Developing Cost Effective CHD Screening Strategies

1
Developing Cost Effective CHD Screening
Strategies

• Leslee J. Shaw, PhD
• Department of Imaging and Medicine
• Cedars-Sinai Medical Center
• Los Angeles, California

2
CHD Detection In Asymptomatic Women Men

• Traditional approach to detection of CHD risk
  assessment of typical risk factors
• Despite many available risk assessment
  approaches, theres a detection gap for
  asymptomatic individuals w/ subclinical
  atherosclerosis.
• Framingham European risk scores - useful
  guides.
• to predict long term risk of CHD events in
  healthy populations.
• Target Population for Screening
• 40 of the US Adult Population (or 36 million)
  Intermediate Risk
• Majority of 1st MIs

Source Abrams, Pasternak, Greenland,
Houston-Miller, Smaha. BC 34 Taskforce 1 -
Identification of CHD and CHD Risk. JACC 2003.,
Blumenthal, Becker, Yanek, Aversano, Moy, Kral,
Becker. Detecting occult coronary disease in a
high-risk asymptomatic population. Circulation
2003107(5)702-707., Wilson, DAgostino, Levy,
Belanger, Silbershatz, Kannel. Prediction of CHD
using risk factor categories. Circulation
1998971837-1847.
3
X
Source Fletcher et al., 33rd Bethesda Conf
Preventive Cardiology How Can We Do Better? JACC
2002404579-651., Wilson et al. Abdominal
aortic calcific deposits are an important
predictor of vascular morbidity and mortality.
Circulation 20011031529-34., Jaffer et al. Age
and Sex Distribution of Subclinical Aortic
Atherosclerosis - A Magnetic Resonance Imaging
Examination of the Framingham Heart Study Art,
Thromb, Vasc Biol 200222849.
4
Estimated 10 Yr. Hard CHD Risk Framingham
Offspring Cohort Women and Men
Percent
Women
Men
Age (years)
Source Abrams, Pasternak, Greenland,
Houston-Miller, Smaha. Bethesda Conference 34
Identification of CHD and CHD risk Is there a
detection gap? JACC 2003
5
Not Qualifying For Pharmacotherapy by CACS
Women as well as young individuals were less
likely to be considered candidates for
pharmacotherapy vs. men older individuals.
Shaw Atherosclerosis (in press) - 45 low risk
reclassified based on CAC
Source Nasir K, Michos ED, Blumenthal RS, Raggi
P. Detection of High-Risk Young Adults and Women
by Coronary Calcium and National Cholesterol
Education Panel-III Guidelines. JACC 2005 (in
press).
6
Estimated Direct Indirect Costs of
Cardiovascular Diseases Stroke United States
2005
Source Heart Disease and Stroke Statistics
2005 Update.
7
Current State of Health Care System

• 50 of health care costs are for end-stage or
  hospital care.
• Avg yrly health expenditure for end stage care is
  5-x higher vs. non-end stage care.
• Shifting care to early, diagnostic or outpatient
  sector potential to reduce cost.

412 Billion Medicare pays 31
286 Billion Medicare pays 21
122 Billion Medicare pays 2
92 Billion Medicare pays 10
60 Billion Medicare pays 0
39 Billion Medicare pays 12
37 Billion Medicare pays 0
32 Billion Medicare pays 29
31 Billion Medicare pays 4
19 Billion Medicare pays 25
Source CMS, Office of the Actuary, National
Health Statistics Group. Access date March 2,
2004.
8
Medicare Spending
- 2/3rds of Spending 5 Chronic Conditions -
1/5th of Spending 3 Chronic Conditions
Source Medicare Standard Analytic File, 1999.
9
The Most Expensive Conditions In America MEPS
Population Estimates
Billion Billion 1. Ischemic
Heart Disease 21.5 9. Cerebrovascular Dz
8.3 2. Motor Vehicle Accidents 21.3 10.
Dysrythmias 7.2 3. Acute Resp.
Infections 17.9 11. Peripheral Vascular
6.8 4. Arthropathies 15.9 12. COPD
6.4 5. Hypertension 14.8 13.
Asthma 5.7 6. Back Problems 12.2
14. CHF 5.2 7.
Mood Disorders 10.2 15. Lung Cancer
5.0 8. Diabetes 10.1
10
The Most Expensive Conditions In America MEPS
Population Estimates
Billion Billion 1. Ischemic
Heart Disease 21.5 9. Cerebrovascular Dz
8.3 2. Motor Vehicle Accidents 21.3 10.
Dysrythmias 7.2 3. Acute Resp.
Infections 17.9 11. Peripheral Vascular
6.8 4. Arthropathies 15.9 12. COPD
6.4 5. Hypertension 14.8 13.
Asthma 5.7 6. Back Problems 12.2
14. CHF 5.2 7.
Mood Disorders 10.2 15. Lung Cancer
5.0 8. Diabetes 10.1
11
Upfront Test Cost
Affected by MD Labor, Lab Volume, /- Add-Ons
(Contrast or Radiopharmaceutical), Equipment
(Lease, Age, Shared)
Low Cost Lab / Office Visit
Cardiac Imaging
Source Mark DB, Shaw LJ, et al. Bethesda
Conference 34- Taskforce 5 - Is atherosclerotic
imaging cost effective? JACC 2003411906.
12
Average Cost Inputs for Adverse Sequelae of CVD

• Out-of-Hospital SCD Lost Productivity
• In-Hospital Death in excess of 50k-100k
• End-Stage Care for CHF 80 of lifetime care
  costs
• AMI or ACS ? 15-20k
• Chest Pain Hospitalization ? 6k
• Stroke ? 50k
• Anti-Ischemic Rx ? 1,500 - 5,000 / yr
• Out-of-Pocket ? 2,000 / yr
• .

