1b s4 i interpreting runcharts final
0 likes783 views
Runcharts are tools used for assessing real-time changes in processes, helping to determine the effectiveness of individual changes or PDSA cycles. Understanding the variations within runcharts, which can be random or non-random, is crucial for making informed conclusions about process stability and improvements. The document outlines four rules to identify these variations and provides guidance on interpreting runcharts to drive decision-making.
1b s4 i interpreting runcharts final
- 1. © ABCi 2019 What is a Runchart? In contrast to data collected before and after a change, runcharts enable you to assess the effectiveness of each individual change or PDSA cycle in real-time and use this information to build knowledge of what works and what doesn’t. It enables you to evaluate whether you are achieving your aim (outcome) or that your processes are reliable. This guide will enable you to understand the variation that lives within your runchart, interpret it accurately in order for you to make the appropriate conclusions. See also ‘Creating Runcharts’ & ‘Using Runcharts’ Skills for Improvement Guides (2/3) See ‘Creating Runcharts (1/3)’ for instructions to construct runcharts in Microsoft Excel, also for further details around the components of a runchart, when you would use these and their benefits. See ‘Using Runcharts (3/3)’ to understand how to use runcharts in practice, plotting new data and recalculating your median to represent non-random variation in your data. Why is understanding variation so important? All systems exhibit variation, this may result from the many interactions or flaws that are built into the system. It is important to understand the variation within your system as this will enable you to make appropriate conclusions and decisions based on your data. Your runchart will demonstrate either random or non-random variation or even both. • Non-random variation is usually caused by a known factor. Something has changed in your process that varies your data. Non-random variation means that your process has become unstable, meaning that it is less predictable. When improving something, you will destabilise the current process and demonstrate non-random variation in your data in the appropriate direction (e.g. reducing delays or increasing compliance), you would want this improvement to be sustained and become stable. In order to interpret the variation in your runchart, you need to understand four rules. Should your data trigger one of these rules, your data is demonstrating non-random variation (see next page). Should your data not trigger one of the rules, your data is demonstrating random variation (see runchart above). • Random variation is data that displays irregular or erratic fluctuations, that are based on chance. This data shows that your process is stable, not changing in any way unless something is applied to change it. You decide if you are happy with the stable process running at this level or if it needs to improve. Interpreting Runcharts (2/3) ABC improvement iSkills for
- 2. © ABCi 2019 What are the runchart rules? There are four rules to determine whether your runchart is exhibiting random or non-random variation. If a rule is triggered, the data is demonstrating non-random variation. A Shift - 6 or more data points above or below the median. Given there is a 50% probability of a data point being above or below the median (p = 0.016), it is extremely unlikely that 6 in a row is due to luck. Any data point falling on the median will not break a shift, but will not be counted as part of the 6 to make the shift. A Trend - 5 or more points consecutively going up or down. If two or more consecutive points are exactly the same value, this is not enough to break the trend but will not be counted as part of the 5 that makes up the trend. Too many or too few runs – A non-random pattern may be signalled by too many or too few runs. A run is the amount of times the line crosses the median plus 1, or how many times the line is above or below the median. Of note, if there are too many runs, consider if the data on the graph is from multiple processes, e.g. day shift & night shift, or weekday & weekend. Count the number of data points on the runchart and look this up on the table on the next page. Lookup the corresponding upper and lower limits for the number of runs. Count the number of runs, this is the number of times the line crosses beyond the median plus 1, or the total number of times the line is either above or below the median. If the number of runs falls within the range between the lower and upper limit – this is random variation. If the number of runs falls outside the range between the lower and upper limit – this is non-random variation The graph above has 25 data points, looking this up on the table below, for the data to show random variation there should be between 8 and 18 runs. There are 6 runs (line crossing the median) in the graph above and because this falls outside the 8-18 range, this graph is exhibiting non-random variation.
- 3. © ABCi 2019 Number data points on run chart Lower limit for number of runs Upper limit for number of runs 10 3 9 11 3 10 12 3 11 13 4 11 14 4 12 15 5 12 16 5 13 17 5 13 18 6 14 19 6 15 20 6 16 21 7 16 22 7 17 23 7 17 24 8 18 25 8 18 Number data points on run chart Lower limit for number of runs Upper limit for number of runs 26 9 19 27 10 19 28 10 20 29 10 20 30 11 21 31 11 22 32 11 23 33 12 23 34 12 24 35 12 24 36 13 25 37 13 25 38 14 26 39 14 26 40 15 27 An Astronomical Point – Anyone eyeballing the chart would agree that an astronomical point is beyond the normal range Caution: Every data set will have a highest or lowest point. This doesn’t mean that it is astronomical