Quality Improvement Methodology Next Steps

Transcription

1 Quality Improvement Methodology Next Steps

2 This breakout Some experience of testing using the Model for Improvement, would like support to take their work beyond testing and onwards to implementation, scale up and spread. If you are stuck or have made some improvements but are wondering what to do next, then this session is for you.

3 Participants will be supplied with a template prior to the learning session asked to bring a brief description of their current work which will form the basis of the practical elements of the session.

4 Purpose of this Session Consider the components of a learning system in your own improvement activity : 4. Sequential testing of new theories 6. Planning for spread at scale What stage are you at in relation to these aspects of the improvement journey? What do you need to do next to ensure that your tests are scaled up and spread correctly?

5 Aim Measures Changes Testing & Implementation The Improvement Guide, API

6 Cycles of Tests Build Confidence S D A P D A P S A P D S Changes that will result in improvement Learning from data S D A P S D A P Proposals, theories, hunches, intuition A S P D S A P D

7 1 person 1 day 1 family 1 setting Start small Move to 3,5,7. as confidence grows

8 You can only learn as quickly as you test! Years Quarters Months Weeks Days Hours Minutes Drop down next two levels to plan test cycle!

9 Components of a Learning System 1. System level measures 2. Explicit theory or rationale for system changes 3. Segmentation of the population 4. Learn by testing changes sequentially 5. Use informative cases: Act for the individual learn for the population 6. Learning during scale-up and spread with a production plan to go to scale 7. Periodic review 8. People to manage and oversee the learning system From Tom Nolan PhD, IHI

10 Sequential Testing & Scale Up Aim Achieve improved communication process for HV and SW handovers using standardised format A P S D A P D S A P D S A P D A P S A P D S Implement Cycle 6: Test all handovers-1 week Hunch/Theory A structured handover will ensure accurate information sharing between prof teams A P S D S D Cycle 5: other team HVs/SWs- 1 day Cycle 4:Test with all handovers-1 HV/1week Cycle 3: Test with 5 more families/ same HV (Mon/Tu) Cycle 2: Test with 3 family handovers/ same HV (Wed) Cycle 1: Test developed handover form 1 HV/1 SW 1 family (Mon)

11 Learning Through Sequential Testing at your tables discussion time Where are you currently? Can you describe an example of the following? Multiple tests with different people/under different conditions? Learning/data captured to describe your testing journey?

12 Testing v Implementation Testing Trying and adapting existing knowledge on small scale. Learning what works in your system. Multiple tests in a variety of conditions Implementation Making the change a part of the day-to-day operation of the system. Permanent change Would the change persist even if its champion were to leave the organisation? Avoid implementation until confident that processes are robust

13 INSERT VIDEO CLIP RE TESTING

14 Components of a Learning System 1. System level measures 2. Explicit theory or rationale for system changes 3. Segmentation of the population 4. Learn by testing changes sequentially 5. Use informative cases: Act for the individual learn for the population 6. Learning during scale-up and spread with a production plan to go to scale 7. Periodic review 8. People to manage and oversee the learning system From Tom Nolan PhD, IHI

15 A Framework for Spread Institute for Healthcare Improvement Leadership Topic is a key strategic initiative Goals and incentives aligned Executive sponsor assigned Day-to-day managers identified Measurement and Feedback Better Ideas Develop the case Describe the ideas Set-up Target population Adopter audiences Successful sites Key partners Initial spread plan Social System Key messages Communities Technical support Transition issues Knowledge Management

16 Things to Consider Checklists For Spread Leadership, Better Ideas & Set Up General Communication & knowledge transfer Developing Measurement, Feedback and Knowledge Management Systems

17 Take a strategic approach to scaling up Level of testing Strategy

18 Increasing uptake of Healthy Start Vitamins Level of testing Strategy 1 1 EY practitioner gives vits to 1 mum Free vitamins in Asda with every pregnancy kit

19 Are You Ready for Scale up & Spread? Table discussion time In the context of your current improvement activity: Have you been testing a theory so that it could be considered ready for implementation in the area you are working? and/or Do you have a strategy for moving to scale up and spread to other sites/teams?

