latest week...week 24 compliance

Post-change median...descriptive, not causal

No signal yet.

Read the interpretation here after moving a slider.

See the actual 24 weekly values Useful for teaching and verification

Week	Compliance	Phase	Compared with baseline median

EXPERIMENT CONTROLS

Change one thing at a time

Keep other sliders fixed to see what each variable does.

1. Baseline compliance 40%

How often the ward gets the process right before the PDSA starts. This is the starting level.

Example: 40% means about 4 in 10 prescriptions are correct.

2. Improvement after change +0 pp

The model's added improvement once the change takes effect. pp means percentage points, not a relative percent increase.

Example: 40% + 30 percentage points = 70%.

3. When PDSA starts week 13

The week when the team tests its first change. Weeks to the left are baseline observations.

We keep at least 10 pre-change points in this teaching example.

4. Week-to-week variation up to ±6 pp

Routine fluctuations around the underlying level. More variation makes a small improvement harder to recognize.

Random deviations are bounded by this number, not guaranteed to reach it.

Every dot is simulated. Re-running changes random variation, not the assumed improvement.

PRESENTER NOTES

How to teach this step

Use the script while the students watch the chart.

SAY THIS

ASK THE ROOM

THREE PRACTICE EXPERIMENTS

Earn your Signal Hunter badge

Try the experiments in order. The tracker recognizes them automatically as you change the chart.

0 of 3 complete

01. Make a shift appear
Increase the modeled improvement until at least 8 post-change dots stay above the baseline median.
02. Make the shift disappear
Reduce the improvement from a setting that showed a shift. Watch what happens to the evidence.
03. Recover a shift despite variation
Set variation to at least 12 points, then increase improvement until you see a shift again.

CHECK YOUR UNDERSTANDING

What can you conclude from a shift?

Award yourself the right interpretation before you present a results slide.

AUTO-GENERATED RESULTS LANGUAGE

Quick teaching reference: every term explained

Run chart

A measure plotted in time order. It helps reveal variation, possible shifts, and whether improvement lasts.

Baseline

The observations collected before a change is tested. This is the comparator, not the ultimate goal.

Median

The middle observation after sorting values. This lab fixes the pre-change median as a reference line.

PDSA cycle

Plan, Do, Study, Act: a structured small test of a proposed change, followed by learning and adaptation.

Shift

A string of observations on the same side of the median. We use an 8-point convention for this 24-point demonstration.

Common-cause variation

Normal variation within a process. A pattern may suggest another influence, but the graph alone cannot identify it.

Percentage point (pp)

The absolute difference between percentages. Moving from 40% to 70% equals +30 pp, or a 75% relative increase.

Run chart versus control chart

A run chart compares values over time with a median. A control chart also uses statistically calculated limits to analyze process variation.

Method and limitations. This is a simulated 24-week dataset, not a result from any of the papers above. For a 24-point chart, this lesson uses the 8-point shift convention from the IHI run-chart reference sheet. Other teaching resources use 6 points and guidance varies with dataset length. Points exactly on the median are skipped when counting a run. This lesson checks the shift rule only; it does not test trends, total numbers of runs, or unusual points. Signal detection is not proof of causation. In real projects, document denominators, measurement definitions, changes in patient mix and other concurrent events. IHI run chart rules · AAP clinical QI interpretation guide.

Build your own project Not started

Now it is your turn. Write a SMART aim, run it through the 3M filter, then sort the measures for an example project.

Direction

What you will measure

Where

Weeks to reach target

Baseline today

Target

The 3M filter

Switch each one on only if it is honestly true for your project.

Sort the measures

Example project: oxygen prescribing on a medical ward. Which kind of measure is each one? Editors always look for a balancing measure.

Now write your own measures

Optional, but it earns bonus points and goes into your blueprint. One of each.

Outcome measure

Process measure

Balancing measure

Where would it go? Not started

Choosing the wrong journal is the fastest way to a desk rejection. Decide where each project belongs.

Your project blueprint

All five stations done. Send this aim to the one person who can approve your change.

Paper studied: Not chosen yet
SQUIRE sort: Not done
Run chart lab: Not done
Your SMART aim: Not written yet
3M filter: 0 of 3
Your measures: Not written yet
Journal fit: Not done

Think like an editor.
Build like a researcher.