New Therapies in Cardiology: Are they CostEffective

1
New Therapies in CardiologyAre they
Cost-Effective?

• David J. Cohen, M.D., M.Sc.
• Director, Interventional Cardiology Research
• Beth Israel Deaconess Medical Center, Boston
• Department of Health Policy and Management,
  Harvard School of Public Health, Boston

2
Potential Economic Impact of New Cardiovascular
Devices in 2003
3
Basic Concepts in Economic Evaluation of Medical
Technologies
4
Cost-Effective? Cost-Saving
Principles of Cost-Effectiveness
5
Cost-Saving ? Cost-Effective
More Effective
Less Effective
_
???
More

???
Less
6
Cost-Benefit Analysis

• Goal
• To determine the net benefit of any health care
  program/treatment
• General Approach
• Health expenditures and benefits both measured in
  the same unit (dollars)
• Net Economic Benefit DBenefit - DCost
• Decision Rule Any treatment program is
  worthwhile if its net benefit is positive

7
Cost-Benefit Analysis (cont)

• Advantages
• Trade-offs between costs and benefits are
  explicit
• Decision rule easy to apply
• Disadvantages
• Difficult to value health benefits in economic
  terms
• Monetary value of life saved?
• Monetary value of pain and suffering?
• Potential approaches
• Income stream approach
• Willingness to pay approach
• Tends to systematically undervalue lives of the
  poor, elderly, and very young (or unborn)

8
Cost-Effectiveness Analysis

• Goal
• Allocate health investments/expenditures so as to
  maximize the aggregate health benefits to
  society, subject to the constraint of a fixed
  health care budget
• General Approach
• Costs and net health benefits measured in natural
  units
• Costs-- dollars
• Health Benefits-- lives saved, life-years gained,
  complications prevented
• C/E ratio DCost/DEffectiveness (vs. best
  alternative)

9
CE ratio the numerator

• Direct costs costs associated with the labor,
  equipment, and supplies necessary to provide
  clinical strategy
• Indirect costs (overhead) rent, depreciation,
  maintenance, etc.
• Induced costs downstream costs incurred or
  avoided due to an up-front clinical strategy
• Productivity costs patient/family time spent on
  receiving/providing care

10
CE ratio the denominator
11
Measurement of Net Health Benefits

• Quality-Adjusted Life Year (QALY)-- metric for
  combining quality and duration of life into a
  single measure of health effectiveness

CABG
Good Health
Quality-adjusted life expectancy (1.0 x 3
yrs) (0.7 x 3 yrs) (0.9 x 5 yrs) (0.3 x
1 yrs) 9.9 QALYs
Mild Angina
Stroke
Severe Angina
12
Cost-Effectiveness Analysis
Cost-Effectiveness Ratio DC / DE

• Decision Rule (Theoretical)
• Calculate C/E ratio for each program to be
  evaluated
• Rank programs in order of increasing C/E ratio
• Select programs for funding in order of C/E
  ratios (lowest to highest), until health care
  budget is exhausted

13
Cost per Quality-Adjusted Year of Life Saved for
Some Health Investments
Investment Lovastatin for 2o prevention CABG
(3-vessel disease) Treatment of severe
hypertension AICD (sudden death
survivor) Treatment of mild hypertension Neonatal
ICU (500-999g) Hospital Hemodialysis Lovastatin
for 1o prevention (CHLlt250, no other risk
factors)
Cost per QALYgained (1992 ) Cost-saving
10,100 15,600 23,000 35,700
55,000 84,000 120,000
14
Cost-Effectiveness Depends on Perspective
Principles of Cost-Effectiveness
15
Cost-Effectiveness Depends on Perspective
Principles of Economic Evaluation

• A 40 y.o. man is admitted to St. Elsewhere
  Hospital with unstable angina. Cardiac cath
  demonstrates a high-grade LAD lesion treated
  successfully with PTCA. He spends 1 day in the
  CCU, 2 days in a cardiac step-down unit and is
  discharged home on hospital day 4.
• Several weeks later, he receives a hospital bill
  for 19,000, which his HMO covers in full.
• Question How much did the hospitalization cost?