20 Seven Spreadly Sins

21 Seven Spreadly Sins 1. Start with a large pilot 2. Find one person to do it all 3. Be assured vigilance and hard work will solve the problems 4. Do one small test and then spread everywhere 5. Get the pilot team to spread the improvement system-wide 6. Only look at the process/quality measures on a quarterly basis (or less frequently) 7. Expect marked improvement in outcomes early on without attention to process reliability

22 1. Start with a large pilot

23 2. Find one person to do it all

24 3. Be assured vigilance and hard work will solve the problems

25 4. Do one small test and then spread everywhere

26 5. Get the pilot team to spread the improvement system-wide

27 6. Only look at the process/quality measures on a quarterly basis (or less frequently)

28 7. Expect marked improvement in outcomes early on without attention to process reliability Memo clearly worked job done! Memo issued to all staff Must do this!

29 Institute for Healthcare Improvement The Seven Spreadly Sins

30 Sustaining Improvement using Measurement Having the correct measures to provide assurance that new processes are reliable Measuring compliance or satisfaction through regular and random sampling of the population Understanding the variation that exists in your data

31 What measures? Outcome measures directly relates to the overall aim what is the result? how is the system performing? Process measures are the processes that contribute to the aim performing as planned? Balancing measures assessing from different dimensions unanticipated consequences, other factors influencing the outcome

32 V R A I A T I O N If we don t understand the variation that lives in our data, we will be tempted to Deny the data as it doesn t fit with our view of reality See trends where there are none Try to explain natural variation as special events Blame and give credit to people for things over which they have no control Distort the process that produced the data Kill the messenger!

33 Quality Improvement (e.g. % of pregnant women offered CO monitor) Action taken on all occurrences Better Quality Worse Source: Robert Lloyd, Ph.D.

34 Measurement of Improvement Define measures that will measure the impact of the Improvement work over time They will guide your progress through and beyond testing to implementation and monitoring for continuous improvement. Different ways of measuring e.g., Per cent compliance with process A count of correct attempts/number of attendances Verbal feedback /surveys

35 Common Cause Variation

36 Guiding Principles for Creating Charts If you have less than 10 data points, make a simple line graph to see where the data points are going. If you have 10 and 12 data points you can convert the simple line graph to a run chart (place the median on the line graph and apply the run chart rules). When you have data points you can calculate a control chart but you should note that the control limits are trial control limits.

37 Measure LOS (minutes) Elements of a Run Chart Annotate 1 The centerline (CL) on a Run Chart is the Median ~ X (CL) 160 Time 2/16/11 3/16 4/13 5/11 6/8 Week Four simple run rules are used to determine if special cause variation is present

38 Calculating the Median (odd number of values) Suppose we have a set of numbers as: 4, 52, 64, 32, 56, 33, 78, 100, 12 Arrange these numbers in ascending order as: 4, 12, 32, 33, 52, 56, 64, 78, 100 Cross off the lowest and highest value (as follows) 4, 12, 32, 33, 52, 56, 64, 78, 100 Therefore, 52 is the median in this case. Keep ALL values!

39 Calculating the Median (even number of values) Suppose we have a set of numbers as: 4, 52, 64, 32, 56, 33, 78, 12 Arrange these numbers in ascending order as: 4, 12, 32, 33, 52, 56, 64, 78 Cross off lowest and highest value (as follows) 4, 12, 32, 33, 52, 56, 64, 78 Find the average of the 2 centre values Therefore, ( )/2 = 42.5 is the median in this case. Keep ALL values!

40

41

42 Measure or Characteristic Rule 1 - Shift A shift in the process is six or more consecutive points either all above or all below the median. Values that fall on the median do not add to nor break a shift. Skip values that fall on the median and continue counting. Rule

43 Measure or Characteristic Rule 2 - Trend Five or more consecutive points all going up or all going down. If the value of two or more successive points is the same, ignore one of the points when counting. Like values do not make or break a trend. 25 Rule Median

44 Measure or Characeristic Rule 3 Runs Too few or too many runs Rule Median 11.4 Data line crosses once Too few runs: total 2 runs

