Quality Improvement
Quality improvement (QI) projects make a specific process or outcome better in a specific setting, usually through a series of small, tested changes. Instead of a single before and after comparison, a QI project tracks its measures over time to see whether a change worked and whether it lasted.
Common mistakes for this type
- Measuring once before and once after, instead of tracking the measure over time.
- Leaving out a balancing measure, so unintended effects go unnoticed.
- Writing an aim without a number and a date, so there's no way to tell whether it was met.
What makes QI different
A specific aim. A QI aim says what will improve, by how much, for whom, and by when. Without a number and a date, you can’t tell whether you met it.
Small, tested changes. QI usually works through repeated cycles of planning a change, trying it, studying what happened, and adjusting. Plan-Do-Study-Act (PDSA) is the most common name for this cycle. Each cycle is a small test, and the project is the sequence of tests.
Measurement over time. Because changes are tested one after another, QI data are collected often, such as every week or month, and plotted in time order. This shows when a change happened and whether it held, which a single before and after comparison can’t do.
Local learning. QI aims to improve care in your setting. That focus is one reason many institutions review QI projects differently from research. Who decides, and how, varies. Read the guide
Three kinds of measures
| Measure | What it tells you | Example |
|---|---|---|
| Outcome | Whether the result you care about improved | Hospital-acquired pressure injuries per 1,000 patient days |
| Process | Whether the change was actually carried out | Percentage of patients with a documented skin assessment every shift |
| Balancing | Whether something else got worse | Nurses’ reported time spent on documentation each shift |
The example is hypothetical. Read the guide
Tracking data over time
Plot each measure in time order with a median line, and mark when each change was tested. A run chart is usually enough, and it needs enough data points to read, often at least 10. Start collecting baseline points as early as you can, because you can’t add them later. Read the guide
Before you start
- Your aim states what will improve, by how much, and by when.
- You have several baseline data points, not just one.
- Each measure has a written definition: what counts, what’s excluded, and where the data come from.
- You know who will collect each measure and how often.
- You have a balancing measure for anything your change could make worse.
Reporting
Many programs and journals expect QI projects to follow the SQUIRE 2.0 reporting guidelines. Check which guideline your program expects. Read the guide
Related guides
- Building and reading a run chart
- Choosing a statistical approach for your DNP project
- DNP project types explained
- Outcome, process, and balancing measures
- Valid, feasible, relevant: Choosing outcome measures
- Do you have baseline data? Testing your problem statement
- QI or research? How project review decisions work
- Setting up your project data
- Statistical vs clinical significance
- Sustaining the change after your project ends
- The proxy trap: When easy data measures the wrong thing
Not sure which type fits? Find your project type