Meeting the Needs of the Community: A System for Redesigning Care

1. Meeting the Needs of the Community: A System for Redesigning Care Mike Hindmarsh Hindsight Healthcare Strategies Steven J. Bernstein University of Michigan Ann Arbor VAMC Winslow Lecture Battle Creek, MI 9 October 2007

2. Overview • Mike Hindmarsh • Burden of chronic illness • Chronic care model • Research findings • Steven J. Bernstein • Model for Improvement • Quality improvement approaches • Application of the chronic care model

3. Mrs. C – We all know one • Ms. C is a 68yo woman with cough and shortness of breath and risk factors for Type II diabetes. She calls her doctor who cannot see her until the following week. • Two days later she is hospitalized with shortness of breath. She is diagnosed with “CHF”, discharged on captopril and a “no added salt diet” with encouragement to see her MD in three weeks • When she sees her MD, he does not have information about the hospitalization • PE reveals rales, S3 gallop, edema and possible depression • Ms. C is told she has “a little heart failure”, encouraged not to add salt, and Captopril is increased. Her depression is not addressed. • She is told to call back if she is no better • Two weeks later Ms. C calls 911 because of severe breathlessness and is admitted. • Fuller history in the hospital reveals that she has been taking the Captopril "as needed" because it seems “strong”, and she has never added salt to her diet, so her diet hasn’t changed. • Further tests reveal elevated blood glucose. She is warned of impending diabetes. • She is discharged feeling ill and frightened.

4. Four Biggest Worries About Having A Chronic Illness (Age 50 +) • Losing independence • Being a burden to family or friends • Affording medical care

5. The Increasing Burden of Chronic Illness For example: Patients with diabetes have * Arthritis (34%), obesity (28%), hypertension (23%),cardiovascular (20%), lung (17%) ** Physical (31%), pain (28%), emotional (16%), daily activities (16%) *** Eating/weight (39%), joint pain (32%), sleep (25%), dizzy/fatigue (23%), foot (21%), backache (20%)

6. Differences between acute and chronic conditions (Holman et al, 2000)

7. Systems are perfectly designed to get the results they achieve The Watchword

8. Problems with Current Disease Management Efforts • Emphasis on physician, not system, behavior • Lack of integration across care settings hindering quality care • Characteristics of successful interventions weren’t being categorized usefully • Commonalities across chronic conditions unappreciated

9. Chronic Care Model Community Health System Health Care Organization Resources and Policies ClinicalInformationSystems Self-Management Support DeliverySystem Design Decision Support Prepared, Proactive Practice Team Informed, Activated Patient Productive Interactions Improved Outcomes

10. Model Development 1993 -- • Initial experience at GHC • Literature review • RWJF Chronic Illness Meeting -- Seattle • Review and revision by advisory committee of 40 members (32 active participants) • Interviews with 72 nominated “best practices”, site visits to selected group • Model applied with diabetes, depression, asthma, CHF, CVD, arthritis, and geriatrics

11. Essential Element of Good Chronic Illness Care Informed, Activated Patient Prepared Practice Team Productive Interactions

12. What is a productive interaction? Prepared Practice Team Informed, Activated Patient Productive Interactions • Assessment of self-management skills and confidence as well as clinical status • Tailoring of clinical management by stepped protocol • Collaborative goal-setting and problem-solving resulting in a shared care plan • Active, sustained follow-up

13. Self-Management Support • Emphasize the patient's central role • Use effective self-management support strategies that include assessment, goal-setting, action planning, problem-solving, and follow-up • Organize resources to provide support

14. Delivery System Design • Define roles and distribute tasks amongst team members • Use planned interactions to support evidence-based care • Provide clinical case management services • Ensure regular follow-up • Give care that patients understand and that fits their culture

15. Decision Support • Embed evidence-based guidelines into daily clinical practice • Integrate specialist expertise and primary care • Use proven provider education methods • Share guidelines and information with patients