13
Medicare Payment Advisory Commission (MedPAC) -
Growth in Physician Services
Growth of All Physician Services
22

Includes all Services in the Physician Fee
Schedule Source MEDPAC Analysis of Medicare
Claims Data March 17, 2005, Executive Director,
Medicare Payment Advisory Commission, Mark
Miller,.htm
14
Trends in CV Operations Procedures United
States 1979-2000
15
Unfolding a Body of Evidence
Building
Building

• Observational
• Data
• Risk identification
• Costs

• Cost Effectiveness
• High Risk CEA
• Reimbursement

• Clinical Trial
• Data
• Vs. Comparators

• Disease Management
• Risk Identification
• Cost Efficiency
• Outcomes Improve Process of Care

• Quality Standards
• Benchmarking / Profiling
• Cost / Charges
• Guiding Providers
• Adherence

Guidelines Practice Guidelines /
Critical Pathways
Source Shaw LJ, Redberg RF. From clinical trials
to public health policy The path from imaging to
screening. Am J Cardiol 2001 Jul
1988(2-A)62E 65E.
16
Basics of CEA

• CEA technique for selecting among competing
  choices when resources are limited.
• Value for Money
• Technique comparing relative value of various
  clinical strategies. Commonly, a new strategy is
  compared w/ current practice (i.e., "low-cost
  alternative") in calculation of CE ratio
• Result "price" of an additional outcome
  purchased by switching from current practice to
  new strategy (e.g., 10,000 / life year). If the
  price is low enough, new strategy is considered
  "cost-effective.

Standard lt50,000 / LYS
Source http//www.acponline.org/journals/ecp/sepo
ct00/primer.htm
17
Critical Cost Effectiveness (CE) Questions

• 1. Vs. usual carei.e., no screeningwhat is the
  CE of CHD screening of asymptomatic adults to
  reduce risk for CHD-specific morbidity /
  mortality?
• 2. What is the CE of selective screening adults
  at increased risk for CHD e.g., those with a
  family history of premature CHD, w/ risk factors
  vs. routine screening usual care?
• 3. How will differences in rx effectiveness
  affect CE estimates for CHD screening?
• 4. Among individuals w/ subclinical disease on
  initial screening exam, what is the CE of
  periodic surveillance vs. one-time screening?
• 5. Among individuals w/out subclinical CAD on
  initial screening exam, what is the CE of
  re-screening at varying intervals vs. onetime
  screening?

18
Screening Criteria Discussed

• Burden
• Prevalence of disease
• Years of life lost
• Disability or quality of life
• Economic burden
• Effectiveness and Efficacy
• Cost effectiveness
• Current delivery rates
• Feasibility of increasing delivery rates

19
Cost Effective CHD Screening

• 1. Detection of Risk
• 2. Early Rx
• 3. Improved Outcome
• Resulting in Reduction in More Costly, End-Stage
  Care
• Improved Societal Productivity

20
Evaluation Criteria

• Burden of disease
• Single measure incorporating mortality
  morbidity
• Effectiveness of Screening
• Cost effectiveness
• Feasibility of Increasing Delivery Rates

21
CHD Screening Framework

• Two Steps
• Burden and Effectiveness into single measure of
  Clinically Preventable Burden (CPB)
• Cost Effectiveness included to account for
  resource consumption

22
Clinically Preventable Burden

• CPB Burden x Effectiveness
• Burden includes all disease targeted by CHD
• Effectiveness of burden reduced
• Measures burden of CHD preventable
• Burden measured in Quality-Adjusted Life Years
  Saved (QALYS) -- approximated
• Uses effectiveness from RCT
• Range of Therapeutic Risk Reduction

23
Clinically Preventable Burden

• Qualitative assessment of CHD screening should
  consider
• CPB - not burden and effectiveness separately
• focus on fatal or high-prevalence, nonfatal
  conditions
• Costs of service medical care, out-of-pocket
• Potential for cost savings

24
Cost Effectiveness (CE) Analysis

• CE costs of screening costs averted
• Net Effectiveness
• ICER
• CHD Screening vs. No Testing / Usual Care
• CHD Screening vs. Global Risk Score
• CHD Screening vs. Alternative Testing
• CAC vs. C-IMT
• CAC vs. BART
• CAC vs. .

