Quality Improvement and Health Services Research in Clinical Settings

1. Quality Improvement and Health Services Research in Clinical Settings Jeffrey Hummel, MD, MPH Medical Director for Clinical Informatics, Qualis Health February 5, 2009

2. Overview of the next 90 minutes Background of the modern Quality Improvement movement – key concepts Evidence-based medicine IHI and the learning collaborative model Lean and where it fits How to write up a QI project How to use a hybrid model using all of the above to improve outcomes in real clinical settings Time for discussion

3. Quality Improvement Methodology Major transformation in late 1990s From measuring mistakes to process redesign From the methodology-heavy machinery of outcomes research… Before and after Intervention and control groups Rigorous statistical analysis …to emphasis on rapid assessment, agile implementation, and simple techniques to measure progress in closing quality gaps Far less academic and more problem-solving driven

4. Evidence-based Medicine Sackett: “The conscientious, explicit and judicious use of the best current evidence” While the standards for what was considered high quality evidence have gone way up, the methods for applying it have become more empiric In the words of David Eddy: “If it works, do it” “If it doesn’t work, don’t do it” “When there is insufficient evidence to decide, be conservative”

5. IHI: Institute for Healthcare Improvement A useful definition of quality in healthcare: “The right care for the right patient at the right time, every time, without fail.” “Every process is perfectly designed to give you exactly the outcome that you get.” - Don Berwick

6. IHI Model for Improvement Step 1: The Three Big Questions: What are we trying to accomplish? How will we know that a change is an improvement? What changes can we make that might result in an improvement? This is last because clinicians usual “jump to solutions” This is where Lean brings particular value

7. Step 2: PDSA Cycle - testing a change in a real world setting Plan: Understand the current workflow Design future state; Identify tools to support the future state; Decide what to measure & how Do: Implement future state Study: Look at what was measured; figure out what it means Act: Fix the things didn’t work the first time and retest until it works right; continue to find ways to improve the process

8. Rapid Improvement Cycle The cycles are linked for continuous improvement

9. But what do we measure? Don’t waste time trying to get perfect data Learn to navigate on minimal data points Don’t wait for the technology Use quick and dirty samples if necessary Examples: Wait times Number of tests ordered

10. Things that can be measured without a computer system Satisfaction Patient satisfaction Provider/Staff satisfaction Times (using a watch or stop watch) Time for Pt to move through the system Time to make/implement decision Limited chart review – keep it small Lab results Treatments (meds or immunizations given) Vital signs (height & weight)

11. Spread It isn’t enough to simply do a demonstration Success lies in spread across the organization Role of leadership is essential Replicate the process of education Replicate the data collection Replicate the PDSA cycles Can start with the perfected workflow from the pilot Try it in other areas, but be prepared for it not to be a perfect fit and to require modification

12. The Collaborative Concept - 1995 Short 6-15 month Collaboratives bringing teams from different settings all seeking improvement on a focused clinical area Teams of 3 to 5 people usually attend 3 learning sessions and report back to additional team members at the local organization Examples of goals: Reduce ED wait times by 50% Reduce hospitalization for CHF Pts by 50% Reduce worker absenteeism by 25%

14. The Collaborative model hasn’t transformed the healthcare system It is very expensive Relies on energy of physician champions It is probably the best kind of CME Difficult to hold the gains and still scale to multiple initiatives Doesn’t address the foundational problems in health care quality, which are rooted in the reimbursement system Yet the collaborative process remains a very powerful model

15. An Introduction to Lean The lessons of assembly line manufacturing can be applied to some workflows Giving immunizations Checking a medication list Sending a patient to the lab for blood work Most processes in health care delivery involve complex workflows involving multiple knowledge domains and several people.

16. Lean Methodology: 7 Wastes; things that don’t add value Waiting: time wasted when material, machines, or information are not ready for people Motion: any movement of people that doesn’t add value to the product Inventory: any material or work on hand other than what is needed for the customer Processing: work that appears important, but actually adds no value from the customer’s perspective Transportation: movement of material that add no value to the product Defects: work that is below the standard requested by the customer (the next process) Over-production: making more products than the customer (the next process step) needs right now

17. Healthcare vs. Manufacturing In manufacturing the 7 wastes tend occur in assembly line processing In healthcare most waste occurs: Trying to locate information Trying to organize the information so that a decision can be made Not actually completing a decision Trying to implement the decision so that the right thing happens

18. Empowering the staff Only the people who do the workflow know how it actually works but they only understand their part in it As a group they will define the entire workflow The people doing the work are the ones who are best able to find the waste and the variation The people doing the work are the best ones to design, implement and improve a future state Standardization is not the end, rather the beginning of quality improvement.

