Harvardx course Practical Improvement Science in Health Care with the Institute for Healthcare Improvement.
Frequently, when you look at a problem, you may have questions about various groupings within your data. For example, is there a difference between new staff members vs. experienced staff? Do we see a difference between shifts or by clinical training? Nurse vs. physician vs. paramedic? This desire to look at the population as various subgroups is called stratification. By stratifying your data, you may be able to identify opportunities for improvement that aren’t obvious from the data as a whole. I worked with a group of pharmacists that were trying to improve the process known as medication reconciliation in hospital. Medication reconciliation is the process of creating the most accurate list possible of all medications a patient is taking and comparing that list against the physician’s orders. The goal is to provide correct medications to the patient at all transition points within the hospital. As the pharmacists were discussing the process, they quickly identified different scenarios within the process worthy of stratifying, including day versus night and weekday versus weekend. It was predicted there were variations in the process in these areas that would be worth understanding.
So let’s take a look at an example of that data. Let’s imagine that their prediction came true. So if we were going to draw ourselves a little bit of a run chart, and we come down here. And imagine we had our days of the week, right. So we have Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, Sunday, Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, Sunday. So we got two weeks here. And imagine we’re coming along and we’re plotting our data and on Monday and Tuesday the data is kind of right here. Natural variation. But then all of a sudden we look up and on the weekend we have a change in our data points that seem really different. It’s almost like there’s a different system that’s occurring on the weekends. These examples here are a different system than the ones down here below. And so this is a very common example of where you might look at a place where it makes sense to stratify out the two different parts of the week because you’re seeing that there are different systems that exist. Another common example is to create a small multiples display of measures. For example, imagine we’re thinking about hand-washing. So we may want to look at our data overall and have a run chart that tells us about our overall data. So overall. But we may have a prediction that there is some variation across our different sites, and so we might have several different places that we’re looking at. And we’ll just call ’em like A, B, C, D, and E. And we’ll stratify those individual units out– or places– so that we can look at their data and see is there variation that exists across sites? And the stratification helps us potentially see where there’s opportunity for improvement that will help change our outcome. So by doing this, it enables us to be able to see where there’s variation in the system. When you think about your work, what are some of the obvious stratification opportunities you see that might help you learn more about your process?
—
If this was helpful, share and include me @DaveWilliamsATX. Sign up here to receive a monthly email from me that includes all my blog posts and other Improvement Science resources I think you’d appreciate.