Clinically Preventable Burden reduced
25
Treatment-Eligible US-Population under NCEP II,
NCEP III, CAC Screening
Men
Women
NCEP II
NCEP III
CAC
Millions of people
Age (y)
Increase 142.5 184.3
124.9 85.9
65.0 50.0
65.0 50.0
Source Fedder DO et al., Circulation
2002105152-156, Nasir K, Michos ED, Blumenthal
RS, Raggi P. Detection of High-Risk Young Adults
and Women by Coronary Calcium and National
Cholesterol Education Panel-III Guidelines. JACC
2005 (in press).
26
Treatment Est. 10-Yr Costs from NCEP III to CAC
Screening
Men
Women
NCEP III
CAC
Millions of
Source Fedder DO et al., Circulation
2002105152-156, Nasir K, Michos ED, Blumenthal
RS, Raggi P. Detection of High-Risk Young Adults
and Women by Coronary Calcium and National
Cholesterol Education Panel-III Guidelines. JACC
2005 (in press).
27
Relative Risk (RR) Ratios (95 CI) by CACS Risk
Events / N
CACS
RR
(95 CI)
p Value
0.01
0.1
1
10
100
Summary RR Ratio
Higher Risk
Low Risk
1.5
(0.8-2.9)
24 / 6931
18 / 8503
0.18
Very Low Risk
1-44
Low Risk
2.1
(1.3-3.3)
46 / 2670
26 / 4600
0.003
1-112
Moderate Risk
4.1
(2.9-6.0)
102 / 4,428
44 / 9,977
lt0.0001
100-400
High Risk
6.7
(4.8-9.4)
179 / 3,550
44 / 6,839
lt0.0001
400-999
Very High Risk
1,000
10.8
(4.2-27.7)
14 / 196
6 / 905
lt0.0001
0.01
0.1
1
10
100
Lower Risk
Higher Risk
Very Low Risk includes Kondos, LaMonte, Taylor
When c/w FRS event rates, ? LYS with CACS ? 0.58
for 35 RR Reduction w/ Rx (0-0.83)
Low Risk includes Arad, Greenland, LaMonte
Moderate Risk includes Arad, Greenland, LaMonte,
Taylor, Vliegenthart
High Risk includes Arad, Greenland, Kondos,
LaMonte, Vliegenthart
Very High Risk includes Vliegenthart
28
CPB Model Inputs Disease Burden
Source MI rates were extrapolated from ARIC,
1987-2000 does not include silent MIs. CVA data
also not included.
29
CPB Model Inputs Disease Burden
Source MI rates were extrapolated from ARIC,
1987-2000 does not include silent MIs. CVA data
also not included.
30
CPB Model Inputs Procedure Burden
Source CDC/NCHS for 2002. http//www.acc.org/advo
cacy/word_files/2005ProposedPhysicianPmtRulev320w
eb.xls.
31
CPB Model Inputs Procedure Burden
Source CDC/NCHS for 2002. http//www.acc.org/advo
cacy/word_files/2005ProposedPhysicianPmtRulev320w
eb.xls.
32
Markov Model Health states - ovals arrows
represent allowed transitions. All pts start
event-free can remain, have MI or angina, or
die.
Event-Free
Post-MI
Post-AP
Death
Post-MI AP
Markov model to estimate the benefits, costs,
incremental cost-effectiveness of CHD screening
followed by targeted statin rx for high risk
subclinical dz, vs. usual care alone, for the
primary prevention of CV events among patients
ages 45-65 years..
Source Blake GJ, Ridker PM, Kuntz KM. Potential
Cost-effectiveness of C-Reactive Protein
Screening Followed by Targeted Statin Therapy for
the Primary Prevention of Cardiovascular Disease
among Patients without Overt Hyperlipidemia. Am J
Med 2003114485 494.
33
Multi-Attribute Cost Markov Model Comparing FRS
vs. CACS for 5 Yrs. Of Observational Follow-up
Estimated LYS in Pts. Ages 45-65 yrs.
Event-Free
Event-Free
FRS
CACS
Death
Death
Post-MI AP
Post-MI AP
lt50,000 / Events Averted
34
Conclusions

• If we can identify w/ a high degree of likelihood
  pts at risk for AMI / SCD, then it is likely that
  a CV screening-driven approach including
  prevention (i.e., risk factor modification) can
  result in improved outcomes aversion of costly
  hospitalizations.
• Preliminary analyses from the CE models reveal
  that subclinical dz screening can be cost
  effective when applied to higher risk or
  appropriate patient candidates.
• When compared with global risk scores that often
  underestimate risk in key patient subsets women,
  young, international cohorts.
• Decision models do not replace RCT comparing an
  array of imaging modalities, laboratory markers,
  or global risk scoring.

35
Potential Evidence for Priority Setting
Priority Criteria Measures
Impact Condition Disability, Mortality
System Costs, Guideline Adherence,
Errors Societal Indirect Costs Improvabilit
y Condition Cost-Effectiveness,
efficacy Disparity Impact on vulnerable
subgroups System Effectiveness of quality
improvement Inclusiveness Diffusion across
subpopulations
36
Unmet Expectations Limitations to CHD Screening

• Many preventive services are recommended
• Delivery of effective services is incomplete
• Resourcestime and moneyare limited
• Preventive services differ in their health impact
  and costs