Art and Science of Medicine Meets Quality improvement

Art and Science of Medicine Meets Quality improvement Elliott Main, MD Medical Director, CMQCC main@cmqcc.org Clinical Professor, Depts of OB/GYN UCSF, and Stanford University

CPQCC and CMQCC California Perinatal Quality Care Collaborative (CPQCC) Multi-stakeholder organization established in 1996 (providers, state agencies, public groups like MOD) Established Perinatal Data Center in 1996, works with VON Data submission for VON Plus data system with 131 out of 136 NICUs with >17,000 annual admissions; Over 10 quality toolkits and related collaboratives Model of working with state agencies to provide data of value California Maternal Quality Care Collaborative (CMQCC) Expertise in maternal data analysis, MMR (2006) Developer of QI toolkits: Early Elective Delivery, OB Hemorrhage, Preeclampsia, CVD in Pregnancy, and First Cesarean Prevention Host of collaborative learning sessions Established Maternal Data Center in 2011

CMQCC Key Partner/Stakeholders State Agencies: MCAH, Dept Public Health OSHPD Healthcare Information Division Office of Vital Records (OVR) Regional Perinatal Programs of California (RPPC) DHCS, Medi-Cal Public Groups California Hospital Accountability and Reporting Taskforce (CHART) California HealthCare Foundation Kaiser Family Foundation March of Dimes (MOD) Professional groups American College of Obstetrics and Gynecology (ACOG) Association of Women s Health, Obstetric and Neonatal Nurses (AWHONN) American College of Nurse Midwives (ACNM), American Academy of Family Physicians (AAFP) Key Medical and Nursing Leaders Universities and Hospital Systems Kaisers, Sutter, Sharp, Dignity, Scripps, Providence, Public hospitals,

CMQCC Key Partner/Stakeholders (con t) Medical Associations: California Hospital Association Regional Hospital Associations California Medical Association Payers Aetna Anthem Blue Cross Blue Shield Cigna Health Net Purchasers CALPERS (State and local government employees and retirees) Medi-Cal (for managed care plans) Pacific Business Group on Health/ Silicon Valley Employers Forum Cover California (ACA entity)

CMQCC: Major Areas of Activity Maternal Mortality and Morbidity Reduction Large- Scale Implementation Maternal Data Center Maternity Quality Measures 5

12- Step Program for Quality Improvement 1) Memory v. Data 2) Defenses 3) Burning Platform 4) External Measures 5) Variation 6) Autonomy 7) Translation 8) Pressures 9) Small Risks 10) Culture 11) Normalization 12) Just Do It : Transforming Maternity Care

1. Memory-Driven vs. Data-Driven QI : Transforming Maternity Care

Limitations of Memory-guided Practice Hard to remember beyond the last 10 cases and the last terrible outcome No denominators, no sense of rates Driven by anecdote and local custom Advantages of Data-driven Practice Full knowledge of rates and outcomes, of yourself, your unit and the state Not overly influenced by isolated cases Driven by evidencebased medicine and basic standards : Transforming Maternity Care

Elimination of Non-medically Indicated (Elective) Deliveries Prior to 39 Weeks Funding Federal Title V block grant from the California Department of Public Health; Maternal, Child and Adolescent Health Division California Maternal Quality Care Collaborative March of Dimes

Percent Affected Adverse Neonatal Outcomes According to Completed Week of Gestation at Delivery: Absolute Risk 16% 14% 12% 37+ Weeks 38+ Weeks 39+ Weeks 10% 8% 6% 4% 2% 0% Any adverse Adverse outcome or death respiratory outcome(overall) RDS TTN Admission to NICU Newborn Sepsis (suspected or proven) Tita AT, et al, NEJM 2009;360:111

Odds Ratios Adverse Neonatal Outcomes According to Completed Week of Gestation at Delivery: Odds Ratios 4.5 4 3.5 37+ Weeks 38+ Weeks 39+ Weeks 3 2.5 2 1.5 1 0.5 0 Any adverse Adverse outcome or death respiratory outcome(overall) RDS TTN Admission to NICU Newborn Sepsis (suspected or proven) Treated Hospitalization > hypoglycemia 5 days Tita AT, et al, NEJM 2009;360:111

Mean EED Rate in California (%) Na onal and California Reduc on of Early Elec ve Deliveries (EED) 15 2010 10 5 >75% Reduction 0 2010 (Baseline, MOD) 2011 (Baseline, LFG) 2014 (Leap Frog Group) 2013 (Joint Commission) 2013 (CMS) 2014 Nationally, CMS estimates an 80% reduction in EED

