Spreading the success to an entire nation: tools for a successful Care Transitions project 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 Outline the essential processes & milestones for a successful initiative Describe the need for impactful measurement strategies Display examples from around the nation 2
Every component is necessary 1. Expertise in an individual area, 2. Experts from different areas coming together to address a specific problem, but 3. Backbone Organization/Group is critical: Common agenda Common measures Structured collaboration 4
Community Coalition Building Information about Intervention Models Social Network Analysis Strategic Plan Develop and Formalize a Charter 5
Personal Narrative Strong commitment to the relationships Thanks to Marshall Ganz, NOI, OfH and others
En
Creating shared relational commitment Relationship as Interest Common Interests New Interests Interests Interests Resources Resources New Resources Common Resources Relationship as Resource
9 Creating shared structure
The roadmap again
A blown-up roadmap Root cause analysis What the intervention strategy should target What to measure Interventions The right intervention for the right target Who gets it, how much of it they get Measure What happens when the intervention reaches its intended target 11
Purpose of the Root Cause Analysis (RCA) Identify the root cause of readmissions within your community Identify patterns of readmissions specific to your community and its providers Use RCA results to guide targeting criteria and intervention selection 12
Community-Specific Root Cause Analysis Patient/family interviews Care coordinator interviews Medical record reviews Data Analysis Process mapping Cause-and-effect diagrams 5 Whys 13
Patient/Family Interviews Semi-structured telephone or face-to-face interviews with patients who were readmitted Helps to identify opportunities for improvement from the patient s perspective 14
Care Coordinator Interviews Conduct individual and/or group interviews with care coordinators Identify patterns, trends, and opportunities for improvement from the staff member s perspective Formulate groups across settings or within provider teams, organizations, or specialties 15
Medical Record Reviews Review randomly sampled hospital discharges and 30-day readmissions. Common finding: Patient education is completed and documented, but patients need more in-depth understanding to be compliant 16
Process Mapping Clarify specific roles and contributions of those involved in the process Observe discharge and admission processes directly, interview process owners, and map the processes Elicit staff perceptions about where communication issues and gaps may occur 17
Cause-and-Effect Diagram (Fishbone Diagram) Visually illustrates potential causes of high readmissions 18
5 Whys This is simple and easy to complete without statistical analysis. Start with asking why readmissions occur at your hospital and record the answer. If the answer provided does not directly identify the root cause of your readmissions problem, ask why again and record the answer. Continue this process until your team agrees the problem s root cause has been identified. 19
5 Whys Example Why are so many Medicare beneficiaries with heart failure being readmitted to our hospital? Because they do not understand or remember the red flags related to their condition after discharge. Why do they not understand the red flags? They do not have the correct documentation or reminder systems in place. 20
5 Whys Example (cont d) Why do they not have the proper documentation or reminders? Because they did not receive a Personal Health Record (PHR) or red flag magnet with documentation of these red flags upon discharge. Why did they not receive the PHR or magnet? Distribution of these materials is not part of the current discharge process. 21
Building a Community-based Program Root Cause Analysis Did Intervention Address Driver ID Driver of Readmission Measure Intervention Select Intervention
Results from the community-specific root cause analysis Existing local programs and resources Funding resources Cost estimates of intervention implementation Estimates for intervention penetration Sustainability Intervention Selection & Implementation Plan Community preferences 23
Intervention Measurement Strategies Involves both process and outcome Measures Providers and CBOs collect most Process data QIOs can help link Outcome Measures from Medicare claims to interventions QIOs can create time series control charts to show intervention progress and to monitor potential effects
Suggested approach 1. Map out a detailed, community-level logic model of the intervention strategy. 2. Select and operationalize outcomes and processes from the logic model. 3. Develop and enforce the system for tracking implementation and outcome. 4. Effectively report time series data.
