How can you make your processes reliable? (Part 2 of 5)

Frank Federico, RPh, IHI Executive Director

Welcome. This is Part 2 of our reliable design methodology. In Part 1, we talked about why we have reliable processes and what we need to do to make them reliable. Now we are going to talk about the methodology and how you can begin to apply it in your processes.

The first [rule] is to start small. That is, start with a subset of your population. We know that in all situations the same process won’t work for every patient, so we have to figure out which ones work best and how do we apply them in the context of the care that we are delivering.

We also have to design a process that is the easiest to work with. That is, we don’t want to add complexity to what the caregivers are doing every day. So we are going to simplify that process, make it as easy as possible to do the right thing. Ultimately by starting small, you can learn about what works well in that population — because if you can’t get it right in that population, what makes you think that you can get it right in the rest of the population? So learning will help you grow and spread your processes to other populations.

What is a subset or a segment? It should be a group of patients that are easy to identify. It should have willing participants on the part of the health care providers who are willing to work with you to test new methodologies, to test the way you want to design the process that make it more reliable. You can learn from that subset so that you can continue to expand and spread your work. The subset has to have enough volume that you are able to test daily or every other day. What we mean by that is that if you are not able to test daily and you can only test once a week, you can only learn once a week. If you run your test daily, every other day, or even multiple times in the same day, each one of those is an opportunity to learn, and so you can move more quickly in your improvement processes.

An example could be patients in one ICU, or it could be just Dr. Smith’s patients, for example. Or it could be just patients in one clinic of a large clinic setting. Or it could be just male patients over 65 who are coming in for a particular screening. Those are all examples.

Now your subset, you have to examine because if it’s too large, then you need to narrow it down a little bit. If your subset crosses many areas of care — that is, many clinics or many units, many wards — then that’s going to be a challenge, too, because each unit has its own culture and its own challenges, so you want to try to limit it to one area. But just remember that it has to be a volume large enough that you can test daily or at least every other day.

The next step is to visualize the steps in your process. We ask you to draw a flow diagram of 3–5 boxes and no more than that. Some of you may be familiar with Lean methodology or others that ask you to draw this very complex driver diagram, and there’s nothing wrong with that, but it’s really complicated, takes a lot of time, and we want you to visualize very quickly the 3–5 steps that are in your process.

Now, the 3–5 steps might look different from one hospital to another or one ward to another, but it’s conceptually what the process looks like for you and what needs to happen. Then within those, identify the defects. Now think of, is there a possible cascade? That is, if we don’t do the first step right, it doesn’t matter how much we make the other steps more reliable. We have to make the first step right.

So let’s take an example of what I mean. If we are developing reliable processes for managing a patient with a pressure ulcer, but we don’t have a good reliable process to assess the patient for pressure ulcer risk, then we can develop all those processes, but we don’t know how best to apply it because we haven’t identified the patients for whom to apply it.

In that case, the flow diagram would point out to you that the first step should be, “Let’s make sure that we have a reliable assessment process.” You might also identify, “What is the biggest defect that we need to work on?” because if we can fix that defect, maybe we can make the rest of the process more reliable. From there, you can continue because you will have to work on the other steps in the process as well.

This is an example of a high-level flow diagram. This is a patient who needed to receive a vaccination. And the patient is identified — again, there’s an assessment done that the patient is eligible for the vaccine — the vaccine is prescribed, the vaccine is administered, it is documented, and then it is communicated with primary care, for example. In this case, this team identified that they failed to identify the patient who should be eligible for that vaccine. Therefore, they needed to start there to identify where they can make the process more reliable.