The PROspective Multicenter Imaging Study for Evaluation of Chest Pain (PROMISE) Trial: Economic Outcomes

1
The PROspective Multicenter Imaging Study for
Evaluation of Chest Pain (PROMISE) Trial
Economic Outcomes
Daniel B. Mark, MD, MPH Professor of
Medicine Vice Chief for Academic Affairs,
Cardiology Division Duke University Medical
Center Director, Outcomes Research Duke Clinical
Research Institute
Co-Investigators/Econ Team Kevin Anstrom Patricia
Cowper Linda Davidson Ray Udo Hoffmann Manesh
Patel Lawton Cooper Kerry Lee Pamela Douglas Jeff
Federspiel Melanie Daniels
Financial Disclosures
Consulting Milestone Medtronic CardioDx St Jude
Medical
Research Grants NIH Eli Lilly
Company AstraZeneca Gilead AGA Medical Bristol
Myers Squibb
March 15, 2015
2
PROMISE Trial BackgroundMoving From Controversy
to Evidence

• Noninvasive ability to directly visualize the
  coronary arteries of patients with chest pain has
  long been on Cardiologys Wish List
• As coronary CT angiography evolved into a test
  that might actually be able to fulfill this wish,
  controversy broke out
• The PRO side CTA would allow precision care -
  only the patients who needed revascularization
  would actually go to cath and the rest would
  avoid it ? invasive testing, ? unneeded
  revascularization, ? false positives, ?
• The CON side CTA would ? non-invasive and
  invasive testing to clarify ambiguous findings, ?
  radiation exposure, ??

3
PROMISE Design Overview
10,003 patients with symptoms of CAD

• New or worsening chest pain or symptoms w/out
  known CAD
• Low to intermediate risk
• Planned noninvasive testing for diagnosis

193 sites (US, CA)
11 Randomization Stratified by site and intended
functional test
Usual Care Arm Pre-selected Functional Testing
Intervention Arm Anatomic Testing 64-slice CTA
Median study follow-up 25.2 months 1 endpoint
composite of death, MI, UA hosp, or major
procedural complication 2 aims incl. cost and
cost effectiveness
4
PROMISE TrialCTA Patient Outcomes Not Superior
to Functional Testing
Strategy of initial CTA, as compared with
functional testing, did not improve clinical
outcomes over a median follow-up of 2 years.
Douglas PS et al NEJM 2015
5
PROMISE Primary Endpoint ResultsDeath, MI,
Unstable Angina, Major Procedural Complications
Douglas PS et al NEJM 2015
6
PROMISE Economic SubstudyPrimary Objectives

• Measure and compare cumulative total costs as
  randomized
• If CTA outcomes superior, estimate cost
  effectiveness of anatomic strategy

7
PROMISE Economic SubstudyCalculation of Medical
Costs

• 96 (9649) of PROMISE cohort in Economic Substudy
• Initial diagnostic test technical fees
• Bottom up estimate (resource-based cost
  accounting methods) from large proprietary
  registry (Premier Research Database)
• Hospital-based facility costs
• UB 04 bill forms provide hospital charges by
  department
• Department-specific ratios of costs to charges
  (RCCs) used to convert charges to estimates of
  cost
• MD professional fees for testing and hospital
  services
• Medicare Fee Schedule

8
PROMISE Economic SubstudyAnalysis Methods

• Comparisons by intention to treat principle
• Costs to 3 years estimated, accounting for
  censoring using inverse probability weighting
  methods
• Bootstrapped confidence intervals 1000
  replications (500 in subgroup analyses), 95
  confidence intervals

9
PROMISE Economic SubstudyBaseline
Characteristics
Functional Anatomic (N4,818)
(N4,831) Anatomic (N4,818)
Demographics    
Age, mean 60.9 8.3 60.7 8.3
Female 54 52
Cardiac risk factors    
BMI, mean 30.6 6.2 30.6 6.2
Hypertension 66 66
Diabetes 22 22
Dyslipidemia 68 67
Family history premature CAD 32 33
Current or past smoking 51 51
Primary symptom chest pain or DOE  88  88
Typical or atypical angina 89 89
Pretest probability of CAD 53 54
10
PROMISE Economic SubstudyEstimation of Initial
Diagnostic Testing Costs
Total 404 514 501 174 946 1132
Dx Test CTA Echo w/ exercise stress Echo w/
pharmacologic stress ECG-only Stress Nuclear w/
exercise stress Nuclear w/ pharmacologic stress
Mean Cost 285 428 415 137 829 1015
MD Fees 119 86 86 37 117 117
based on costs in Premier database based on
Medicare Fee Schedule
11
PROMISE Economic SubstudyCumulative Total Costs
by ITT and Mean Cost Difference (95CI)
Difference in Cost (Anatomic Functional)
Cumulative Cost
694
388
358
279
12
PROMISE Secondary Endpoints90-Day
Catheterization and Revascularization Rates
Invasive cath Revascularization No CAD
on cath
CTA (n4996) 609 (12.2) 311 (6.2) (51 of
cath patients) 170 (3.4) (28 of cath
patients)
Functional (n5007) 406 (8.1) 158
(3.2) (39 of cath patients) 213 (4.3) (52
of cath patients)
13
PROMISE Economic SubstudyCost Differences by
Categories 0-3 and 4-12 Months
-378
68
203
17
12
-17
49
43
-10
8
8
81
357
279
14
PROMISE Economic SubstudyCost Differences by
Categories Years 2 and 3
10
20
-35
-97
311
97
306
35
-12
-69
7
15
53
29
15
PROMISE Economic Substudy2-Year Cost Difference
Thresholds From Bootstrap Analysis
Cumulative distribution of mean cost difference
CTA-FXN from 1000 bootstrap replications out to
24 months
Cost difference lt 500 62 of samples lt 750
81 of samples lt 1000 93 of samples
16
PROMISE Economic Substudy Pre-Randomization MD
Choice of Functional Test Subgroups
Months 0-36
Months 0-3
17
PROMISE Economic SubstudyCaveats

• Costs of initial testing from external data
  source
• Significant deviations by centers from testing
  costs used in this analysis might alter relative
  cost positions of the two strategies
• Outpatient medications not counted
• QOL and employment status still being analyzed

18
PROMISE Economic SubstudySummary

• In stable patients with new chest pain, CTA
  strategy improved efficiency of use of invasive
  cath (fewer normal caths, higher proportion of
  caths also getting revasc)
• But despite lower testing costs for CTA compared
  with stress echo (100 less) and stress nuclear
  (630 less), net effect was to drive a small
  (lt500), statistically non-significant increase
  in cost
• After 90 days, very little test strategy-related
  differences in costs out to 3 years
• Coronary CTA may not be the holy grail of
  diagnostic testing once hoped for, but its more
  liberal use following PROMISE standards will
  improve some aspects of care without causing a
  major new economic burden on the health care
  system

19
(No Transcript)