QSIR knowledge exchange - Matt Tite presentation
Download as PPTX, PDF1 like1,258 views
The document outlines a quality improvement program aimed at enhancing healthcare delivery based on the five-year forward view, focusing on diagnosis and treatment for cancer and other health services. It emphasizes the importance of using statistical process control (SPC) and Pareto analysis to identify issues in healthcare quality and capacity management. Additionally, it discusses capacity and demand management in primary care, suggesting that effective planning must take into account seasonal variations and different demand patterns throughout the week.
QSIR knowledge exchange - Matt Tite presentation
- 1. Organising for Quality and Value Delivering Improvement Programme Quality, Service Improvement and Redesign: Practitioner Programme Applying Quality Improvement to the Five Year Forward View
- 2. Let me train you to be an Analyst….
- 4. Toss a coin…. What are the chances?
- 5. LCL UCL MEAN 7 Points above centre line SPC rules – A run of Seven A run of seven points all above or all below the centre line, or all increasing or all decreasing X X X X X X X X X X X X X X X X X X X X 7 Points below centre line
- 6. 0 20 40 60 80 100 120 140 Falls Ulcers VTE UTI NumberogPatients Type of Harm Anywhere NHS Trust Harms (5 Months) 80% of the Harms 0 5 10 15 20 25 30 35 40 Ward A Ward H Ward Y Ward L Ward I Ward P Ward B Ward Z Ward Q Ward K Ward S Ward U NumberogPatients Anywhere NHS Trust - Falls(5 Months) The Pareto Principle
- 7. Combining SPC and Pareto…….
- 8. 0 20 40 60 80 100 120 140 Falls Ulcers VTE UTI NumberogPatients Type of Harm Anywhere NHS Trust Harms (5 Months) 80% of the Harms 0 5 10 15 20 25 30 35 40 Ward A Ward H Ward Y Ward L Ward I Ward P Ward B Ward Z Ward Q Ward K Ward S Ward U NumberogPatients Anywhere NHS Trust - Falls(5 Months) Mtweek_2 Mtyear_2 1885114420105316649 20142014201320132013201220122011201120112010 60 50 40 30 20 10 0 NumberofAttenders _ X=29.05 UCL=52.83 LCL=5.27 1 1 5 COPD pathways - Spring and Winter Attenders (weekly)
- 10. Aims measurements change ideas The Improvement Guide Langley et al (1996) What are we trying to accomplish? How will we know that a change is an improvement? What changes can we make that will result in the improvements that we seek ? Model for Improvement Act Plan Study Do testing ideas before implementing changes
- 11. The Five year view for Cancer…. Faster diagnosis. We need to take early action to reduce the proportion of patients currently diagnosed through A&E— currently about 25% of all diagnoses. These patients are far less likely to survive a year than those who present at their GP practice. Currently, the average GP will see fewer than eight new patients with cancer each year, and may see a rare cancer once in their career. Better treatment and care for all. It is not enough to improve the rates of diagnosis unless we also tackle the current variation in treatment and outcomes. Leydig tumors account for 1% of all testicular cancer
- 12. Question 1 What should we do first? A) Set some targets. Set Cancer targets that seem sensible, that stretch, but are achievable. B) Write a clear aims statement, which does not contain any solutions. C) Create SPC charts. D) Create Pareto charts.
- 13. Question 2 What’s next? After working up the Aim, what’s the next two tool needed? 1) Cause and effect diagrams and SBAR 2) Driver diagrams and 6 thinking hats 3) Pareto and SPC 4) Spaghetti Diagrams and Fresh Eyes Create Pareto charts to define and understand the problem. Use SPC charts to see the 80%’s if they have always been the problem, are growing or shrinking.
- 14. If I had an hour to save the world, I Would Spend 55 Minutes Defining the Problem and then Five Minutes Solving It (Einstein – Misquoted)
- 15. The Five Year Forward View Outcomes vs Processes The Donabedian Model Outcome Process Balancing Measures
- 16. Creating measures at all levels System Team Service Balance of Health and Social Care Outcomes Line of sight across the levels of integrated care • Type of measure by domain • Timeline (baseline – 1, 3, 5 and 10 years) Adapted from AquA
- 17. How capable are your processes of achieving targets?
