Quality and Variation in Medical Practice: Why are Doctors so Different? Objectives List 3 examples of significant variation in pediatric practice Describe the relationship between variation and quality of a process Describe one method of improving quality of care by addressing variation Mark W. Shen, M.D. November 19, 2010 A Mad Lib 1. Pick a pediatric practitioner Generalist, pulmonologist, rheumatologist, hematologist, neonatologist, ID, GI, ENT 2. Pick a management scenario: ITP, post-op T&A, HSP, bronchiolitis, protein-losing enteropathy, post-op cardiac surgery, bacterial meningitis 3. Pick a word pair: Given & stop OR Not given & start Your Mad Lib You are a pediatric [insert type of practitioner] and begin your busy Monday by seeing a patient cared for by your partner over the weekend The patient has been receiving care for [insert disease] and was [given/not given] steroids. You completely disagree with this approach and [stop/start] the medication. What a frustrating start to the day. Sometimes you wonder how 2 physicians could practice such different medicine. As you leave the room, you notice a look of puzzlement on the family s faces Encountering Variation: The 5 Stages of Grief Denial Is that person board-certified? Anger It s my patient, I can do what I want Bargaining Let me try to use the family to get my way (I ll tell them my side) Depression I m an accomplice in providing poor care Acceptance Just do whatever the other MD wants The History of (the study of) Medical Variation
Variation in Incidence of Tonsillectomy: J Alison Glover Puzzling as is the geographical distribution, the social distribution is yet more of an enigma. Tonsillectomy is at least three times as common in the well-to-do classes. Tonsillectomy Variation: Back Across the Pond 1934, American Child Health Association 1000 New York City School Children 40% had not yet undergone tonsillectomy School physicians: 45% needed an operation Of those not selected, another group of physicians recommended that 46% receive tonsillectomy Of the twice-rejected children, a third group of physicians recommended operation in 44% After 3 exams, only 65 children remained Int J Epidemiol 2008;37:9 19 Sci Am 1982;246:120-34 The Beginnings of Modern Day Variations Research Extreme Variation in Tonsillectomy Rates John Wennberg s House Science 1973;182:1102-1108 Probability of Having Surgery in 11 Vermont Hospitals Surgical Rates for the Most Populous Hospital Areas: Maine Sci Am 1982;246:120-34 Int J Epidemiol 2008;37:26 29
The Surgical Signature International Differences in Surgical Rates Int J Epidemiol 2008;37:26 29 NEJM 1982;307:1310-14 Proof of Preference- Sensitive Care Preference-Sensitive Care aka: Sci Am 1982;246:120-34 Terminology Medical Variation: The Present Unwarranted Variation: Care that is not consistent with a patient s preference or related to their underlying illness Preference-Sensitive Care No right rate (T&A) Misuse Effective Care Evidence-based care not provided Underuse
Evidence for Underuse of Effective Care: Adults Evidence for Underuse of Effective Care: Children Terminology Unwarranted Variation: Care that is not consistent with a patient s preference or related to their underlying illness Preference-Sensitive Care (Misuse) No right rate Effective Care (Underuse) Evidence-based care not provided Supply-Sensitive Care (Overuse) Systems supply creates demand The Dartmouth Atlas: Medicare Spending Varies Dramatically http://www.dartmouthatlas.org/downloads/reports/spending_brief_022709.pdf Unwarranted Variation in Medicare Spending Preference Sensitive Care Effective Care Supply Sensitive Care 12% 25% 63% Source: John E. Wennberg and Dartmouth Atlas http://www.ahrq.gov/about/annualconf09/brownlee.htm
A Trending Topic Trending in Pediatrics www.nejm.org www.pediatrics.org Examples of Data Sources in Pediatrics Databases PHIS (Pediatric Health Information Systems) Propietary administrative database Maintained by Child Health Corporation of America (CHCA) Business Alliance of 42 children s hospitals Collaborative Networks VIP (Value in Inpatient Pediatrics) Grassroots collaborative improvement network Data: administrative and chart review Converted to oral therapy (%) 100 90 80 70 60 50 40 30 20 10 0 Osteomyelitis: Variability in Early Conversion to Oral Therapy 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Hospital Pediatrics 2009;123;636-642 UTI in Infants: Variability in Length of IV Therapy Corticosteroid Use After Congenital Heart Surgery Pediatrics 2010;126;196-203 Circulation. 2010 Nov 8. [Epub ahead of print]
Bronchodilator Doses Per Patient in Acute Bronchiolitis I See Variation Center Source: VIP Network Unwarranted Variation: The Losers Isn t Variety the Spice of Life? Patients Learners System (everyone loses) Preference-Sensitive Losers: The Patients Patients lose when not involved Recent Dell Children s patient comments: Doctors, deliver a consistent message. We heard different plans from different doctors. Lack of communication between doctors I was given conflicting info, on which I had to make a judgment call. I didn t know who to talk to. Too many doctors involved Preference-Sensitive Losers: The Learners What do you want to do? (everyone does things differently so just tell me what you want to do) I don t care
Adverse Effects of Unmeasured Variation Adverse Systems Effects of Unwarranted Variation Medicare: Cost vs Quality The Value Equation Quality Value = --------------------- Cost Variation in Annual Total Cost and Quality for Chronic Disease Beyond Just The Numbers Average Quality of Care Score 1.60 1.40 1.20 1.00 0.80 0.60 0.40 0.20 A Greenville, NC B Saginaw, MI Best Practice Curve Ft. Lauderdale, FL East Long Island, NY Orange County, CA Boston, MA C Melrose Park, IL Manhattan, NY D Newark, NJ 0.00 $- $10,000 $20,000 $30,000 $40,000 $50,000 $60,000 $70,000 $80,000 $90,000 Average Annual Reimbursement per Beneficiary (Wage-Index Adjusted) * Based on percent of beneficiaries with three conditions (diabetes, chronic obstructive pulmonary disease, and congestive heart failure) who had a doctor s visit four weeks after hospitalization, a doctor s visit every six months, annual cholesterol test, annual flu shot, annual eye exam, annual HbA1C test, and annual nephrology test. Source: G. Anderson and R. Herbert for The Commonwealth Fund, Medicare Standard Analytical File 5% 2001 data.
