PDSA stands for Plan, Do, Study, Adjust—also commonly written as Plan, Do, Study, Act. It is an iterative method for testing a change, comparing what happens with what you predicted, and deciding what to do next. It applies to improvement work in manufacturing, healthcare, services, software, and office processes.
A proposed change is a hypothesis, not a guaranteed improvement. PDSA helps a team learn before committing to a broad rollout. A cycle can be small and quick when the problem and risks are straightforward; more complex changes require more preparation and evidence.
The four steps · A small PDSA example · A real example from Mark Graban's book · Plan your next test
Start with the gap between what is happening and what customers or employees need. Observe the work and involve the people who do it. Agree on what you want to improve, a change worth trying, and what you expect to happen.
Define the smallest useful test: where it will happen, who will take part, how long it will run, and what will be observed or measured. Consider possible side effects and the conditions under which the test should stop or be reversed. The preparation should fit the risk and uncertainty.
A useful prediction: “If we put the frequently used tools at this workstation, we expect less searching without creating crowding or making the correct tool harder to identify.” An observable prediction may be enough for a simple local change; consequential or uncertain changes need stronger evidence.
Carry out the agreed test, rather than treating the idea as a finished solution. Explain what is changing to the people affected. Record what actually happened, including departures from the plan and unexpected effects.
A test might involve one workstation, one form, one team, or another carefully chosen setting. Small does not automatically mean safe. Follow the appropriate safety, quality, and approval requirements for the work.
Ask whether the change had the effect you expected. Use direct observation, relevant data, and feedback from the people affected. Look for both benefits and problems, including effects on other departments or customers.
In Healthcare Kaizen, Mark Graban and Joseph E. Swartz emphasize verifying results honestly, through qualitative or quantitative evidence. If the idea was intended to reduce walking or waiting, check whether that happened. Completing the action is different from confirming the benefit.
One encouraging result does not automatically justify applying a change everywhere. Consider whether the test was representative and what remains uncertain.
The next decision might be to adopt the change in the tested setting, adapt it and run another cycle, or stop that approach and try something different. Explain why, and carry the learning into the next test.
When a change works, update the relevant standard work, communicate it, and continue checking that it helps. Testing under additional conditions may be appropriate before spreading it. A successful first cycle can lead to further refinement rather than ending the learning.
Small, low-risk ideas should not require an elaborate project or an A3 form. But moving quickly should still include checking whether the change helped.
In his article Maybe “Just Do Its” Should Be Called “Just PDSA Its”?, Mark Graban uses a simple example: a trash can is in an inconvenient location. The team can follow a quick cycle:
The same thinking applies to relocating a tool, simplifying a routine form, or clarifying a handoff. A short conversation and direct observation may be sufficient for some changes. The point is to preserve the learning without adding paperwork that costs more than the problem warrants.
In Chapter 7 of The Mistakes That Make Us, “Iterate Your Way to Success,” Graban describes a nurse's idea at Mary Greeley Medical Center: attach a caddy to a bedrail to help patients keep personal items within reach.
Instead of immediately purchasing one for every bed, the team considered a smaller test. Would patients use the caddies? Which design would work best? Would different patients or units have different needs? How practical would they be to clean?
The hospital tried the idea, observed use, and gathered patient feedback. It ultimately left units the option to purchase caddies rather than buying the same model for every bed. The initial proposal became an opportunity to learn about suitability before making a larger commitment.
The principle transfers to other industries: test a new fixture at one appropriate workstation, trial a revised document with a small group, or evaluate a service change in a limited setting before expanding. Starting small reduces the cost of discovering that an assumption was wrong. Further tests may still be needed before deciding to scale.
PDSA is a way of thinking that can support different sizes of improvement. It does not require the same documentation, team structure, or meeting schedule for every idea.
Appropriate authorization and safeguards still apply. A trial should not remove a protective control simply to see what happens.
Chapter 7 of The Mistakes That Make Us emphasizes learning through experiments rather than defending an idea as though it must succeed. A thoughtful test that challenges a hypothesis gives the team information. Leaders should make it possible to report that result honestly.
Discuss what was expected, what actually happened, and what the difference suggests. Was the hypothesis wrong? Was the change carried out as planned? Did it help one group while creating work for another? Decide what to change in the next cycle.
Healthcare Kaizen illustrates this system perspective with supply-room organization: a layout that makes restocking easier may make it harder for nurses to find what they need. Evaluate the whole effect rather than celebrating one department's local improvement.
Before starting, agree on these questions with the people involved:
Afterward, record what happened, what you learned, and the next decision. KaiNexus can support managing improvement work; the essential discipline is making the prediction explicit, checking the result, and following through. Learn more about supporting PDSA with KaiNexus.
.svg)
Copyright © 2026
Privacy Policy