16
All Cost-Effectiveness is Relative
Principles of Cost-Effectiveness
17
Cost-Effectiveness is Incremental
Z
Cost
Y
X
Do Nothing
Effect
18
Cost-Effectiveness is Incremental
Avg. cost-effectiveness
Z
Costz Effectz
Cost
Y
X
Do Nothing
Effect
19
Cost-Effectiveness is Incremental
Avg. cost-effectiveness
Z
Costz Effectz
D Cost
IncrementalCostZ
IncrementalEffectivenessZ
Y
X
D Effect
20
Incremental Cost-EffectivenessDoes it Really
Make a Difference?
Per 10,000 pts screened
Neuhauser D et al. NEJM 1975293226-8
21
Incremental Cost-EffectivenessDoes it Really
Make a Difference?
Per 10,000 pts screened
Neuhauser D et al. NEJM 1975293226-8
22
Treatments are not cost-effective unless they
are effective
Principles of Cost-Effectiveness
23
Case Study 1 ICDs
24
If the results of this trial were incorporated
now into our clinical practice, the consequences
would be staggering . We estimate that the cost
would be over 1 billion annually .... Should
this economic burden be added to an already
overladen health care delivery system on the
basis of a single clinical trial involving fewer
than 200 patients? - Friedman PL and Stevenson
WG (NEJM 1996)
25
CMS presented a specious analysis to discredit
the rigor of the studys landmark findings I am
not so naïve a physician as to believe cost
shouldnt be a factor in medical decision
making Medicare pays for dozens of therapies
that are not as efficacious as an ICD in directly
saving lives. - Ronald Berger, MD (Wash Post,
May 15 2003)
26
Are ICDs Cost-Effective?
27
Secondary Prevention AVID

• AVID
• NIH-funded trial
• 1016 patients with resuscitated VF or symptomatic
  sustained VT
• Mean age 65 yrs
• Mean LVEF 32
• 3-year survival
• ICD 75
• Amio 64

P0.02
N Engl J Med 1997 3371576-1584
28
Cost-Effectiveness in AVID
Cumulative Cost (US )

• AVID C-E Analysis
• Time horizon 6 yrs
• D Cost 26,166
• D LE 0.33 yrs
• Reductions in LOS with ICD placement projected
  to reduce C/E ratio to lt45K/LY

C/E ratio 79,291/LYS
ICD group
AAD group
Conclusion ICDs for secondary prevention are
moderately cost-effective by US standards
Larsen et al. Circ 20021052049
29
Are ICDs Cost-Effective?
30
Primary Prevention MADIT
Survival (all-cause)

• 196 patients with
• Prior MI
• EFlt35
• NSVT (3-30 beats)
• AND inducible VT at EPS
• Randomized to ICD vs. usual care (74 amio)
• Main Result All cause mortality reduced by 54
  in ICD group (NNT 5 at 3 years)

Moss AM et al. N Engl J Med 19963351933-40
31
Cost-Effectiveness in MADIT

• MADIT C-E Analysis
• Time horizon 4 yrs
• D Cost 21,500
• D LE 0.80 yrs
• Use of transvenous defibrillator projected to
  reduce C/E ratio to 23,000/LY gained

Annual Cost (US )
C/E ratio 27,000/LY
Conclusion ICDs for primary prevention in
patients with EFlt35 and inducible VT are
reasonably cost-effective
Mushlin AI et al. Circ 1998972129-35
32
Are ICDs Cost-Effective?
33
MADIT II In-trial cost-effectiveness
Impact of Time Horizon
P0.007
Zwanziger J et al. JACC 2006472310-8
34
MADIT-2 Modeled Life Expectancy
Survival

• Life expectancy after ICD projected for up to 12
  yrs
• Used 3 alternative models of long-term survival
• ICD1 continued benefit (HR 0.68)
• ICD2- gradual loss of efficacy (HR 1 at 12 yrs)
• ICD 3- accelerated loss of efficacy (HR 1.4 at
  12 yrs)

ICD1
Survival
ICD3
ICD2
Time (yrs)
Zwanziger J et al. JACC 2006472310-8
35
ICDs for Primary PreventionIs There a Role for
Risk-Stratification?

• Stanford PORT Model
• Computer simulation (Markov) model
• Suggests that ICDs for primary prevention are
  only reasonably cost-effective for patients with
  both high cardiac mortality and high proportion
  of sudden death (vs. pump dysfunction)
• ? Can we prospectively identify these patients

Owens DK et al. Am Heart J 2002144440
36
Summary ICD Cost-Effectiveness

• ICDs are clearly cost-effective for primary
  prevention for patients at high risk of SCD on
  the basis of reduced EF and markers (NSVT, EPS)
• Attractiveness of ICDs for primary prevention in
  lower risk populations depends critically on
• Absolute risk of sudden cardiac death and
  relative risk of SCD vs. all-cause mortality
• Implant cost and need for generator replacement
  in subsequent years
• Ability to reliably risk-stratify patients would
  substantially enhance the overall
  cost-effectiveness of this technology

37
Case Study 2 Drug-Eluting Stents
38
Drug-Eluting Stents

• Drug-eluting stents allow high levels of
  inhibitory drugs to be administered in a
  controlled fashion directly at the time and
  location of injury
• DES have dramatically reduced angiographic and
  clinical restenosis rates
• Sirolimus (Cordis) 75-90 risk reduction
• Paclitaxel (BSC) 60-80 risk reduction
• Numerous other drugs and stents in testing

Bx Velocity
Sirolimus-1XTC
39
Cost-Effectiveness of Drug-Eluting Stents

• Do they save money?
• Do they save lives?
• Do they improve quality of life?
• Are they cost-effective (for which patients)?

