Using a Positive Deviance Approach in a Regional Health Improvement - PowerPoint PPT Presentation

Reducing Disparities in Diabetes Care: Using a Positive Deviance Approach in a Regional Health Improvement Collaborative Randall D. Cebul, MD, Thomas E. Love, PhD, Douglas Einstadter, MD, MPH, and Shari Bolen, MD, MPH Better Health Partnership Center for Health Care Research and Policy Case Western Reserve University at MetroHealth Medical Center Cleveland, Ohio

Better Health Partnership’s Vision and Mission To help Northeast Ohio become a healthier place to live and a better place to do business By creating a safe space for health care competitors to collaborate 2

Positive Deviance and Positive Deviants • Positive Deviance (PD) is an approach to social change based on the observation that in any community there are subgroups whose successful strategies enable them to find better solutions to a problem than their peers, despite facing similar challenges. – These ( “ high outlier ” ) subgroups are referred to as positive deviants. – Better Health Partnership (BHP) uses a PD approach. – Uses clinical data from members to identify high outliers

BHP ’ s Positive Deviance Approach 1. Establish common quality of care goals among peers who have the same challenges 2. Measure and share data – publicly report 2/year 3. Use data to find the positive deviants 4. Interview/visit Key Informants among the positive deviants – describe what they did 5. Disseminate the Process in a safe space for competitors to collaborate. 6. Re-measure, publicly report change

What Better Health Measures and Reports Measurement: • Nationally endorsed, locally vetted standards • Electronic Medical Records • Focus on primary care, chronic conditions Report: • Care and Control of Diabetes (2007-), Heart Failure (2008-), and Hypertension (2009-) • Disparities across the region

EMR-based Disparities Measures on all patients • Insurance • Race/Ethnicity • Language Preference • Household Income • Educational Attainment • Geographic Location

Objective and Methods • Objective : – To examine changes in racial/ethnic disparities in diabetes care associated with using a PD approach in a regional collaborative, 2008-2014. • Methods: – Design: Observational – Sample: All clinics reporting > 7 times during 2008-2014 – Analyses: Regression models, weighted by clinic sample size, estimating changes in gaps (highest-lowest by race/ethnicity category) over the 7-year period

Diabetes Care Measures • A1c test Once past 12 months • Kidney mgmt. ACE/ARB or U.A./12 mos • Eye exam Documented past 12 mos. • Pneumococcal Vacc. Documented received Scoring: All or none credit = % all achieved Illustrative PD Approaches: MU of EHRs; PCMH: staffing, workflow re-design, cross-disciplinary teamwork, standing orders

Find High Outliers, Interview, Describe & Disseminate (Pneumococcal Vaccine) 2007 2008-2010 Who are those guys? How ’ d they do this? How they did this

The Results of Sharing Best Practices Regionwide improvement 2007 2007 2014 Who are those guys? How did they do this?

Pneumovax: Everyone Improves Example: Income

Main Results

RESULTS: Patient Characteristics in 53 Clinics 2008 2014 Total Total # Patients 21,213 34,185 % White 57.4 53.7 % A-A 37.2 40.3 % Hispanic 5.4 6.0 % Medicaid + 18.0 17.5 uninsured % Inner City 37.8 46.8 % prefers English NA 95.1

Changes in Care by Race/Ethnicity 55 50.4 White % meeting Diabetes Care standard 50 45 40 35 30 2008 2009 2010 2011 2012 2013 2014 Better Health Partnership Reporting Period

Changes in Care by Race/Ethnicity 55 White 50.4 % meeting Diabetes Care standard 50.5 50 45 40 35 30 2008 2009 2010 2011 2012 2013 2014 Better Health Partnership Reporting Period

Changes in Care by Race/Ethnicity 55 White 50.4 % meeting Diabetes Care standard 50.5 50 47.0 African-American 45 40 35 30 2008 2009 2010 2011 2012 2013 2014 Better Health Partnership Reporting Period

Changes in Care by Race/Ethnicity 55 52.5 White 50.4 % meeting Diabetes Care standard 50.5 50 47.0 African-American 45 40 35 30 2008 2009 2010 2011 2012 2013 2014 Better Health Partnership Reporting Period

Changes in Care by Race/Ethnicity 55 52.5 White 50.4 % meeting Diabetes Care standard 50.5 50 47.0 African-American 45 40 35 Hispanic 34.1 30 2008 2009 2010 2011 2012 2013 2014 Better Health Partnership Reporting Period

Changes in Care by Race/Ethnicity – Disparities Gone 55 52.5 White 50.4 % meeting Diabetes Care standard 50.5 50 47.0 48.2 African-American 45 40 Hispanic 35 34.1 30 2008 2009 2010 2011 2012 2013 2014 Better Health Partnership Reporting Period

Changes in Diabetes Care & Gaps by Race/Ethnicity 4.3 pts 16.3 points

Disparities Gap Reductions/Year: Regression Results Measure Gap Change 95% CI P-value Points/Year Composite -1.72 -2.74, -0.71 0.002 Eye Exam -1.14 -1.81, -0.47 0.002 Pneum. -2.23 -3.49, -0.98 0.001 Vacc A1c Done -0.20 -0.30, -0.11 0.001 Kidney Care -0.06 -0.30, -0.47 0.627 Safety Net

IMPROVEMENT in Meeting All DM Care Standards Across Non-medical Characteristics 2008-2015

Summary • Over a 7-year period, racial/ethnic disparities in diabetes care – gaps in % of patients reaching all four standards of care, by R/E, declined – declined signficantly • Improvements also were observed: – for 3 of 4 individual measures

Implications • Regional health improvement collaboratives exist in regions covering > 120M (40%) of U.S. residents: – Opportunities exist for widespread dissemination of best practices discovered by PD approaches • Regional health improvement collaboratives should be fostered by federal/state policies for care delivery and payment transformation. • Collaboration with the VA can accelerate improvement bi-directionally (non-VA VA)

Next Up: Eye Exams in Diabetes Best Practice - Courtesy of Our VA Colleagues Top 30 of 74 Clinics VAs : 13 of Top 15!

Q & A THANK YOU 26