16. ClinicalInformation System • Provide reminders for providers and patients • Identify relevant patient subpopulations for proactive care • Facilitate individual patient care planning • Share information with providers and patients • Monitor performance of team and system

17. Health Care Organization • Visibly support improvement at all levels, starting with senior leaders • Promote effective improvement strategies aimed at comprehensive system change • Encourage open and systematic handling of problems • Provide incentives based on quality of care • Develop agreements for care coordination

18. Community Resources and Policies • Encourage patients to participate in effective programs • Form partnerships with community organizations to support or develop programs • Advocate for policies to improve care

19. Advantages of a General System Change Model • Applicable to primary and secondary preventive issues, prenatal and pediatric, mental health and other age-related chronic care issues • Once system changes in place, accommodating new guideline or innovation much easier • Fits well with other redesign initiatives – such as improved access • Approach is being used comprehensively in multiple care settings and countries

20. Research and QI Findings about The Chronic Care Model

21. The Evidence • Randomized controlled trials (RCTs) of interventions to improve chronic care • Studies of the relationship between organizational characteristics & quality improvement • Evaluations of the use of the CCM in Quality Improvement • RCTs of CCM-based interventions • Cost-effectiveness studies

22. The IHP Chronic Care Collaborative • 15 participating practice teams • 6 months into process • Process measures moving; clinical outcomes close behind • Broad acceptance of QI process among team members • Spreading successful change is next step.

23. The Mrs. C We Want to Know • Mrs. C is discharged after her first bout of breathlessness with information about CHF, risk factors for diabetes, and assurance of rapid PCP follow-up • The discharge nurse notes Mrs. C’s conditions and care in the EHR and then sends an email to PCP’s office about her recent hospitalization. • The PCP's nurse ensures the physician sees the information; calls Mrs. C to schedule a follow-up within 48 hours; and adds her to the care team’s registry which prompts team to her future care needs. • Mrs. C is scheduled for 30 minutes: 15 minutes with her physician and 15 minutes with the nurse. The physician explains CHF and diabetes to her. He orders the appropriate tests for diabetes and assures her that all will be fine recognizing her fear. He closes the loop with her to make sure she understood his recommendations and then briefly explained the concept of self-management support. • Mrs. C then visits with the nurse who steps her through a collaborative goal setting and action planning process. While Mrs. C is a bit overwhelmed, she is assured that her care team will follow-up in the next couple of days by phone to make sure she understands her clinical and self-management care plan and to report on the results of diabetes test. • The nurse calls within 48 hours and informs Mrs. C that she should be able to manageher blood sugar by better diet and exercise. She reviews the CHF medications with Mrs. C and adjust dosage since it seems to be bothering her. • She is scheduled for a follow-up visit in one week to discuss her blood glucose in moredepth and encouraged to call her team with any concerns or symptoms in the meantime. • Mrs. C understands the hard work she needs to do to manage her conditions but is thankful for such a caring team.

24. Overview • Mike Hindmarsh • Burden of chronic illness • Chronic care model • Research findings • Steven J. Bernstein • Model for Improvement • Quality Improvement Approaches • Application of the chronic care model

25. For every problem, there is a solution that is simple, neat, and wrong. - H.L. Mencken

26. Model for Improvement What are we trying toaccomplish? How will we know thatachange is an improvement? What change can we make thatwill result in improvement? Act Plan Study Do From: Associates in Process Improvement

27. Model for Improvement What are we trying toaccomplish? How will we know thatachange is an improvement? What change can we make thatwill result in improvement? A P S D D S P A A P S D A P S D Repeated Use of the PDSA Cycle Changes that result in improvement Implementationof Change Wide-Scale Tests of Change Follow-up Tests Hunches Theories, Ideas Very Small Scale Test From: Associates in Process Improvement

28. …if only they had done more PDSA cycles...