Logic model Visual representation, roadmap How a program is expected to work Context of the real world where the program is implemented Conceptual Utilized in program planning, management, evaluation and communication
Logic model components Inputs Resources, contributing factors Outputs Activities (interventions) Participation (processes) Outcomes Short-, medium, and long-term Assumptions External factors
Logic Model Outcomes Short-, medium-, and long-term changes and improvements Short-term Specific improvements in the targeted driver or root cause Medium-term Related outcomes along the causal path Long-term Improved care transitions Avoided readmission Improved health care utilization Implications of potential negative changes or non-changes
A few ideas CTI a) Pre- vs. post-coaching patient activation b) % coached patients with increased patient activation c) % coached patients who achieve 30-day health goal RED a) Percentage of RED patients reporting no problems at telephone follow-up b) Care Transitions Measure (CTM) BOOST a) Pre- vs. post-implementation patient satisfaction with discharge b) Rate of discharge summaries completed w/in 48h or discharge. INTERACT a) Pre- vs. post-implementation detection of acute change in condition using early warning and/or communication tools (re: "Stop and Watch" and care paths) 30
Business case for measurement Facilitate rapid cycle improvement See indicators of immediate, ground-level effect Make efficient, informed program decisions Do it without onerous data collection Leverage what s already being collected Incorporate new data elements into existing processes Demonstrate program effectiveness Tell the real story How people actually benefitted from the intervention That a goal was set and reached 31
Guiding the discussion What is the goal of the intervention? Are there multiple goals? How will you know the goal was achieved? Are there several observable components? What data would you need to show it? Is it something already being collected? Could you insert new data elements into an existing work process? How burdensome would it be to collect entirely new data? Who would be responsible? 32
Guiding the discussion, cont d What would improvement look like? A score or rate that changes over time People who get the intervention have a better outcome than those who don t People are better off after getting the intervention, compared to before getting it How large of a change would be needed to confidently say that improvement occurred? Over what time frame should this happen? 33
Aiming for SMART-ness SMART objectives dial in the measurement strategy Specific Measureable Quantify the amount of change expected. Provide a reference point for measuring change. Achievable Realistic Time-phased http://www.cdc.gov/healthyyouth/evaluation/pdf/brief3b.pdf 34
Intervention How often did it happen? Process measure = number of times intervention occurred Is the intervention working? Interim/Proximal Outcome measure = Number of times the intervention did what it was intended to do Are we getting the expected outcome? Final Utilization measure = readmissions/1000 in the entire target population Is it enough to make a difference? Proximal Utilization Outcome measure = Utilization among those receiving the intervention
Let s see some real examples Examples of intervention measurement strategies Long-term outcomes Short-term outcomes Process / Reach measures Success stories Interim QIO Data 36
Lateral Cluster: 30day hospital readmission rate from SNFs in Harlingen http://www.cfmc.org/caretransitions/files/feb24_2011%20learning%20session_final.pdf
Short Term (% of HF patients who can TeachBack) 120 Percentage of CHF patients Who Can Teachback 100 80 81 78 90.9 77.5 85.7 88.1 92.2 60 62.7 68.3 66 R² = 0.8262 40 39 44.6 20 Intervention began March 2012 0 Jan-12 Feb-12 Mar-12 Apr-12 May-12 Jun-12 Jul-12 Aug-12 Sep-12 Oct-12 Nov-12 Dec-12
CTI Reach Process /Reach (CTI) Patient Referral Acceptance Percent 100.0% 90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% Completion Rates among Accepted Cases 39
A FEW STORIES OF SUCCESS ACROSS THE COUNTRY 40
41 Washington County, Rhode Island