EED Success: Collective Impact Prof Orgs (Natl and Local) Public Policy Public advocates Performance measures Data-driven QI EED Public Reporting Evidence 70-80% Reduction Nationally! Payment Incentives

2. Recognize the Defenses : Transforming Maternity Care

Self-Defense Manual for Medical Professionals 1970, 1980, 1990, 2000, 2010. : Transforming Maternity Care

Table of Contents: The best defense is a good offense. Chapter One: Attack the Data Chapter Two: Attack the Messenger Chapter Three: Attack the Premise When all else fails, there is always. Chapter Four: My Patients are Higher Risk : Transforming Maternity Care

Appendix: Counter Strategies Data: Clean carefully before presentation- Be very certain about case attribution Example: every obstetrician who covers midwives or FP s will have higher CS rates Premise: Good to have backing of national organization(s) Risk Adjustment: simple strategies best- Risk stratification v. logistic regression Process measures do not need risk adjustment! : Transforming Maternity Care

CMQCC Maternal Data Center PDD Discharge Diagnosis File (ICD9/10 Codes) Monthly uploads: mother and infant PDD (participating hospitals) Automated Linkage of all 3 files Birth Certificate (Clinical Data) Monthly uploads: electronic files for all CA births Chart Review (select metrics/qi projects) Maternal Data Center Limited manual data entry for these measures Interactive Analytics Guide QI Practice Links over 1,000,000 mother/baby records each year!

3. Build the Burning Platform : Transforming Maternity Care

Kotter s Eight Steps of Change

Rate per 100,000 Live Births Number of Maternal Deaths Maternal Mortality Rate California Residents and United States: 1991-2006 18 16.9 100 16 ICD-9 ICD-10 90 14 12 10 8 6 10.7 7.9 7.8 9.1 10.9 9.9 9.7 9.1 9.9 7.7 8.3 8.5 9.8 9.7 7.1 7.7 8.4 7.5 7.6 6.7 7.1 5.6 14.6 15.1 13.3 13.1 11.7 10 12.1 11.7 California Rate 8.9 United States Rate HP2010 Objective 80 70 60 50 40 30 4 2 0 HP 2010 Objective 4.3 Deaths per 100,000 Live Births 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 20 10 0 SOURCE: State of California, Department of Public Health, California Birth and Death Statistical Master Files, 1991-2006. Maternal mortality for California (deaths 42 days postpartum) calculated using ICD-9 cause of death classification (codes 630-638, 640-648, 650-676 ) for 1991-1998 and ICD-10 cause of death classification (codes A34, O00-O95,O98-O99) for 1999-2006. United States data and HP2010 Objective were calculated using the same methods. The break in the trend line represents the change from ICD-9 to ICD-10. Produced by California Department of Public Health, Maternal, Child and Adolescent Health Program, June 2009. : Transforming Maternity Care

THE CALIFORNIA PREGNANCY-ASSOCIATED MORTALITY REVIEW Report from 2002-2003 Maternal Death Reviews This project was supported by federal Title V block grant funds received from the California Department of Public Health; Center for Family Health; Maternal, Child and Adolescent Health Division

Causes of Maternal Mortality and Morbidity Cause Mort. ICU Serious Morbid VTE and AFE 10% 5% 2% Infection 15% 10% 5% Hemorrhage 15% 30% 45% Preeclampsia / CVA 20% 30% 30% Cardiac Disease 20% 15% 10% : Transforming Maternity Care

CMQCC Hemorrhage Task Force: 5 meetings in 2008-2009 Developed a Tool Kit for OB services: Set of Best Practices (short summaries of key aspects of OB hemorrhage) Checklist for managing OB hemorrhage Flow-Chart and Table Chart Summaries of approach Implementation tools such as sample policies, procedures, charting examples, implementation hints All resources on-line at: www.cmqcc.org/ob_hemorrhage CMQCC has sponsored an IHI-like Learning Collaboratives to help sites implement : Transforming Maternity Care

CMQCC California OB Hemorrhage Project 26 Hemorrhage Taskforce (2008-2009) QI Toolkit/Best Practices CHW QI Project (2009) 1 st CMQCC Statewide Collaborative (2009-2010) 30 hospitals (110,000 annual births) Large/small, urban/rural New CMQCC Collaboratives (2011) Statewide: 20+ hospitals (still open) LA County: 11 hospitals Systems: Kaiser North and South; Sutter Enhanced Website resources

Obstetrics & Gynecology April 2015 Pregnancy-related mortality should not be considered a single clinical entity. The five leading causes exhibit different characteristics, degrees of preventability, and contributing factors, with the greatest improvement opportunities identified for hemorrhage and preeclampsia.

