A3 thinking is a problem solving approach that is at the heart of Lean. The Simpler problem-solving A3 includes nine steps or “boxes”. As a physician, I like following the A3 process because it takes the same logical approach that physicians use to evaluate and treat patients. Both are deeply rooted in the scientific method.
Let’s take a look the nine boxes of the A3 and see how they compare to the steps in working up a patient.
A3 Thinking in Patient Care
Box One: We start in Box 1 – “Reason for Action” or “Problem Statement”. In patient care, this is the “Chief Complaint”. What is the problem that was significant enough for the patient to present for care? The more clearly we define the problem, the better we will be able to ensure that our solution will meet the patient’s needs. It’s common in patient care to hear the chief complaint and quickly start moving toward a treatment plan.
If a patient’s chief complaint is shortness of breath, it’s common that we can quickly assess the root cause from across the exam room, whether this is due to pneumonia, CHF, asthma, anemia, or possibly some other less likely diagnosis. We can do this because we have seen tens of thousands of patients. We couldn’t do this with the first patient we worked up in our third year of medical school. That took us hours and we often missed important aspects of the patient’s problems.
In A3 Thinking we constantly fight the temptation to “jump to solutions”, and consciously work on Box 1 until it is well-defined. This takes a bit longer up front, but saves time and prevents problems in the long run. If we don’t identify the correct problem, it’s likely that our solution will be wrong. We are like third year medical students when it comes to diagnosing and solving workflow problems.
Box 2: the “Current State”. Clinically, we perform our initial history, physical, and diagnostic studies to get a working diagnosis. Much of this is metric based – vital signs, physical findings, lab data – that define the patient’s condition. There is also the subjective component of how the patient feels physically and is feeling about their illness. To solve workflow problems we also do a diagnostic evaluation – going to the “Gemba” (where the work is done) to examine the problem for ourselves rather than rely on reports, and we use a combination of objective data and subjective assessments.
Box 3: the “Target State”. In Lean problem solving this specifies improvements we plan to achieve in the Current State metrics – shorter times to complete a task, fewer defects, less waiting, or greater staff or patient satisfaction. In patient care, the target state is the patient’s baseline or optimal health.
Box 4: Defining the difference between Box 3 and Box 2 leads us to Box 4 – the “Gap Analysis”. We calculate the difference between the current state performance and the target state goal to quantify the gap. In our example of dyspnea, if the patient’s respiratory rate is 28 and their baseline is 16, the gap is 12 breathes per minute. In order to make the right treatment decision, we need to identify the root cause of the problem. This is our differential diagnosis. We use a number of Lean tools that help us identify the root cause. A fishbone analysis assists in making sure we are not missing possible causes of the gaps. We apply a Pareto analysis to decide the most likely gaps that deserve deeper analysis to determine root cause. We “ask why five times” to get to the true root causes of those gaps.
Box 5: With a good sense of the root causes, we use Box 5 to decide on our “Solutions Approach”, following an “if we, then we” format. In our clinical example, if our root cause analysis identified CHF as the root cause, the “if we, then we” approach would include “if we give the patient a diuretic, then we expect that the pulmonary edema will decrease leading to a decrease in the respiratory rate.”
Box 6: The “if we, then we” solutions guide Box 6, in which we perform “Rapid Experiments”, actually testing the root cause and solutions approach. If our assumptions are correct, the metrics we are following in Box 2 and 3 should improve. If our assumptions are wrong, our experiments can lead to worsening performance. If the root cause is pneumonia, the dyspnea may well worsen with a diruetic, so we would reconsider our original diagnosis of CHF. It the root cause is CHF, and the patient improves, we are ready to commit to a treatment plan.
Box 7: – the “Implementation Plan” Once our experiments have validated the solution to the problem, we document the new treatment plan, or Standard Work in Lean parlance, including the steps to be completed, who is responsible for each item, and when it is expected to be completed. This is tracked over time to ensure that indeed each item is completed. In our clinical example, this may include a plan to change from IV to oral diuretics along with other adjuncts.
Box 8: We then periodically follow up in Box 8 – the “Confirmed State” metrics – tracking against the initial and target state metrics, to look for improvement. Clinically we will see the patient daily in the hospital, then perhaps weekly for a while as an outpatient. If we don’t see the improvements we expect, we re-evaluate and change the treatment plan. In workflow improvement, if we don’t achieve target state metrics, we use a PDCA cycle to analyze and adjust the standard work.
Box 9: “Insights”. What did we learn from this process that can help us when we encounter a similar situation in the future? Clinically we incorporate the experience we gained from caring for this patient into our assessment of subsequent patients. In workflow improvement, we similarly recognize what we have learned that we will apply in our next improvement effort.
The Lean A3 Thinking approach to problem solving and the approach that physicians use to evaluate and treat patients are strikingly similar. This makes the A3 Thinking approach to solving problems a natural fit for physicians.