- 18. How capable are your processes of achieving targets? Example - Door to needle times 75% of heart attack patients will receive thrombolysis within 20 minutes of their arrival in hospital
- 19. Imagine this is your process - Is it capable of achieving 75% of patients treated in 20 minutes? Door to Needle Times 0 10 20 30 40 50 60 70 80 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Consecutive patients Minutes Door to needle time Average UCL LCL Target
- 20. Door to needle time Average UCL LCL Moving range Average MR Target 28 31 68 0 14 20 45 31 68 0 17 20 50 31 68 0 5 20 40 31 68 0 10 20 20 31 68 0 20 20 15 31 68 0 5 20 46 31 68 0 31 20 30 31 68 0 16 20 20 31 68 0 10 20 45 31 68 0 25 20 46 31 68 0 1 20 34 31 68 0 12 20 20 31 68 0 14 20 10 31 68 0 10 20 30 31 68 0 20 20 50 31 68 0 20 20 30 31 68 0 20 20 23 31 68 0 7 20 10 31 68 0 13 20 30 31 68 0 20 20 28 31 68 0 2 20 30 31 68 0 2 20 9 31 68 0 21 20 40 31 68 0 31 20 42 31 68 0 2 20
- 21. Capability = Target - Average 3 * sigma Formula
- 22. Calculation (example) Target 10 mins Average 20 Sigma 18.7 10 - 20 = 3 x 18.7 -0.18 Target Average Sigma
- 23. Interpretation Value Capability of Achieving Target more than 1 100% 0-1 50-100% less than 0 0-50% How capable is the process of achieving the target?
- 24. Capability value % Capability Capability value % Capability 0 50 0.42 89.6 0.02 52.4 0.44 90.7 0.04 54.8 0.46 91.6 0.06 57.1 0.48 92.5 0.08 59.5 0.5 93.3 0.1 61.8 0.52 94.1 0.12 64.1 0.54 94.7 0.14 66.3 0.56 95.4 0.16 68.4 0.58 95.9 0.18 70.5 0.6 96.4 0.2 72.6 0.62 96.9 0.22 74.5 0.64 97.3 0.24 76.4 0.66 97.6 0.26 78.2 0.68 97.9 0.28 80 0.7 98.2 0.3 81.6 0.75 98.8 0.32 83.2 0.8 99.2 0.34 84.6 0.85 99.5 0.36 86 0.9 99.7 0.38 87.3 0.95 99.8 0.4 88.5 1 99.9 Table of Capability Values
- 25. Is your process Capable? = 20 - 31 = -0.30 37 The figure is negative so this process is not capable of achieving 100% within 20 minutes. A minus figure means more than 50% of patients will not meet target. The maximum this process can deliver is 18.4% and the target is 75% Therefore, the process needs significantly redesigning to achieve the target
- 26. Question 3 Before you can attempt capability calculations the data needs to meet certain criteria? A) No, You can do capability analysis on any data B) Yes , The data needs to be “in control and stable”
- 27. Question 4 How do you know if your data is stable, predictable and in control? A) All the points are inside of the UCL and LCL B) There are no runs of 7 points C) The distribution of the data is as expected D) All of the above
- 28. Understanding GP Capacity and Demand
- 29. 252321191715131197531 20 15 10 5 0 Observation Minutes _ X=10.68 UCL=20.54 LCL=0.82 Patient Time (minutes) - GW 252321191715131197531 20 15 10 5 0 Observation Minutes _ X=9.48 UCL=20.56 LCL=-1.60 Patient Time (minutes) - SP 252321191715131197531 25 20 15 10 5 0 Observation Minutes _ X=11.36 UCL=25.43 LCL=-2.71 Patient Time (minutes) - SF 252321191715131197531 30 20 10 0 -10 Observation Minutes _ X=11.64 UCL=29.70 LCL=-6.42 Patient Time (minutes) - JB 252321191715131197531 50 40 30 20 10 0 Observation Minutes _ X=16.16 UCL=38.88 LCL=-6.56 1 1 Patient Time (minutes) - SYA
- 30. 4 are the same, one seems different… SYAJBSPGWSF 20 18 16 14 12 10 8 Minutes Interval Plot of SF, GW, SP, JB, SYA 95% CI for the Mean SYA is statistically different from the other 4
- 31. Activity, Backlog, Capacity & Demand….. Demand: All the requests for a service – from all sources Capacity: All that we can do Bottleneck: Constraint is the cause of the bottleneck Activity: What we actually did Backlog = Queue = Waiting List
