The CostEffectiveness of Primary Angioplasty compared to Thrombolytics for Acute Myocardial Infarcti

1
The Cost-Effectiveness of Primary Angioplasty
compared to Thrombolytics for Acute Myocardial
Infarction

• Yolanda Bravo, Christian Asseburg, Steve Palmer,
  Liz Fenwick and Mark Sculpher
• Centre for Health Economics, University of York

Clinical advisors Mark de Belder, David Gray,
Rob Henderson, Jim MacClenan Statistical
advisor Keith Abrams.
Research funded under an unrestricted educational
grant from Cordis UK
2
Overall Aim

• Previous meta-analyses that compare thrombolysis
  and primary angioplasty (PCI) have shown
  significant clinical benefits from angioplasty in
  terms of reducing major adverse clinical events
• Thrombolytic treatment remains the default
  treatment option in many countries (including the
  UK). Possible reasons include
• practical reasons (shortage of cardiac catheter
  facilities and skilled staff)
• additional delay in initiating treatment
  (PCI-time delay)
• limited evidence related to long-term
  cost-effectiveness
• Establish whether the additional cost of PCI is
  justified in terms of long-term generic outcomes
  (QALYs) relative to medical management with
  thrombolytics, using a decision analytic model.
• Address two of the major sources of uncertainty
• the impact of PCI-time delay
• maintenance of benefits in the long term.

3
When is PCI Cost-Effective?
Benefits of primary PCI
More
Less
?
More
PCI dominated
Costs of primary PCI
Less
?
PCI dominates
What is the opportunity cost of finding the
additional resources to provide PCI? Can the
extra resources be obtained from doing less of
something else and TOTAL benefits increase?
4
What do we need to know about PCI to make
decisions about resources?

• Outcomes in terms of mortality and morbidity
• Measured in generic units comparable with other
  services
• Time horizon of a patients lifetime

• Costs on initial procedures
• Costs of further cardiac procedures
• Costs over a patients lifetime

5
Previous Cost-effectiveness evidence
Lieu TA, Gurley RJ, Lundstrom RJ et al.
Projected cost-effectiveness of primary
angioplasty for acute myocardial infarction.
JACC 199730(7)1741-50

• QALYs
• Lifetime time horizon
• Considered key scenarios
• BUT... US costs based on evidence to mid-1990s
• PCI Dominant

Hartwell D, Colquitt J, Loveman E et al. 2003.
Clinical and cost-effectiveness of immediate
angioplasty for acute myocardial infarction.
HTA, vol.9(17)

• Only short-term costs
• No link with QALYs
• ICER 8,707 to 12,171

6
Methods

• Update most recent meta-analysis (Keeley et al.,
  Lancet 2003 361) scope and statistical rigour.
• Evaluate the relationship between the treatment
  effects and the time delay involving initiation
  of primary angioplasty, for a number of outcomes
  (Death, NF MI, NF stroke).
• Bayesian random-effects meta-regression model
• Measurement uncertainty in the time delay
  covariate is modelled explicitly
• Lifetime extrapolation (Markov model, 40 years)
  to quantify the lifetime costs and QALYs
  associated with primary angioplasty and
  thrombolysis

7
Overview Model Structure
8
Evidence Synthesis - RCT data
D RI St
D RI St

• 22 trials
• All trials report mortality at 4-6 weeks
• There are fewer data on longer term events (6
  months) and non-fatal outcomes
• The additional time delay for angioplasty can be
  estimated from summary statistics on the
  time-to-needle and time-to-balloon.
• PCI-related time delay difference in time to
  initiation of treatment between the two
  reperfusion strategies

4-6 weeks
4-6 weeks
9
Mortality
10
Reinfarction (Non-Fatal)
11
Stroke (Non-Fatal)
12
Base-Case (Absolute probabilities)
Average time delay 54.3 minutes
13
Alternative Time-Delays
6-month treatment effect of PCI compared to
Lysis (mean and 95 Cr.I.)
14
Extrapolation

• Lifetime extrapolation of costs and outcomes
  (QALYs) based on long-term observational data
• Extrapolation data derived from 1992 cohort of
  the Nottingham Heart Attack Registry (NHAR)
• 627 patients with 5 years follow-up of survival
  and subsequent events (MI/ Stroke)
• Transition probabilities derived using survival
  analysis
• Resource use and costs from NHAR and UK reference
  costs
• Utility data from external sources (systematic
  review)

15
Cost-Effectiveness Plane
16
Cost-Effectiveness Results (ICER)
BASE-CASE
SENSITIVITY ANALYSES
17
Cost-Effectiveness Acceptability Curves
___Base_case (54.3 min) ---- 30 min ..
60min -.-.-. 90 min.
18
Discussion

• Results indicate that PCI is cost-effective for
  the treatment of AMI based on a lifetime horizon.
  
• ICER 9,241 for base-case analysis (54.3 min
  average time delay), in spite of conservative
  assumptions (stents, GPAs, hospital LOS).
• These findings are explained by
• superior mortality benefit associated with PCI
  for delays up to 60 min
• prevention of NF MI and NF stroke for delays up
  to 80 min.
• Primary PCI appears superior, on average, for
  delays of between 90-120 minutes across all the
  outcomes considered.
• At delays gt 60 minutes, the cost-effectiveness
  threshold is important. From a UK perspective,
  primary angioplasty does not appear
  cost-effective at a delay of 90-minutes.

19
Appendix 1. EVSY - Model
20
Appendix 2. Previous Meta-Regression
Death 1-month Source Nallamothu and Bates, Am J
Cardiol 2003 92 824-826.
21
Appendix 3. Short-term Model (6 m.)
22
Appendix 4. Long-term Model