Professional Documents
Culture Documents
Economic Analysis in Clinical Research
Economic Analysis in Clinical Research
Economic Analysis in Clinical Research
Martin L. Brown, Ph.D. Applied Research Program Division of Cancer Control and Population Sciences National Cancer Institute mb53o@nih.gov
Introduction to the Principles and Practice of Clinical Research October 31, 2005
Overview
Introduction Concepts Technical
Examples Cost of Illness Cost-Benefit Analysis Marginal Analysis Cost Effectiveness Analysis
Clinical Trials plus Modeling CEA Along Side a Clinical Trial
INTRODUCTION
care resource allocation decisions understanding determinants of technical and organization efficiency of health care delivery understanding determinants of distribution, access and equity issues in health care delivery
For
For
Services Research: observational data in community settings or quasicontrolled experimental settings Economics Modeling: economic modeling combined with results of randomized clinical trials through modeling Studies Along-side Clinical Trials: used especially in Pharmacoeconomics
Health
Economic
CONCEPTS
Cost-of-illness
Cost
Cost Benefit Analysis (CBA) Cost Effectiveness Analysis (CEA) Cost Utility Analysis (CUA)
relevant costs and benefits should be counted of evaluation is to compare alternative used of resources
is incremental
Purpose
Measurement
costs and health effects are expressed in monetary terms (i.e., must put a $ value on a year of life) Benefit -> all benefits minus all costs, sometimes call Social Return on Investment
Cost
Cost
Benefit Ratio -> All benefits divided by all costs, sometimes called Social Rate of Return
terms
Benefits
are expressed in natural units, e.g., life-years Effectiveness Ratio -> Cost divided by life-years (or other measure of benefit)
Cost
terms
Benefits
Cost
Let
Intervention
B A
0 50
C
may be costeffective
100 150
lose/lose
Cost
-1 5 0 -1 0 0 -5 0
50
clinically ineffective
-5 0
win/win
-1 0 0 -1 5 0
Health Effects
Dominance
Intervention
A produces more health benefits at lower cost than intervention B: A dominates B. Intervention C produces more health benefits than intervention A, but at higher cost: the ICER of A relative to C can be computed. ICER of A relative to C is the slope of the dotted line
Transformations of ICER
ICER may be transformed to: Net Benefit (NB) = E (C * 1/Vs) Net Monetary Benefit (NMB) = (Vs * E) C
Vs
= the Social Value of health (e.g., $100,000 per life-year, but this can also be varied) transformations are useful when dealing with uncertainty in the estimates of ICER
These
Comparison of Interventions
Micro-analysis
comparison is no treatment or status quo treatment Marginal analysis comparisons are different intensities of the same intervention Macro-analysis
Comprehensive league table, e.g.
Disease Control Priorities Project Vs as determined by rule of thumb, e.g. $100,000 rule Vs as determined by economic analysis as function of wealth, distribution and preferences
of Cost Data
Technical
Statistical
trial forms/medical record abstraction Hospital bills Health system cost-accounting systems (e.g. HMOs) Administrative claims data (e.g. Medicare, Medstat) Patient/provider survey (e.g. MEPS) Cost scenario Time-motion study Engineering study
Discounting Pricing
non-market goods
data are complex Economic data tend to be highly skewed and censored - special estimation techniques have been developed Trials designed for clinical end-points may be under-powered for economic and/or cost-effectiveness results Cost-effectiveness or Cost-utility ratio estimates pose specific problems for analyzing and presenting confidence intervals (regions)
EXAMPLES
Cost Domains
Cost
domains refers to categories of costs according to whether they are directly or indirectly related to the provision of marketed health care services. cost domain may also determine whether accessible and/or high quality cost data is available to the researcher and what degree of effort is required to obtain data.
The
health care costs (e.g. Medicare payments) non-health care costs (e.g., paid child care)
Direct
Patient
Morbidity
Mortality
Source: H. Varmus, Disease Specific Estimates of Direct and Indirect Costs of Illness and NIH Support
Billions of Dollars
For colorectal cancer, time costs (valued by average wage rates) during initial treatment were $4655, 20% of direct medical expenditures in that period.
Cost-Benefit
Source: PS Romano, et al. Folic acid fortification of grain: an economic analysis. AJPH 1995:85:667-676.