- 32. Face to Face Appointments 13/06/2014 11/06/2014 09/06/2014 05/06/2014 03/06/2014 30/05/2014 28/05/2014 26/05/2014 22/05/2014 20/05/2014 16/05/2014 14/05/2014 12/05/2014 300 250 200 150 100 50 0 Date NumberofAppts _ X=129.2 UCL=267.1 LCL=-8.8 Number of face to face appts (daily) 24/11/2014 04/11/2014 15/10/2014 25/09/2014 05/09/2014 18/08/2014 29/07/2014 09/07/2014 19/06/2014 30/05/2014 12/05/2014 160 140 120 100 80 60 40 20 0 Date NumberofPatients _ X=92.8 UCL=148.5 LCL=37.1 11 Number of Patients on triage (daily) Patients who were Triaged
- 33. Demand for GP Triage, is it the same each day? 65432 160 140 120 100 80 60 40 20 0 Mtday Numberontriage Boxplot of Number on triage 65432 160 140 120 100 80 60 40 20 0 Mtday Numberontriage Individual Value Plot of Number on triage vs Mtday Monday is statistically different from the other days of the week (the other 4 days are the same as each other)
- 34. Conclusions for planning C&D (triage patients only) You must plan for Mondays Separately Tue, Wed, Thurs & Fri are the same You cannot plan capacity for the whole year. It is affected by seasonality (runs of 7 points on SPC charts) – Possibly need more weeks to prove this You should plan one system from Oct onwards (winter) We only have data from May, so not sure when the seasonality ends in the data) – We can guess on the weather, but the years worth of triage data would be better
- 35. Planning C&D Take the 80% time (1.5 sigma above the average) – Half way between the average and the UCL. This is the time that it takes each Doc to do an appointment: – GW = 15 minutes – SP = 15 minutes – SF = 18 minutes – JB = 20 minutes – SYA = 27 minutes
- 36. Planning C&D part 2 – Demand 80% Monday’s 80% Triage = 139 people each Monday The rest of the week Triage = 105 people each day. Then times 80% times by 80% volumes to work out the minutes you require to meet demand… – If SF did all the work.. • 18 minutes X 139 people = 2502 minutes (41.7 hours each Monday) • 18 minutes x 105 people = 1890 minutes (31.5 hours each Tue, Wed, Thu & Fri) – Total for the week = 10,062 minutes per week (167.7 Hours a week)
- 37. This needs additional calculations Not all triaged work is a face to face appointment in the GP practice Some are ‘visits’ – Take 25 timings for each GP for visits, we can adjust the proportions accordingly.
- 38. Same Day Appts 24/11/2014 04/11/2014 15/10/2014 25/09/2014 05/09/2014 18/08/2014 29/07/2014 09/07/2014 19/06/2014 30/05/2014 12/05/2014 70 60 50 40 30 20 10 0 Date NumberofSameDayAppts _ X=39.22 UCL=69.22 LCL=9.22 11 Same day appts (Daily) 65432 70 60 50 40 30 20 10 0 Mtday Samedayappt Boxplot of Same day appt Monday and Tuesday are peak days for requests for Same day Appts, Wed, thurs & Friday are about the same. If the tails overlap they are not statistically different
- 39. Number of Visits 09/10/2014 23/09/2014 05/09/2014 20/08/2014 04/08/2014 17/07/2014 01/07/2014 13/06/2014 28/05/2014 12/05/2014 25 20 15 10 5 0 Date NumberofVisits _ X=11.9 UCL=25.60 LCL=-1.80 Number of visits (daily) 65432 25 20 15 10 5 0 Mtday numberofvisits Boxplot of number of visits Monday is statistically higher than the other 4 days of the week
- 40. Question 5: Do you routinely apply these techniques to your work? a) No, I didn’t know the techniques existed b) No, I know the techniques, but don’t have the time c) Yes, to some extent d) Yes all the time
- 41. Question 6: How many people need to be ‘experts’ in Quality Improvement in an NHS organisation of 4000 staff? a) 6 people b) 63 people c) 648 people d) 2129 people
- 42. Calculate the square root of the total number of people in your organisation. This is the number of QI ‘Expert’ that you need.
- 43. If I had an hour to save the world I Would Spend 55 Minutes Defining the Problem and then Five Minutes Solving It (Einstein – Misquoted) In God we Trust, everybody else bring data (W Edwards Deming)