19. Qualis Primary Care Lean Project Planning session: Select high pain workflow of proper scale Select 8-10 team members Three 2-hour workshops in three days Plan implementation in series of PDSAs

20. Topics addressed so far Getting paper results into EMR in time Accurate med list in primary care visits Patients in lab without orders Well Child Visit immunizations Emergency Department Med Reconciliation Anticoagulation Clinic patient flow Operating Room turn-over time Nurse advice calls to the clinic

21. Getting the staff to lead the improvement effort – Part 1 Run a lean workflow simulation: Several available, we use sewing kit Measure productivity, inventory, cycle time Discuss the 7 kinds of waste Let staff redesign it Run simulation a second time with new workflow Productivity improves an order of magnitude with fewer people, less space

22. Getting the staff to lead the improvement effort – Part 2 Go and See: Walk through the process in the clinic letting each team member explain their part Instruct everyone to look for waste Have staff map out the current workflow Assign one person to write sticky notes Assign another to post the sticky notes Have staff note areas of variation Have staff note areas of waste Use sticky notes on butcher paper

23. Supporting the team by documenting their work Take a digital photo of the workflow diagrams Turn these into diagrams for the team

24. Drawing the work flow in Visio

25. Getting the staff to lead the improvement effort – Part 3 Eliminate waste where possible Even out flow where possible Standardize to eliminate variation. Standardization is the starting point for quality improvement Let everyone give input Let the team member find and defend their solutions

26. Identify metrics for improvement Most groups need help with this If there is an EHR is in place, look for easy-to-get numbers Tests ordered, Results of clinical measures e.g. blood pressure Time/date stamps For systems with no EHR use limited time/motion studies Satisfaction surveys This isn’t about perfection

27. Create an implementation plan and accountability What are the tasks that need to be done before a pilot can take place Each task has a responsible person Each task has a due date Plan a next meeting to follow up on progress

28. How do you publish this stuff? This is applied science, not scientific discipline About changing human performance, not generalizable knowledge It is driven by experiential learning It is a social process & therefore context driven What part of the intervention worked? For whom? Under what circumstances?

29. Standards for Quality Reporting Improvement Excellence SQUIRE Highlights the unique properties of improvement interventions The Practice of Improvement Planning Implementing interventions Evaluation of Improvement: Did an intervention work?Why did it (didn’t it) work? SQUIRE: Annals of Internal Medicine, Nov 2008; Vol. 149 Number 9

30. An HIV Clinic In Mozambique Approximately 500 new HIV positive patients each month and increasing Only 10% are having their CD4 counts done within 1 month of enrollment There is a registry to track patients Resources to buy reagents for CD4 testing are scarce Only those patients with resources to obtain ART get CD4 test

31. PDSA Cycle in Beira, Mozambique What are we trying to accomplish? All HIV positive patients will have a CD4 count within 1 month of presenting to the clinic How will we know that a change is an improvement? The percent of patients with CD4 count will rise from current level and approach 100% What changes can we make that will result in an improvement? Remove barriers to testing Remove non-value added steps from the workflow

32. Steady enrollment growth

33. Outcome of a process perfectly designed get 10% CD4 Testing

34. Plan: Standardize Initial Workflow

35. Do: New Standardized Workflow

36. Study: See if the new workflow is an improvement

37. Act: Remove Waste

38. Do: New Standardized Workflow

39. Study: After CD4 count order is “automated”

40. Act: Even out the flow by identifying pace-maker steps

41. Recap of the tools we used Evidence-based medicine: – target was designed to identify everyone who needs ART as early as possible Improvement Methodology: Clear articulation what we are trying to accomplish Changes tried out, adjusted to get them to work better, all of them required overcoming resistance, Measurement to track improvement Spread to other clinics This is “what to do”, now let’s talk about “how to do it”

42. What parts of “Lean” do we see in this case study? Involve the people doing the work in decisions about how to improve the process Go and see for yourself to thoroughly understand the situation Look for non-value added steps to identify waste. In healthcare it is usually waiting, motion and processing in order obtain and organize information Standardized tasks and processes are the foundation for continuous improvement and employee empowerment Identify the pace-maker process and schedule your flow to match the pace-maker

43. Questions?