Washington County, Rhode Island Intervention Send a complete communication document at the time of patient transfer (Transfer information sheet with reason for ER visit, current medication list, face sheet-demographics, DOB,PCP, Insurance, Emergency Contact, IC, Advance Directives) # Beneficiaries Touched 630 Schedule outpatient follow-up appointment prior to discharge 2600 Provide PCP with summary clinical information at discharge 2600
Percent of Patients with Communication Form that is 100% Complete 120% 100% 80% 88% 90% 77% 78% 75% 83% 92% 98% 86% 94% R² = 0.2019 89% 60% 40% 20% 0% Feb 12 Mar 12 Apr 12 May 12 Jun 12 Jul 12 Aug 12 Sep 12 Oct 12 Nov 12 Dec 12
% Percent of Patients who have MD Appt Booked at Discharge 100.0 80.0 60.0 40.0 20.0 0.0
% Percent of patients whose summary info is provided to PCP 100 90 96 98 97 98 95 96 95 96 97.0 100.0 100.0 99.0 97.0 95.0 95.0 80 70 60 50
12.7%* 31.1%* *10.1.10-3.31.11 compared to 10.1.11-3.31.12
47 Lufkin, TX
Lufkin, Texas Intervention # Beneficiaries Touched Follow Up Appointment Scheduled for CHF Patients 151 Follow Up Appointments Scheduled for CHF, AMI, and PNE patients 190 Patient Education: Use of CHF Zone tool for CHF patients 50
Follow-up Appointments Scheduled for CHF Patients 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 75.0% 63.6% (12) 54.5% (7) (6) 43.8% 33.3% 40.0% (7) 30.8% 27.2% (3) (4) 27.8% (4) 20.0% 20.0% (3) (5) 12.5% (2) (2) (2) n=11 n=9 n=13 n=16 n=18 n=10 n=16 n=10 n=11 n=11 n=10 n=16 Jan-12 Feb-12 Mar-12 Apr-12 May-12 Jun-12 Jul-12 Aug-12 Sep-12 Oct-12 Nov-12 Dec-12
30-day Readmission for CHF Patients 50.0% 40.0% 44.4% 30.0% 20.0% 10.0% 7.7% 16.7% 13.6% 0.0% Aug-12 Sep-12 Oct-12 Nov-12 Dec-12
Follow-up Appointments for CHF, AMI, PNE Patients 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 52.6% (10) 52.4% 40.0% (4) 57.7% (15) (11) 50.0% (5) 40.0% 40.0% (6) (6) 33.3% (7) 27.3% 26.3% (3) (5) n=19 n=10 n=21 n=15 n=21 n=26 n=11 n=19 n=15 n=10 n=23 43.5% (10) Jan-12 Feb-12 Mar-12 Apr-12 May-12 Jun-12 Jul-12 Aug-12 Sep-12 Oct-12 Nov-12
8.9%* 5.3%* *10.1.10-3.31.11 compared to 10.1.11-3.31.12
53 San Francisco, CA
The San Francisco Transitional Care Program Intervention # Beneficiaries Touched Care Transitions Intervention (CTI) 201 Enrollment for In-Home Support Services (IHSS) 90 Provision of Meal Services 45 Provision of Transportation Services 26
CTI Enrollment Rate
In-home Support Services Enrollment Rate
Meal Services Enrollment Rate
5.5%* 13.3%* *10.1.10-3.31.11 compared to 10.1.11-3.31.12
59 The State of Montana
The State of Montana
Billings, Great Falls, Missoula-Rivalli, MT Community Intervention # Beneficiaries Touched Missoula/Ravalli Care Transitions Intervention (CTI) 661 Billings Check In 600 Billings Community coalition engagement intervention strategy ~2600 Billings End of Life Culture Change ~100 s Great Falls Safe Landing 89
% Accepting Check In 120.00% 100.00% 91.30% 93.10% 96.97% 100.00% 80.00% 82.35% 60.00% % Accepting Check In 40.00% Expon. (% Accepting Check In) 20.00% 0.00% Aug-12 Sep-12 Oct-12 Nov-12 Dec-12 Aug-12 Sep-12 Oct-12 Nov-12 12-Dec % Accepting Check In 91.30% 82.35% 93.10% 96.97% 100.00% 62
Billings Level of Coalition Development 7 6 5.40 5.75 5 4.70 4.12 4 3 2 1 1.00 0 Baseline 12/1/2011 Jun 2012 Jul 2012 Aug 2012 Oct 2012 63
End of Life Culture Change 7 6 5 4 3 Stage of development of Provider Education Tool Baldrige ADLI Linear (Stage of development of Provider Education Tool Baldrige ADLI) 2 1 0 64
Statewide Measures 7.6%* 15.2%* *10.1.10-3.31.11 compared to 10.1.11-3.31.12
Wisdom from A. Einstein Learn from yesterday, live for today, hope for tomorrow. The important thing is to not stop questioning. Not everything that counts can be counted, and not everything that can be counted counts If we knew what it was we were doing, it would not be called research, would it? Logic will get you from A to Z; imagination will get you everywhere.
Questions? Kimberly Irby, MPH Program Manager National Coordinating Center (NCC) Colorado Foundation for Medical Care (CFMC) Email: kirby@cfmc.org Phone: 303-695-3300 x3026 If you want to go quickly, go alone. If you want to go far, go together. -African proverb