PDSA stands for Plan-Do-Study-Act, a four-step cycle used to test and improve processes. It’s essentially the scientific method applied to real-world problem solving: you form a hypothesis, run a small experiment, analyze the results, and decide what to do next. Originally developed for manufacturing, PDSA is now one of the most widely used frameworks in healthcare, education, and business for making incremental improvements without overhauling entire systems at once.
Where PDSA Came From
The roots of PDSA go back to the 1930s, when statistician Walter Shewhart proposed that quality control follows the same logic as the scientific method. In his 1939 book, Shewhart laid out a three-step process: specification, production, and inspection, which he compared directly to hypothesizing, experimenting, and testing. He initially described these as a straight line but later revised them into a repeating cycle, recognizing that improvement is never a one-and-done process.
W. Edwards Deming, who had edited Shewhart’s early lectures, took that cycle and reshaped it. In 1950, he introduced his version during an eight-day seminar in Japan sponsored by the Japanese Union of Scientists and Engineers. That presentation helped launch Japan’s postwar quality revolution. Deming continued refining the idea for decades, and in 1993 he formally named it the PDSA cycle, calling it the “Shewhart Cycle for Learning and Improvement.”
The Four Stages
Plan
The Plan phase is where you define what you’re trying to improve, design a small test, and make a specific prediction about what will happen. This isn’t just goal-setting. You need to spell out the who, what, where, and when of your test, and you need a concrete way to measure results. The prediction is arguably the most important part: it forces you to clarify exactly what you expect to change and by how much. A prediction like “50 percent of children in one kindergarten class will receive a vision screening” is far more useful than “we’ll screen more kids,” because it gives you something measurable to compare your results against and helps you figure out the people and resources needed to make it happen.
Do
In the Do phase, you carry out the test on a small scale. This might mean trying a new process with one patient, one classroom, or one shift. The point is to keep things contained so you can learn quickly without much risk. You also collect data during this step and note anything unexpected, whether it’s a barrier you didn’t anticipate or a side benefit you hadn’t considered.
Study
Study is where you look at what actually happened and compare it to what you predicted. Did the results match your expectations? Were they better, worse, or completely different? You analyze and interpret the data, then summarize what you learned. This phase is the intellectual core of the cycle. It’s what separates PDSA from just “trying stuff and hoping it works.”
Act
Based on what you learned in the Study phase, you choose one of three paths. You can adopt the change and begin scaling it up if the results were strong and the process ran smoothly. You can adapt the change by tweaking your approach and running another cycle if the results were promising but the execution needs work, or if there were unintended consequences to address. Or you can abandon the change entirely if it clearly didn’t produce improvement, and go back to brainstorm a different approach. Then the cycle starts again.
How It’s Used in Healthcare
The Institute for Healthcare Improvement (IHI) built PDSA into what it calls the Model for Improvement, which pairs the cycle with three guiding questions: What are we trying to accomplish? How will we know that a change is an improvement? What change can we make that will result in improvement? Teams move back and forth between these questions as they learn from each cycle.
In practice, a hospital team might use PDSA to reduce patient wait times in an emergency department. The first cycle could test a new triage process with one nurse on one shift, measure how it affects wait times for that group of patients, study whether the results match predictions, and then decide whether to refine the process or try something else. Each cycle builds knowledge. After testing under varying conditions (morning versus night, weekdays versus holidays), the team gains enough confidence to implement the change more broadly.
The key principle is starting small. A single PDSA cycle might cover just one patient or one day. This keeps the stakes low, the feedback fast, and the learning concrete.
PDSA vs. PDCA
You’ll often see PDSA and PDCA (Plan-Do-Check-Act) used interchangeably, but Deming himself preferred “Study” over “Check” for a specific reason. “Check” tends to be interpreted as verifying whether people followed the new process correctly. “Study” signals something deeper: analyzing the results of your experiment and understanding why things turned out the way they did. The distinction matters because checking compliance is a narrow activity, while studying outcomes drives genuine learning and better decision-making in the next cycle.
Common Mistakes
A systematic review in BMJ Quality & Safety found that the core principles of PDSA are often poorly executed in practice, with “substantial variability” in how cycles are designed, carried out, and reported. The most common failure modes fall into a few patterns.
In the planning stage, teams frequently skip defining clear success criteria, don’t identify key stakeholders, or design data collection plans that can’t actually answer their questions. Without a theory connecting the change to its intended outcome, the whole exercise becomes guesswork.
Many teams get stuck in the Do phase and never progress to Study. They implement a change but don’t follow through on analyzing the results. When teams do reach the Study phase, they sometimes collect too little data or the wrong kind of data to draw meaningful conclusions. And in the Act phase, a common error is jumping too quickly from a small-scale test to full-scale implementation without enough evidence, or failing to act on what was learned at all.
One particularly subtle pitfall is what researchers call a failure of “double loop learning,” where teams never step back to question whether their overall goals still make sense in light of what they’ve discovered. If the data suggests the original problem was framed incorrectly, a good PDSA process should prompt you to rethink the goal itself, not just the solution.
Why It Works
PDSA’s power comes from its simplicity and its insistence on learning before scaling. Rather than designing a massive overhaul and hoping it works, you test one small change, see what happens, and build from there. Each cycle adds knowledge. Over multiple iterations, you develop a change that’s been refined by real-world evidence, not just theory. The cycle also creates a built-in discipline: predictions force clarity, data collection forces honesty, and the adopt-adapt-abandon decision at the end forces accountability. When done well, even a “failed” cycle is valuable because it tells you what doesn’t work before you’ve invested significant resources.