QIO Care Transitions Activity: the Good News so far Kim Irby, MPH; kirby@cfmc.org Senior Project Director Colorado Foundation for Medical Care www.cfmc.org/integratingcare This material was prepared by CFMC (PM-4010-046 CO 2012), the Medicare Quality Improvement Organization for Colorado under contract with the Centers for Medicare & Medicaid Services (CMS), an agency of the U.S. Department of Health and Human Services. The contents presented do not necessarily reflect CMS policy.
Objectives Why we know it worked JAMA publication Describe the need for better transitions Discuss leading interventions in this work Outline current and future work and implications National and Alabama data 2
http://jama.jamanetw ork.com/article.aspx? articleid=1558278 It Worked!!
August 2008-July 2011: 14 QIOs working in 14 Communities AL: Tuscaloosa CO: Northwest Denver FL: Miami GA: Metro Atlanta East IN: Evansville LA: Baton Rouge MI: Greater Lansing area NE: Omaha NJ: Southwestern NJ NY: Upper capital PA: Western PA RI: Providence TX: Harlingen HRR WA: Whatcom county 4
Totals among Communities 70 Hospitals 277 Skilled Nursing Facilities 316 Home Health Agencies 89 Other types of Providers (Dialysis, Hospice, etc.) 666 ZIP Codes 1,125,649 Fee-for-Service Medicare Beneficiaries 5
The Strategy Define a community Identify service patterns associated with readmission Recruit and convene providers/partners To reduce unplanned 30d hospital readmissions for the community Using evidence based interventions and tools 6
Interim Quarterly Results Baseline Quarter Readmissions = 12,926 First quarter after intervention readmissions = 12,151 20.00% 19.80% 19.60% 19.68% 19.40% 19.48% 19.20% p=0.0024 19.00% 18.80% Jan07- Mar07 N = 66590 Apr07- Jul07-Sep07 Jun07 N = 62060 N = 64621 Oct07- Dec07 N = 62822 Jan08- Mar08 N = 65689 A Apr08- Jun08 N = 61781 Jul08-Sep08 N = 59098 B Oct08- Dec08 N = 59962 Jan09- Mar09 N = 61517 C Apr09- Jul09-Sep09 Jun09 N = 56395 N = 58825 Oct09- Dec09 N = 57766 Jan10- Mar10 N = 60616 D Apr10- Jul10-Sep10 Jun10 N = 57984 N = 59422 Oct10- Dec10 N = 59630
5,000 10,000 15,000 20,000 25,000 30,000 Numerator and Denominator Quarterly 68,000 Numerator (readmissions) Denominator (admissions) 63,000 58,000 53,000 48,000 43,000 Jan07-Mar07 N = 66590 Apr07-Jun07 N = 64621 Jul07-Sep07 N = 62060 Oct07-Dec07 N = 62822 Jan08-Mar08 N = 65689 A Apr08-Jun08 N = 61781 Jul08-Sep08 N = 59098 B Oct08-Dec08 N = 59962 Jan09-Mar09 N = 61517 C Apr09-Jun09 N = 58825 Jul09-Sep09 N = 56395 Oct09-Dec09 N = 57766 Jan10-Mar10 N = 60616 D Apr10-Jun10 N = 59422 Jul10-Sep10 N = 57984 Oct10-Dec10 N = 59630 Quarter The unit N represents target community eligible beneficiaries. MIlestones: A) baseline quarter; B) Care Transitions theme initiation (Aug 2008); C) intervention implementation (Jan 2009); and D) 28-month follow up quarter.
