1. Measurement for Improvement

3. Why we look at data graphed over time Change Made Change to process made in June

4. Improvement Science Consulting O1O1 P1P1 Measure Types O = Outcome Measure P = Process Measure B = Balance Measure S = Process Step Measure PDSA = Learning Cycle Measure O2O2 P2P2 BB P1P1 P2P2 S3S3 S2S2 S1S1 System of Feedback PDSA 1 PDSA 2 PDSA 3 PDSA 4 S3S3 S2S2 S1S1 PDSA 1 PDSA 2 PDSA 3 PDSA 4 © Improvement Science Consulting

5. Break out – developing a useful project level dashboard Spend some time developing the measures you would like to include in a project level dashboard Which ones are outcome measures? Which ones are process measures? Do you have balance measures? Process step measures? Note any gaps in your measures – where would you like to add measures?

6. Improvement Science Consulting O1O1 P1P1 Measure Types O = Outcome Measure P = Process Measure B = Balance Measure S = Process Step Measure PDSA = Learning Cycle Measure O2O2 P2P2 BB P1P1 P2P2 S3S3 S2S2 S1S1 System of Feedback PDSA 1 PDSA 2 PDSA 3 PDSA 4 S3S3 S2S2 S1S1 PDSA 1 PDSA 2 PDSA 3 PDSA 4 © Improvement Science Consulting

7. As the scale of the test increases we move from qualitative to quantitative evidence Improvement Science Consulting Very small scale test of a change idea Large scale test of change idea or Implementation of a change idea Evidence primarily Qualitative Evidence primarily Quantitative with noticeable impact on process measures Sequence of learning and change

8. Run Charts Making and Interpreting

9. So what is a run chart? Murray and Provost, Pg 3-4

10. Defining elements of a run chart contains at least 10 data points must have a median tells the story through careful use of annotation

11. What time is it? Stop what you are doing Look at your watch Write down the time

12. How do we prevent this?

13. How do we interpret variation? Distinguishing between random variation and non- random variation Four rules for discovering non-random variation Shift Trend Too many or too few runs Astronomical values

14. Shift Rule: Six or more consecutive data points either all above or all below the median (skip values on the median and continue counting data points. Values on the median DO NOT make or break a shift.) Murray and Provost, 3 (11-15)

15. Why do we need 6 data points? What is the probability of a coin landing heads or tails?.5.5 x.5 =.25.5 x.5 x.5 =.125.5 x.5 x.5 x.5 =.0625.5 x.5 x.5 x.5 x.5 =.03125.5 x.5 x.5 x.5 x.5 x.5 =.015625

16. Trend Rule: Five or more consecutive data points either all going up or all going down. (If the value of two or more consecutive points is the same, ignore one of the points when counting; like values do not make or break a trend.) Murray and Provost, 3 (11-15) Median=11

17. Run Rule: Too many or too few runs (A run is a series of points in a row on one side of the median. Some points fall right on the median, which makes it hard to decide which run these points belong to. So, an easy way to determine the number of runs is to count the number of times the data line crosses the median and add one. Statistically significant change signaled by too few or too many runs). Murray and Provost, 3 (11-15) Median 11.4 10 Data points not on median Data line crosses once Too few runs: total 2 runs

18. Run Rule Reference Table Table is based on about a 5% risk of failing the run test for random patterns of data. Adapted from Swed, Feda S. and Eisenhart, C. (1943). “Tables for Testing Randomness of Grouping in a Sequence of Alternatives. Annals of Mathematical Statistics. Vol. XIV, pp.66 and 87, Tables II and III. Murray and Provost, 3 (11-15)

19. Astronomical Data Point (For detecting unusually large or small numbers: Data that is a Blatantly Obvious different value. Everyone studying the chart agrees that it is unusual. Remember: Every data set will have a high and a low – this does not mean the high or low are astronomical). Murray and Provost, 3 (11-15)

20. When should we transition from using a Run Chart to a Shewhart Control Chart? Perla, Provost and Murray