45 Jan-08 Feb-08 Mar-08 Apr-08 May-08 Jun-08 Jul-08 Aug-08 Sep-08 Oct-08 Nov-08 Dec-08 Jan-09 Feb-09 Mar-09 Apr-09 May-09 Jun-09 Jul-09 Aug-09 Sep-09 Measure of Characteristic Rule 3 Runs Too many runs what is this data telling you? Rule 3 Too Many Runs

46 Too few or too many runs

47 Measurement or Characteristic Rule 4: Astronomical Value For detecting unusually large or small numbers: Data that is 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 25 Rule

48 Why use charts One of the most powerful but simple tools for improvement Describe a process captured over time (as opposed to being a single sample) Reveal any trends a process might be experiencing When combined with careful annotation they track the impact of change

49

50

51 Let s Build a Run Chart

52 How to Make a Run Chart 1. Identify the question you would like to answer using a Run Chart 2. Develop the horizontal scale (x-axis) 3. Develop the vertical scale (y-axis) 4. Plot the data points 5. Label the graph Note: Be sure you have gone out and collected the relevant data before trying to construct your chart 6. Calculate and place a median center line on the chart 7. Add any additional information which will communicate a more complete picture to the intended audience (including annotations on change efforts) Murray and Provost, Pg 3-4

53 RUN CHART EXAMPLE Months Percentage Notes Jan Feb Mar Event 1 Apr Event 2 May Jun Event 3&4 Jul Aug-12 8 Event 5 Sep Oct Event 6 Nov Dec Jan Feb Implementation Start Mar Apr-13 47

54

55 Trend

56 Trend Astronomical Point

57 Trend Shift Astronomical Point

58 Trend Shift Run 1 Astronomical Point

59 Trend Run 2 Shift Run 1 Astronomical Point

60 Trend Run 2 Shift Run 1 Astronomical Point Run 3

61 Trend Run 2 Run 4 Run 1 Shift Astronomical Point Run 3

62 Too few or too many runs

63 Types of Variation Common Cause Variation Is inherent in the design of the process Is due to regular, natural or ordinary causes Affects all the outcomes of a process Results in a stable process that is predictable Also known as random or unassignable variation Special Cause Variation Is due to irregular or unnatural causes that are not inherent in the design of the process Affect some, but not necessarily all aspects of the process Results in an unstable process that is not predictable Also known as non-random or assignable variation

64 Point Common Cause does not mean Good Variation. It only means that the process is stable and predictable. For example, if the percentage of pregnant women offered CO monitoring averaged around 30% and was usually between 25% and 35%, this might be stable and predictable but unacceptable..

65 Point Similarly Special Cause variation should not be viewed as Bad Variation. You could have a special cause that represents a very good result (e.g. a high % of pregnant women being offered CO monitoring in some particular weeks) which you would want to emulate. Special Cause merely means that the process is unstable and unpredictable.

66 Point You have to decide if the output of the process is acceptable!

67 Measurement Principles Develop aims before measuring Design measures around aims How Good By When Develop process and balancing measures Be clear on your operational definitions

68 Measurement Principles Establish a reliable baseline Track progress over time using annotated run charts Teams need measures to give them feedback that the changes they are making are resulting in improvement Need to understand common cause and special cause variation to ensure we don t over/under react to situations

69 Deming s System of Profound Knowledge Appreciation of a system Theory of Knowledge Psychology Understanding Variation

70 Appreciation of a System Complex system of interaction between people, procedures and equipment Success depends on integration, not performance of individual parts Theory of Knowledge Change is prediction of improvement based on knowledge of the system Learning from theory, experience Operational definitions are the basis for improvement with PDSA cycles for learning Improvement Psychology Interaction of people with systems Motivation & will of individuals & teams Situation awareness/decision making Managing stress and fatigue Helps planning for change management Understanding Variation Variation is to be expected everything we measure varies We make decisions based on interpretation Data over time data story of what has been happening

71 What s the Scope of Change? System Targeted for Implementation (Defined by Aim) Single-unit prototype: segments Spread to Total System (Additional units, sites, organisations) As you move from pilot testing to implementation to spread, your population of interest will need to be adjusted.

72 Thank You