Rehospitalization Trends, Intervention and Comparison Communities -5.7% (p<.001) -2.1% (p=.08) P=.03 (difference)
Hospitalization Trends, Intervention and Comparison Communities -5.7% (p<.001) -3.1% (p<.001) P=.01 (difference)
Statistical process control Assesses variation in an outcome presumed to be related to system functioning A change worth investigating: Reduced variation (increased control) Significant change in the value of the outcome Process control limits = 3sd from the mean variation during baseline Significant : 8 points in a row above/below the mean with at least one point in the during intervention time period OR A single point above/below the process control limit in the during intervention time period
Community Results Rehospitalizations Intervention Comparison Special cause decrease 10/14 (71%) 22/50 (44%) Special cause increase 2/14 (14%) 13/50 (26%) Hospitalizations Intervention Comparison Special cause decrease 13/14 (93%) 31/50 (63%) Special cause increase 0/14 (0%) 8/50 (16%)
Control Charts an innovative way to measure progress in healthcare
Readmissions per 1000 Benes Target Community - AL_0 Readmissions per 1000 Benes 24 3s Limits For n=2: 22 UCL=22.01 The 20 18 16 Test1 Test1Test1 Test1 Test1 Test1 Test1 Test1 X=20.08 LCL=18.14 Improvement continues. 14 0 2 4 6 8 10 12 14 16 18 20 22 24 Target Community - TX_0 22 3s Limits For n=2: 20 UCL=20.66 18 X=18.54 16 14 Test2 Test2 Test1 Test1 Test1 Test1 Test1 Test1 Test1 Test1 0 2 4 6 8 10 12 14 16 18 20 22 24 LCL=16.42
What s important about this publication? Intervention communities avoided twice as many rehospitalizations (1 hospitalization for every 1000 Medicare beneficiaries) and hospitalizations (5 for every 1000 beneficiaries) as comparison communities Improvement for whole communities is a promising strategy Providers engaged based on relevance QIOs in the role of convener/supporter Included community and social services Unadjusted geographic population data allows easy data display/sharing
What Else was Important? Allowing flexibility leverages local resources/context Shewhart control charts published in a major peer-reviewed journal Rehospitalizations/1000 and hospitalizations/1000 metrics proved useful for improvement work
Why are people readmitted? Provider-Patient interface Unmanaged condition worsening; Use of suboptimal medication regimens; Return to an emergency department
Why are people readmitted? Provider-Patient interface Unmanaged condition worsening; Use of suboptimal medication regimens; Return to an emergency department Unreliable system support Lack of standard and known processes; Unreliable information transfer; Unsupported patient activation during transfers
Why are people readmitted? Provider-Patient interface Unmanaged condition worsening; Use of suboptimal medication regimens; Return to an emergency department Unreliable system support Lack of standard and known processes; Unreliable information transfer; Unsupported patient activation during transfers No Community infrastructure for achieving common goals
Why are people readmitted? Provider-Patient interface Unmanaged condition worsening; Use of suboptimal medication regimens; Return to an emergency department Unreliable system support Lack of standard and known processes; Unreliable information transfer; Unsupported patient activation during transfers No Community infrastructure for achieving common goals
Why Engage a Community? Every readmission begins with hospital discharge Every transition has 2 sides Isolated information is not safe medical management Inevitably need to share The problem of home Patients are people too Visibility to drive improvement and mission Providers are people too
It s not a hospital project
It s a Community Problem HHA SNF
Ways to Convene a Community
System-Level Drivers of Readmissions Provider-Patient interface Unmanaged condition worsening; Use of suboptimal medication regimens; Return to an emergency department Unreliable system support Lack of standard and known processes; Unreliable information transfer; Unsupported patient activation during transfers No Community infrastructure for achieving common goals
Interventions and Drivers Intervention Patient Activation Standard Process Information Transfer Care Transitions Intervention Transitional Care Model INTERACT II HHQI Best Practices Project BOOST Bridge model Project RED GRACE Model STAAR Initiative
CMS Table of Interventions Available at: www.cfmc.org/integratingcare
Hospital discharge standardization protocols RED 11 item checklist BOOST QI support for hospitalists and discharge planners Evidence: weak for readmissions; insufficient but probably necessary Driver: standard known process, information transfer Setting: Hospital
The CMS Discharge Planning Checklist Description: CMS developed checklist for patients and families to prepare for care capability after transition Resource: http://www.medicare.gov/public ations/pubs/pdf/11376.pdf
Care Transitions Intervention SM (CTI) Description: Transitions coaches support selfmanagement capacity Personal Health Record Medication discrepancy tool 5 contacts Evidence: RCTs and the Care Transitions Theme (30-50% 30d; 50% 180d) Driver: patient activation, information transfer Setting: hospital to home
Transitional Care Nursing Model Description: Transitional Care Nurses follow patients from the hospital into the home; work with a multidisciplinary team to develop and deliver comprehensive care plan; risk assessment tool(s) Evidence: RCTs (45% 90d) Driver: information transfer, creates a new standard process that individualizes services Setting: hospital to home health
Interventions to Reduce Acute Care Transfers (INTERACT) Description: Toolkit for SNF personnel to reduce avoidable hospital admission. Three types of tools: 1) communication; 2) clinical care paths; 3) advance care planning. Evidence: Ouslander (2008): Higher hospitalization rates associated with larger facilities, more Medicaid and Medicare skilled care residents, lower percentage of Caucasian residents and higher percentage of residents with impaired decision making; 68% of hospitalizations were avoidable, per expert panel record review.