40
Repeat Revascularization
Medicare PCI 1998-9
20
N11,920
Any
16.2
15
Estimated clinical restenosis rate 0.85
16.2 13.8
12.7
PCI
10
5
4.2
CABG
41
Mean 1-year F/U Cost per Patient
Medicare PCI 1998-99
23,808?19,506
Attributable cost of restenosis 18,944 (95
CI, 18,410-19,477 ) After risk
adjustment

• But.
• Drug-eluting stents cost 1400/stent more than
  bare stents
• We use 1.6-1.7 per procedure
• And they only reduce restenosis by 70-80
• And they require at least 1200 of additional
  dual antiplatelet therapy

Economic Burden of Restenosis 18,944
13.8 2550/PCI patient
4,087?9,528
Restenosis
No Restenosis
42
Costs and Health Effects of Recent Advances in
Interventional Cardiology
Costs
2b/3a inhibitors
Stents
DMR/TMR
Embolic Protection
IVUS
Health Benefits
Direct PTCA for AMI
43
Cost-Effectiveness of Drug-Eluting Stents

• Do they save money?
• Do they save lives?
• Do they improve quality of life?
• Are they cost-effective (for which patients)?

44
Does Restenosis Increase Late Mortality?

• Emory Study
• 3363 pts who underwent angiographic restudy after
  successful PTCA
• No difference in 6 yr mortality according to
  status at angio. f/u
• No restenosis 5
• Restenosis 7

Implications Except for highly selected subsets
(e.g., unprotected LM, single remaining vessel)
DES unlikely to improve survival
Weintraub W et al. Circulation 199387831-40
45
Impact of Restenosis on Quality of Life
Stent PAMI
No TVR
TVR
D10.1 Plt0.001
D10.0 Plt0.001
D7.8 P0.006
D6.4 P0.03
D14.5 P0.008
Rinfret SA et al. J Am Coll Cardiol
2001381614021
46
Cost-Effectiveness of Drug-Eluting Stents

• Do they save money?
• Do they save lives?
• Do they improve quality of life?
• Are they cost-effective (for which patients)?

47
Cost-Effectiveness Measures
Economics of Restenosis

• Cost per quality-adjusted year of life gained
• Standard metric for CEA
• Allows comparison across different diseases
• C/E threshold of 50,000 reasonably
  well-accepted (dialysis benchmark)

• Cost per repeat revascularization avoided
• Readily measured in both clinical trials and
  observational studies
• Interpretable to both patients and clinicians
• Appropriate threshold within U.S. healthcare
  system less well-defined (? 10,000 per repeat
  revasc avoided)

48
Disease-Specific C/E ratios Cost per Repeat
Revascularization Avoided
49
Cost-Effectiveness of DES (2008 data)Impact of
Bare Metal Stent Restenosis Rate
Healthcare system perspective

• Model Assumptions
• Incremental cost per DES 1400
• DES per case 1.6 (case-mix adjusted)
• 70 reduction in TVR with DES vs. BMS (New SIRIUS)

Conclusions DES reasonably cost-effective when
bare stent TVR rate gt 10
50
Predicted Clinical Restenosis Rate
Diabetes
No Diabetes
Adapted from Ho KKL et al. ACC 1999
51

SIRIUS
1-Year Medical Care Costs
D 309 (p0.64)
16,813 9737
16,504 11,511
C/E ratio 1650 per repeat revasc
avoided 27,000 per QALY gained
Cohen DJ et al. Circulation 2004
52
Relative Impact of Restenosis and Mortality on
Quality-Adjusted Life Expectancy
Reduction in Quality-Adjusted Life Expectancy
(per event)
125x
Cohen DJ et al. Circulation 2001
53
Take Home Messages

• Cost-effectiveness analysis is not about saving
  money
• Treatments are not cost-effective unless they are
  effective
• Cost-effectiveness is not an intrinsic property
  of a drug or device? can vary considerably with
  clinical setting (i.e., primary vs secondary
  prevention ACS vs. stable CAD) as well as
  patient characteristics

54
Take Home Messages- 2

• C/E analysis can be used to supplement
  traditional mechanistic and outcomes studies in
  order to develop evidence-based guidelines for
  device utilization and treatment strategies
• Technologies that increase the overall cost of
  medical care may still be acceptable in the
  future, but only if we can demonstrate improved
  clinical outcomes that are meaningful to our
  patients (i.e., increased life expectancy,
  reduced symptoms, or improved QOL)

55
(No Transcript)