Note:
Cost
Proportion
Cases
benefit of birth defects averted: $121.5 million benefit of fortification program = $93.6 million Ratio = 4.3
Net
Benefit/Cost
What
Is
Cost-effectiveness analysis of folic acid supplementation vs. fortification Both found to be cost-savings compared to doing nothing Fortification (dominantly) costeffective relative to supplementation
Source: Kelly AE, et al. Appendix B in Gold, et al. Cost-Effectiveness in Health and Medicine. Oxford U. Press, 1996.
of NTD prevalence pre- and post-fortification More fortification that anticipated Original estimate did not take doseresponse into account NTDs averted:
520 Spina Bifida 92 Anancephaly
Net
large should National Cord Blood Bank be (beyond 50,000 inventory initially proposed)? Human leukocyte antigen match rate (and therefore transplant benefit) increases with size Cost increases with size
Source: D. Howard, et al. Institute of Medicine, 2005.
Cord transplants by match level for patients age <20 as a function of inventory
250 5/6 match 200 150 100 50 0 50,000
Number of transplants
6/6 match
4/6 match
150,000
350,000
79,500
Life years
79,000
78,500
78,000 50,000
150,000
250,000
350,000
Inventory size
Costs in millions
150,000
250,000
350,000
Inventory size
Cost per life year gained of increasing inventory from 150,000 to 200,000 units is $105,000 = $12,000,000 114
ICER
$150,000
$100,000
ICER ICER
$50,000
ICER
Linking
Analyzing
differences in economic resource utilization meaningful from a societal perspective? Will adding economic component to the analysis influence clinical practice or health policy? Is collection of good economic data feasible and affordable within context of overall trial design? Does trial design have external validity from an economic perspective?
JM et al. Decision framework for chemotherapeutic interventions for metastatic nonsmall-cell lung cancer. Journal of the National Cancer Institute, 2000 Aug 16;92(16):1321-9 trial data with a modeling approach to costs and longer term outcomes
Combining
of chemotherapy vs. best supportive care for advanced stage lung cancer benefits of treatment based on survival curves modeled (out to 48 months using Weibull survival function) from RCT results and community survival data based on Canadian cost scenarios (POHEM model)
Survival
Costs
cost of best supportive care = $25,904 cost of chemotherapy = $25,105 $41,576 depending on regimen
Total
Hospital/Clinic
costs are higher for best supportive care compared to chemotherapy e.g. intervention results in down-stream cost savings
of chemotherapy ranges from costsavings to $37,800 / quality adjusted life year
ICER
Cost-effectiveness analysis of adding early hormonal therapy to radiotherapy for locally advanced prostate cancer Clinical Trial EORTC 22863
CEA of Hormonal Treatment Construction of 95% Confidence Region Using standard Monte Carlo simulation methods, (sampling with replacement) 5000 replications of the Incremental Cost Effectiveness Ratio, were calculated
bootstrap simulation Examine all results that fall within 95% confidence intervals for the cost effectiveness ratio Compare to reference values for social value of health (Vs) Calculate probability that: CER < Vs
It represents the value that a society or an individual places on one extra unit of health It reflects the a societys level of economic wealth and the relative distribution of that wealth to the health sector Example: consider approaches to cervical cancer control in poor and rich societies.
Costs and Benefits of Cervical Cancer Screening United States Every 3 yr Pap @ $60,000/YLS United States Every 2 yr Pap @ $174,000/YLS
100% 90%
Reduction in Lifetime Cancer Incidence
80% 70% 60% 50% 40% 30% 20% 10% 0% $0 $200 $400 $600 $800 $1,000 $1,200 $1,400 $1,600 $1,800 $2,000 $2,200 $2,400 $2,600 $2,800 $3,000 Lifetime Costs ($U.S.) South Africa Screening 2x Lifetime $50 - $250/ YLS South Africa Screening 1x Lifetime cost saving to <$50/YLS South Africa Screening 3x Lifetime $250 - $500/YLS
Question: How much health is gained in the U.S. per $50,000 invested in this strategy? Answer: 23 days of life-expectancy
Question: How much health is gained in South Africa per $50,000 invested in this strategy? Answer: 1000 years of life-expectancy