Provider Contributing Factors in Maternal Deaths (California) From detailed chart reviews of maternal deaths (CA-Pregnancy Associated Mortality Review Committee; CDPH-MCAH) Main EK, McClain CL, Morton CH, Holtby S, Lawton ES. Pregnancy-related mortality in California: Causes, characteristics and improvement opportunities. Obstet Gynecol 2015

California Approach to Reduce Maternal Mortality and SMM Hemorrhage Taskforce (2009) Hemorrhage QI Toolkit (2010) Multi-hospital QI Collaborative(s) (2010-11) Test the tools and implementation strategies State-wide Implementation (2013-2014) Preeclampsia Taskforce (2012) Preeclampsia QI Toolkit (2013) Multi-hospital QI Collaborative (2013-2014) Cardiovascular Detailed Case Analysis (2013) Cardiovascular QI Toolkit (2015)

Maternal Deaths per 100,000 Live Births Maternal Mortality Rate, California and United States; 1999-2013 24.0 21.0 California Rate 19.3 22.0 18.0 15.0 12.0 9.0 6.0 9.9 7.7 10.9 9.8 United States Rate 9.9 9.7 10.0 8.9 14.6 12.1 13.1 15.1 11.8 11.7 16.9 13.3 12.7 11.1 15.5 16.6 16.9 14.0 11.6 9.2 7.4 19.9 6.2 7.3 3.0 HP 2020 Objective 11.4 Deaths per 100,000 Live Births 0.0 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Year SOURCE: State of California, Department of Public Health, California Birth and Death Statistical Master Files, 1999-2013. Maternal mortality for California (deaths 42 days postpartum) was calculated using ICD-10 cause of death classification (codes A34, O00-O95,O98-O99). United States data and HP2020 Objective use the same codes. U.S. maternal mortality data is published by the National Center for Health Statistics (NCHS) through 2007 only. U.S. maternal mortality rates from 2008 through-2013 were calculated using CDC Wonder Online Database, accessed at http://wonder.cdc.govon March 11, 2015. Produced by California Department of Public Health, Center for Family Health, Maternal, Child and Adolescent Health Division, March, 2015.

National Partnership for Maternal Safety: 3 Maternal Safety Bundles in 3 Years What every birthing facility in the US should have Obstetric Hemorrhage Preeclampsia/ Hypertension Prevention of VTE in Pregnancy Note: The bundles represent outlines of highly recommended protocols and materials important to safe care BUT the specific contents and protocols should be individualized to meet local capabilities. Example materials are available from perinatal collabortives and other organizations. 31

5 Key Complimentary Strategies: 1) QI projects for labor management practices 2) Payment reform to eliminate negative or perverse incentives 3) Education for the value of normal birth (culture) 4) Transparency with Public Reporting 5) Continued public engagement Main EK et al. Obstet Gynecol Nov 2012;120(5):1194 1198.

5. Variation Reflects Opportunity : Transforming Maternity Care

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101 106 111 116 121 126 131 136 141 146 151 156 161 166 171 176 181 186 191 196 201 206 211 216 221 226 231 236 241 246 251 80% 70% Large Variation of Total CS Rate Among 251 California Hospitals: 2013 60% 50% 40% Range: 15.0 71.4% Median: 32.5% Mean: 32.8% 30% 20% 10% Will this degree of variation remain after risk adjustment? 0% Hospitals

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101 106 111 116 121 126 131 136 141 146 151 156 161 166 171 176 181 186 191 196 201 206 211 216 221 226 231 236 241 246 80% 70% Even Larger Variation of NTSV CS Rate Among 251 California Hospitals: 2013 60% 50% 40% 30% Range: 10.0 75.8% Median: 27.0% Mean: 27.7% National Target =23.9% 20% 10% 0% 36% of CA hospitals meet national target Large Variation = Improvement Opportunity Hospitals

CHCF Infographic Released November 2014 CHCF reports over 11,000 page views in first week and very positive feedback calqualitycare.org