In Reality Adapted models/components of models 2 tracks: activation plus provider process
What s he saying? I sure hope my wife is getting this.. No I m good to go. Whatever you say is what we ll do Doctor Blah blah blah, blah blah. Any questions? 1. Patient activation trumps all
The PAM is very helpful to guide interventions
2. Local adaptation is inevitable Adapt gold standard models Do not adapt others adaptations
3. Ask the community to help Brought to you by your Community Partners
Organize a Community Tie participation to values Include personal narratives Develop flexible tactics Community champion Align with other federal and local initiatives Develop a leadership team/advisory group
4. Measuring is important p=0.0024 Insist on a population based measure of progress 41
Recurring Themes in Successful Communities Community cohesiveness Provider activation/will Strategic Partners Cross-setting Work Coaching as an intervention Strong community leadership (e.g., physician champions)
QIO Care Transitions: Good News continues Integrate Care for Populations & Communities August 2011-present
August 2011 Integrating Care for Populations & Communities Aims: Improve the quality of care for Medicare beneficiaries as they transition between providers Reduce 30 day hospital readmissions (nationally) by 20% within 3 years 44
QIO assistance Toolkit Root cause analysis Learning and Action Networks Learning Sessions Community Convening Social Network Analysis Diagrams Hot-spotting maps Data, data, data (e.g., readmission/admission metrics; reach/intervention effectiveness measures) 45
QIO Accomplishments as of March 31, 2013 # of Engaged Communities 375 # of Beneficiaries Living there 12,455,368 # Formally Recruited Communities 227 # Communities with Signed Coalition Charter 221 # Applications Submitted 125 # Communities Receiving Formal Funding 81 # Recruited Hospitals 859 # Recruited Nursing Homes 1,533 # Recruited Home Health Agencies 901 # Recruited Hospice Facilities 342 # Recruited Dialysis Facilities 91 # Recruited Outpatient Physicians > 1900
National Coalition of QIO-recruited Communities Early Progress 6.8%
National Coalition of QIO-recruited Communities Early Progress 9.1%
50
Alabama Data Tied for 3 rd highest number of recruited communities over LOE; 5 th highest number of engaged communities over LOE Among 17 states having ALL interventions showing improvement 38 th in readmissions reduction 41 st in admissions reduction
Interim Reductions!!! 10/1/10-3/31/11 compared to 10/1/11-3/31/12 Community Readmissions Admissions Statewide 2.9% 2.8% Community Care Coalition of East Alabama Top of Alabama Care Transitions Innovation Coalition Healthy Gulf Coast Care Transitions 14.9% 9.4% 13.1% 5.2% 7.2% 6.5% 53
Alabama
Top of Alabama Care Transitions Innovation Coalition CCTP Partner as of March 2013 Intervention # Beneficiaries Touched PCP follow-up appointments arranged 600 CHF program readmission rates 1500 INTERACT II 42
Percent of those contacted Patients Contacted Who Had Hospital Arranged Follow-up Appointments 100% 90% 80% 70% 60% 50% 40% R² = 0.7375 Median 41% 30% 20% <-Intervention 10% 0% Oct-11 Nov-11 Dec-11 Jan-12 Feb-12 Mar-12 Apr-12 May-12 Jun-12 Jul-12 Aug-12 Sep-12 Oct-12 Month
Readmission Rate Monthly CHF Readmission Rates 35.00% 30.00% 25.00% 20.00% CHF Program Median 18.67% 15.00% 10.00% R² = 0.1116 5.00% 0.00% Month
Rate Rate of Unplanned ER Visit + Hospitalization 0.45% 0.40% 0.35% 0.30% 0.25% 0.20% 0.15% 0.10% 0.05% 0.00% R² = 0.1113 Median 0.25% Jan-12 Feb-12 Mar-12 Apr-12 May-12 Jun-12 Jul-12 Aug-12 Sep-12 Oct-12 Nov-12 Dec-12 Month
October 1, 2010-March 31, 2011 compared to October 1, 2011-March 31, 2012 5.2% Relative Improvement in Admissions/1000 Medicare FFS Beneficiaries 13.1% Relative Improvement in Readmissions/1000 Medicare FFS Beneficiaries
Community Care Coalition of East Alabama Intervention # Beneficiaries Touched PCP follow-up appointment scheduled by patient 256
Compliance Percent of Patients Who Made Their PCP Appointment 100% 90% 80% R² = 0.7991 Mean 92% 70% 60% 50% <-Intervention 40% 30% 20% 10% 0% March 26-31, 2012 Apr-12 May-12 Jun-12 Jul-12 Month
October 1, 2010-March 31, 2011 compared to October 1, 2011-March 31, 2012 9.4% Relative Improvement in Admissions/1000 Medicare FFS Beneficiaries 14.9% Relative Improvement in Readmissions/1000 Medicare FFS Beneficiaries
Seeing Ourselves in the System First image of the entire Earth - 1968 Each of us is a system citizen in that we are (potential) change agents in the systems of which we are a part.
How to Get Started Contact your QIO http://www.cfmc.org/integratingcare/files/icpc_contacts.pdf Join (and listen to archived) Care Transitions Learning Sessions http://www.cfmc.org/integratingcare/learning_sessions.htm Browse our Toolkit http://www.cfmc.org/integratingcare/toolkit.htm 65
Additional Resources Medicaring an independent website for improving care transitions www.medicaring.org Partnership for Patients www.healthcare.gov/compare/partnership-for-patients/ Community-based Care Transitions Program http://go.cms.gov/caretransitions The AoA Toolkit www.aoa.gov/aoaroot/aoa_programs/hcltc/adrc_caretransitions/toolkit/index.aspx 66
Thank You & Questions NCC website www.cfmc.org/integratingcare Kim Irby kirby@cfmc.org 303-784-5710 67