McAllen vs El Paso: Medicare Spending http://www.whitehouse.gov/omb/blog/09/06/04/mcallenredux/ http://www.newyorker.com/reporting/2009/06/01/090601fa_fact_gawande?currentpage=all CABG Rates in California To Vary is Human http://www.managedcaremag.com/archives/0311/0311.variation.html Public Perception: CABG Rates, Redding vs CA Doctors Decisions and Impact on Medical Care
Lenses Under Which to Analyze Doctors Decisions Uncertainty & limits of the human brain Medical decision-making, clinical problemsolving Different Disciplines Clinical, economic, sociological, psychological Components Patient, physician, system We Are Surrounded By Uncertainty Defining a Disease Making a Diagnosis Selecting a Procedure (e.g., test or intervention) Observing Outcomes Assessing Preferences Health Affairs 1984;3:74-89 Colorectal Experts: Consensus??? Question: What is the effect of screening annual fecal occult blood and flexible scope on colorectal cancer? Eisenberg: Determinants of Medical Decision-Making Physician as a self-fulfilling practitioner Physician as patient s agent Physician as guarantor of social good Health Affairs 1984;3:74 Doctors Decisions and the Cost of Medical Care. Michigan, 1986 Eisenberg: Determinants of Medical Decision-Making Older Doctors Use Fewer Laboratory Tests Physician as a self-fulfilling practitioner 1. Desire for income 2. Desire for a style of practice 3. Personal characteristics 4. Practice setting 5. Standards established by clinical leadership Doctors Decisions and the Cost of Medical Care. Michigan, 1986 Medical Care 1981;19:297-309
Eisenberg: Determinants of Medical Decision-Making Physician as a self-fulfilling practitioner 1. Desire for income 2. Desire for a style of practice 3. Personal characteristics 4. Practice setting 5. Standards established by clinical leadership David Eddy on Practice Setting s Impact on Variation This tendency to follow the pack is the most important single explanation of regional variations in medical practice. If uncertainty caused individual physicians to practice at random, or to follow their personal interpretations and values, without any attempts to match the actions of their neighbors, the variations in practice patterns would average out, and no significant differences would be observed at the regional level. Differences between regions are observed because individual physicians tend to follow what is considered standard and accepted in the community. Doctors Decisions and the Cost of Medical Care. Michigan, 1986 Health Affairs 1984;3:74 Eisenberg: Determinants of Medical Decision-Making Physician as patient s agent 1. Economic agent 2. Clinical agent 3. Patient demand 4. Defensive medicine 5. Patient characteristics 6. Convenience Eisenberg: Determinants of Medical Decision-Making Physician as guarantor of social good Duty to the patient vs steward of resources Tension between the prisoner s dilemma and the tragedy of the commons Classic scenario: end-of-life care (flat of the curve medicine) Doctors Decisions and the Cost of Medical Care. Michigan, 1986 Doctors Decisions and the Cost of Medical Care. Michigan, 1986 Does Genotype Determine Medical Decision-Making? Knowledge is Paralyzing
Learn from Patients Paternalistic Craft-based Silos are archaic How Do We Improve? Learn from Improvement Science 1. Measure the process 2. Analyze the data 3. Intervene: Control the process (Research is a Slightly Different Order) 1. Measure the process 2. Control everything 3. Intervene 4. Analyze the data Learn from Improvement Science 1. Measure the process 2. Analyze the data 3. Intervene: Control the process Control unwarranted variation through standardization Continue to measure and analyze 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 11/15/2009 11/22/2009 Quality Improvement in Action: Control Chart Step 2: Analyze 11/29/2009 12/6/2009 12/13/2009 12/20/2009 12/27/2009 Step 1: Measure Step 3: Intervene - Standardize 1/3/2010 1/10/2010 1/17/2010 1/24/2010 1/31/2010 2/7/2010 2/14/2010 2/21/2010 2/28/2010 3/7/2010 3/14/2010 3/21/2010 3/28/2010 4/4/2010
Pediatric Oncology Learn from High Performers Minimal unwarranted variation All variation is measured & patient-level Enormous success Pediatric cancer transformed from uniformly fatal disease in 1950s to 78% five-year survival for all types Better outcomes than adult groups for adolescents and young adults (AYA) Pediatric vs Adult Trials in AYA with ALL Reasons Better compliance on pediatric protocols Better enrollment in pediatric trials Next Steps Enrolling adults in trials with pediatric protocols Blood 2008;112:1646-1654 If to vary is human, then only through collaboration will we truly divine