29. Quality Improvement Efforts R Grol. JAMA 2001:284:2578-85

30. Interventions and their targets Ann Intern Med 2002;136:642

31. Effectiveness of strategies to improve physician performance R Grol. JAMA 2001:284:2578-85

32. The University of Michigan experience with the model for improvement • Chronic disease management is a priority • Establish multi-specialty/department steering committee • Develop multi-payor chronic disease registries • Asthma, CAD, CHF and Diabetes • Collect relevant clinical information • Report information to provider & patient • Identify a responsible provider • Modify delivery system to improve care

33. Clinical Information SystemsReport to provider their patients data (diabetes) • Prioritize report and highlight items needing attention • Decide whether to usecoloror black & white ($$$) • Focus on information that can be easily obtained and that has clinical relevance

34. Educate providers and staff on improving clinical documentation • Click on: • Diabetes eye exam • Click on: • Foot exam – visual, sensory, pulses • Click on: • Self-management goal • Then enter the goal in • Additional information

35. Decision Support • Report detailed clinical data to the provider at patient's visit via an auto- mated system • Identify items that need attention; provide detailed action steps

36. Decision Support Identify registry patients for automated reminders • Dear John Smith, • At the East Ann Arbor Health Center, we want you to have the best diabetes care. To improve our care and to keep our medical record up-to-date, we are sending this letter with information regarding diabetes-related tests and exams…. After reviewing your medical record, we have the following recommendations: •  Please take the enclosed slip to the lab for a: • blood test to check your average sugar control (A1c) • blood test to check your cholesterol levels • urine test to check your kidneys • You do not have to fast for this test. We will contact you about the results and suggest follow-up if needed.

37. Decision SupportReporting • Reports • Monthly high priority patient reports sent to nursing and pharmacists for case management • Semi-annual physician, health center comparative, and leadership reports • Discussed at health center and department leadership meetings to determine resource allocation, educational needs • Incorporated into physician annual performance evaluation

38. UMHS Diabetes Report by Provider at one Health Center Physician Name % • - - - - • J. Smith • - - - - • - - - - • - - - - • - - - - 74 39 50 • Be willing to show all their data to a provider • Do notsingle out an individual provider to other providers

39. Primary care only Pulmonary / allergy Jointly managed Leadership Report on Asthma Care Patients age 18 years or younger < 18 years old all patients

40. Delivery System Design

41. Delivery System Design: PDSA • Use PDSA cycles at pilot sites to determine if an intervention is both feasible and effective • For example, improving hypertension control • Medical assistants at East Ann Arbor Health Center trained to add bright green sticker to patient encounter form if blood pressure > than target goal • Sticker prompts MD to note elevated BP and act • Sticker also prompts clerk to automatically schedule 2-4 week follow up with MD or pharmacist

42. Delivery System Design Patients with diabetes with blood pressure<135/80 UM average = 58% Health Center

43. Delivery System DesignIdentify Achievable Benchmarks 90%tile for BP < 130/80 across 25 academic teams participating in an Asso. American Medical CollegesChronic Care Collaborative is 56% Population studied by each academic team

44. Activate Patients • Activate and educate patients by providing them with information on how they are doing at the time of their visit • Insert patient data onto a take-home educational sheet

45. Self-management support • Handout is given to patient when put in the exam room • MA asks patient to think about a goal • MD supports goal • MA documents in medical record • MA calls patient in 2 weeks

46. Reinforce Self-management Goals

47. Self-management goals documented

48. University of Michigan Diabetes Care Improvements over time, 2004 – 2007 (n=9170)

49. Spreading Success • Horizontal from pilot clinics to the remaining fourteen primary care health centers • 131 PCPs, 157 residents • Standardize practice at the health centers • Vertical from the primary care clinics to geriatric and endocrine specialty clinics • Involvement of the multi-specialty team members assist in implementation • Across conditions from diabetes to other chronic conditions with registries (asthma, CAD, CHF)

50. For more information regarding the chronic care model and its application, please see: www.improvingchroniccare.org or contact Mike Hindmarsh……… hindmarsh.m@ghc.org Steven J. Bernstein….. sbernste@umich.edu Thank you