Reducing Regional Disparities in Health Spending: Framing the Debate

1
Reducing Regional Disparities in Health Spending
Framing the Debate
David Wennberg and Friends Maine Medical Center
Center for the Evaluative Clinical Sciences
2
Regional disparities in health care spending

• Part 1 -- Unwarranted variations in U.S. health
  care findings from the Is More Better?
  studies
• Part 2 -- What can be done about it?

3
Elliott Fisher, MD, MPH Therese Stukel, PhD Dan
Gottlieb, MSF. L. Lucas, PhD Etoile Pinder, MS
4
Unwarranted variations in medical practice a
framework for thinking about the delivery (or
non-delivery) of care

• Unwarranted? Variations that cannot be
  explained by
• Illness or need --- and dictates of evidence
  based medicine
• Patient Preferences
• Categories of variation
• Effective care
• Preference sensitive care
• Supply-sensitive services
• Causes and remedies differ for each category

5
Dartmouth Atlas of Health CareUnited States
Hospital Referral Regions
6
Elderly (U.S. Medicare) Study Design
Myocardial Infarction Colorectal Cancer Hip
Fracture Medicare Population (MCBS)
Step 1 Select Cohorts
Step 2 Group by regional spending level --
assigned based upon End-of-Life Expenditure Index
Step 3 Validation (1) are patients the same at
baseline? (2) does subsequent treatment differ?
Process / Quality of Care / Survival
Step 4 Assess outcomes Follow cohorts for up to
five years.
7
Regional Variations in the End-of-Life
Expenditure Index (EOL-EI) and average
per-capita Medicare spending
Spending
EOL-EI
3,922
9,074
4,439
10,636
4,940
11,559
5,444
12,598
6,304
14,644
EOL-EI highly correlated (r 0.81) with average
per-capita Medicare spending
8
Effective Care

• Services of proven effectiveness.
• It involves no significant tradeoffs--all with
  specific needs should receive them
• Conflict between patients and providers is
  minimal

9
Effective Care Ratio of Rates in Highest vs
Lowest Spending Regions
Acute MI
Reperfusion in 12 hours for AMI
Aspirin at admission
Beta Blockers at admission
Aspirin at Discharge
Beta Blockers at discharge
Exercise Test w/in 30 d
Lower in High Spending Regions
Higher in High Spending Regions
10
Preference-Sensitive Care

• Involves tradeoffs among outcomes
• Decision should reflect preferences of patient
• Scientific uncertainty often substantial

11
Preference-Sensitive Care Highest vs Lowest
Spending Regions
Procedures after AMI
Angiography
Angiography among appropriate cases
Coronary Angioplasty
Coronary Artery Bypass Surgery (CABG)
Major Surgery (all cohorts combined)
Cholecystectomy
Cataract Extraction
Hernia Repair
Total Hip Replacement
Total Knee Replacement
Back Surgery
Carotid Endarterectomy
Lower in High Spending Regions
Higher in High Spending Regions
12
Supply Sensitive Services

• Care strongly correlated with supply
• Generally provided in absence of strong clinical
  theory
• Evidence weak or non-existent on benefits.

13
Supply-Sensitive Care Highest vs Lowest
Spending Regions
Diagnostic Cardiology Procedures
Electrocardiogram
Echocardiogram
Ambulatory ECG (Holter)
Lower in High Spending Regions
Higher in High Spending Regions
14
Supply-Sensitive Care Highest vs Lowest
Spending Regions
Lower in High Spending Regions
Higher in High Spending Regions
15
FindingsMortality
16
Relative Risk of Death across Quintiles of
Spending
Decreased Risk
Increased Risk
ColorectalCancer
Q1
Q2
Q3
Q4
Q5
MyocardialInfarction
Q1
Q2
Q3
Q4
Q5
1.00
1.05
1.10
0.95
17
Change in relative risk of death per 10
increment in regional practice intensity Acute
Myocardial Infarction Cohort
Decreased Risk
Increased Risk
Age lt 80
Age gt 80
Female
Male
Black
Non-black
Non-Q MI
Anterior MI
Inferior MI
Other location
Low risk (lt15 1yr)
Moderate (15-30)
High Risk (gt 30)
1.00
1.02
1.04
0.98
18
Summary of Findings

• Increased spending across regions is largely
  devoted to supply-sensitive services
• Visit frequency, specialist services, tests,
  inpatient and ICU care.
• Residents of higher spending regions
• Slightly worse basic access to care
• Equal use of major (potentially beneficial)
  procedures
• Quality measures generally somewhat worse
• No gain in function, survival or satisfaction

19
Implications

• Costs reflect the capacity of the system

20
Spending and capacity the role of beds and
medical specialists
High MDHigh Bed
1.35
1.34
High MDLow Bed
Low MDHigh Bed
1.59
1.18
Low MDLow Bed
21
Implications

• Costs reflect the capacity of the system
• Greater capacity is not necessarily better

22
Implications

• Costs reflect the capacity of the system
• Greater capacity is not necessarily better
• Were wasting 30 of current spending on supply
  sensitive care alone

23
Regional disparities in health care spending

• Part 1 -- Unwarranted variations in U.S. health
  care findings from the Is More Better?
  studies
• Part 2 -- What can be done about it?

24
Principles to Guide Interventions
Variation
Cause
Remedy
Effective Care
Poorly understood care processes Failure to
learn
Develop systems of care capable of
improvement Reward those who provide high quality
care Construct benefits to incent beneficiaries
to become active consumers and to seek high
quality providers
25
Principles to Guide Interventions
Variation
Cause
Remedy
26
Principles to Guide Interventions
Variation
Cause
Remedy
27
Regional disparities in health care spending

• Part 1 -- Unwarranted variations in U.S. health
  care findings from the Is More Better?
  studies
• Part 2 -- What can be done about it?