New National Guidelines for Defining Labor Abnormalities and Management Options

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101 106 111 116 121 126 131 136 141 146 151 156 161 166 171 176 181 186 191 196 201 206 211 216 221 226 231 236 241 246 80% 70% Even Larger Variation of NTSV CS Rate Among 251 California Hospitals: 2013 60% 50% 40% 30% Range: 10.0 75.8% Median: 27.0% Mean: 27.7% National Target =23.9% 3 Pilot Hospitals for Interventions 20% 10% 0% 36% of CA hospitals meet national target Large Variation = Improvement Opportunity Hospitals

This is the same Orange County as depicted in the popular television show. This is the hospital where most of these mothers deliver Not the easiest population to start with

3 Major Drivers of the Primary CS Rate

3 Major Drivers of the NTSV CS Rate

Provider-Level Cesarean Rates G5xxxx G6xxxx G7xxxx G8xxxx A8xxxx A6xxxx A5xxxx A4xxxx A8xxxx A9xxxx

Data-Driven QI: NTSV CS Pilot Hospital: Orange County 35% 33% 32.9% 33.6% 31.2% 31.8% NTSV CS Rate 30% 28% 25% 23% 20% 18% 15% National Target for NTSV CS = 23.9% 2011 2012 2013 Jan-14 Feb-14 Mar-14 Apr-14 Apr14 May-14 14 43

Data-Driven QI: NTSV CS Pilot Hospital: Orange County 35% 33% 32.9% 33.6% 31.2% 31.8% NTSV CS Rate 30% 28% 25% 23% 20% QI Project Started: Jan 16 28.3% 18% 15% National Target for NTSV CS = 23.9% 2011 2012 2013 Jan-14 Feb-14 Mar-14 Apr-14 Apr14 May-14 14 44

CMQCC Data-Driven QI: NTSV CS Pilot Hospital: Orange County 35% 33% 32.9% 33.6% 31.2% 31.8% NTSV CS Rate 30% 28.3% 28% 25% 23% QI Project Started: Jan 16 24.3% 20% 18% 15% National Target for NTSV CS = 23.9% 2011 2012 2013 Jan-14 Feb-14 Mar-14 Apr-14 Apr14 May-14 14 45

Data-Driven QI: NTSV CS Pilot Hospital: Orange County 35% 33% 32.9% 33.6% 31.2% 31.8% NTSV CS Rate 30% 28.3% 28% 25% 23% QI Project Started: Jan 16 24.3% 25.0% 20% 18% 15% National Target for NTSV CS = 23.9% 2011 2012 2013 Jan-14 Feb-14 Mar-14 Apr-14 Apr14 May-14 14 46

Data-Driven QI: NTSV CS Pilot Hospital: Orange County 35% 33% 32.9% 33.6% 31.2% 31.8% NTSV CS Rate 30% 28.3% 28% 25% 23% QI Project Started: Jan 16 24.3% 25.0% 23.4% 20% 18% 15% National Target for NTSV CS = 23.9% 2011 2012 2013 Jan-14 Feb-14 Mar-14 Apr-14 Apr14 May-14 14 47

No Change in Baby Outcomes: Rate of Unexpected Newborn Complications Hoag Hospital Intervention Period Dec - Feb 2015

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101 106 111 116 121 126 131 136 141 146 151 156 161 166 171 176 181 186 191 196 201 206 211 216 221 226 231 236 241 246 80% 70% 60% Low-Risk First-Birth (Nuliparous Term Singleton Vertex) CS Rate (endorsed by NQF, TJC PC-02, CMS, HP2020) Among 249 California Hospitals: 2011-2012 (Source: CMQCC--California Maternal Data Center combining primary data from OSHPD and Vital Records) 50% Hoag Hospital 40% 30% Range: 10.0 75.8% Median: 27.0% Mean: 27.7% National Target =23.9% 20% 10% 36% of CA hospitals meet national target For the last 3 mos, the rate was 22.5% 0% July 24, 2013 49

Collaborative Action: Collective Impact Clinical Leaders Datadriven QI Public Reporting Public Policy Quality measures NTSV CS Public advocates Strong Evidence Payment Incentives Multiple Pressure Points are much more effective than one or two alone

Thank you, from all of us at CMQCC! Elliott Main, MD David Lagrew, MD Kathryn Melsop, MS Christine Morton PhD Anisha Abreo, MPH Andrew Carpenter Jeffrey Gould, MD MPH Barbara Murphy, RN MSN Julie Vasher, DPN, CNS Nancy Peterson, RN MS Valerie Cape Allana Moore