Quality Improvement Methods
31. Pareto Analysis
a. Definition: An analysis, typically represented by a bar chart, whereby items or problems are listed along the horizontal axis of the chart and are sorted in order of importance or frequency. This can help in prioritizing future work or identifying where attention should be focused. The chart is based on the Pareto principle or the idea that often when many factors affect a situation, usually just a few will account for most of the impact. Often 80% of the impact, such as errors or defects, is caused by just 20% of the factors - few are important, many are not. It is named after the Italian economist Vilfredo Pareto who first noticed in the early 1900s the fact that often an 80-20 proportion occurs.
c. Example: Data is gathered over several weeks regarding the incidences of patient falls and the unit or location of the patients when they fell. The frequencies are totaled by location and then sorted. This is shown in a bar chart below or cumulative line graph. The locations are arranged by frequency with the most frequent on the left. This highlights which few areas deserve further study and which do not.
1) Define the topic or scope of the analysis. It may be a quality improvement issue or a problem related to performance in a particular area.
2) Identify related data. This may be existing historical data or new data which must be gathered. Generally, we are interested in the frequency of a particular item or occurrence of a problem. The data set should be of sufficient size or cover a sufficient time period to give a complete picture of the situation.
3) Gather the data into categories. The number of categories, such as problem types, should not be too few, generally not less than about five, nor too many to be actionable.
4) Sort the data by frequency and put into a table or graph for clarity.
A spreadsheet program such as Excel is often handy to help construct a Pareto chart.