Video Transcript: Pareto Analysis Bob Lloyd, PhD, Executive Director Performance Improvement, Institute for Healthcare Improvement A frequently used tool in quality improvement is something that is traced back to a fellow by the name of Vilfredo Pareto. It’s a great name to say actually, Vilfredo Pareto. And Pareto was a sociologist and economist in Italy who developed what is classically known as the Pareto concept, or the old 80/20 rule. That is, when he looked at the economics of Italy, the discovered that 80% of the wealth of Italy was owned by 20% of the people. Now, the Pareto concept was turned later on into a Pareto chart, a diagram, that was able to show you essentially what are called the vital few, and separate those from what are classically known as the trivial many, and so I want to show you what that tool looks like. It is again grounded in that principal that Pareto established back in the late 1800’s. Now that we’ve briefly explained the Pareto concept, let’s show what a Pareto chart actually looks like. I’ve laid one out here to save a little time. Interestingly enough when you make these by hand it’s actually a rather interesting task. The computer does them in the blink of an eye, but let me show you the pieces that you have to set out in order to make this particular graph. We have two Y-axis on a Pareto chart – the first one is dealing with the number or frequency of a reason why something didn’t happen, or happen properly. On the right side, the Y-axis is the cumulative percentage of these individual reasons. So, let’s see how that plays out. First of all, let’s think of looking at med errors as our topic. What are the reasons for a medication error? It could be wrong dose, wrong time, wrong patient, it could be dose repeated, or dose omitted. Now, we have a bunch of med orders, all stacked up, and we’re going to review them and so this one is okay, this one is not okay, this one is not okay, and this one is okay. And, so let’s say that we have a whole pile of orders and now we’re going to take those that were not okay, and look at our reasons - wrong dose, wrong time, patient, repeated, omitted. Out of this stack of orders, let’s say that we find that there are about two hundred errors that we’ve detected. So what we do is we go through all the charts and then we start to look at the reasons, and then we put little tick marks to figure out the frequency, and which one is the highest number of frequency becomes the first box. This is a rank ordered decreasing array of the data, and let’s say that our first box here contains the wrong dose, that was the major reason for why there were medication errors, and then it’s the wrong time, let’s say that this was because the dose was repeated, this is the wrong patient, category four, and finally, the dose omitted as the last one. When we work out the numbers, let’s first of all look at the frequency. There were ninety reasons that the dose was detected as being the wrong dose. We had seventy in terms of the frequency of wrong time. We had twenty repeated, sixteen the wrong patient, and finally, four in terms of being a med that was omitted. Now, if we compute the absolute percentage that each of these is of the total, of two hundred errors, we see that the first column is forty-five percent, the second is thirty-five percent, this next column is ten percent, then eight percent, and finally two percent. So now we have the absolute, the numbers in dark, the frequency, and the percent that each of those is of the total of two hundred errors that we detected. Now what we do is we start accumulating these. So forty-five, plus thirty-five, plus ten, eight, two eventually all add up to one hundred. We start to make a curve coming off this first bar, and then we add the next one to it. It starts to rise quickly but then it starts tapering off because the percentages get smaller, then to apply the Pareto principal, we look at the spot where eighty percent on the cumulative percentage, this is the point at which eighty percent of all the errors are detected, and if we want to apply that Pareto principal of 80/20 we’ll come across and find the point at which that horizontal line strikes the curved line, the cumulative percentage, and then we drop a line down. Now we find very quickly, that these two categories, classically known as the vital few, are the two areas that if we want to work on improving this procedure, medication detection and errors, these are the two things, wrong dose and wrong time. These [remaining] things, while interesting, are called the trivial many. It’s not that they’re not important, but they are less important than these two main categories. So the Pareto principal, two axis’, the categories are the reasons why something didn’t work properly, are laid out according to the frequency and then the absolute percentage, and then we accumulate those to find where this 80/20 split occurs; the vital few versus the trivial many.