COST-EFFECTIVENESS ANALYSIS (CEA)
Mihail Samnaliev, PhD Health Economist Children's Hospital Boston
[email protected]
_______________ The speaker for this session has reported no financial relationships with a commercial entity producing healthcare-related products and/or services
1
Outline
Why use health economics? Definition and applications of CEA Using CEA to improve policy making and health outcomes Definition, measurement & analysis of costs and effectiveness Proposing and designing a CEA
2
C2
Why use health economic evaluations / CEA?
Health care resources are limited and need to be used efficiently
New health technologies emerge every year (drugs, interventions, programs, tests, software, new models of care)
Payers (providers) are faced with the question which to reimburse from available budgets (prescribe for use)
CEA: Interventions with greatest incremental value to patients receive priority
3
Slide 3 C2
CH147121, 10/19/2011
Goal of health economic evaluations / CEA Cost - Effectiveness
Resource use (costs)
Value (effectiveness)
EFFICIENCY
4
Incremental cost-effectiveness ratio (ICER)
ICER = ( COST new strategy – COST current practice ) ( EFFECT new strategy– EFFECT current practice )
ICER: the cost of additional unit of effectiveness
Incremental cost-effectiveness ratio (ICER)
Incremental cost-effectiveness ratio compares health technologies to current practice:
• • •
alternative intervention to treat the same condition do nothing current frequency of an intervention (e.g. HIV screening)
6
ICER of a new health technology vs. current practice
$900,000
ICER= $100k/year
Incremental costs
I
II
I II Less effective & more costly More effective & more costly
$400,000
(dominated) -$100,000 IV III
-$600,000
More effective &less costly
Less effective &less costly
(dominant)
-$1,100,000 -12
-9
-6
-3
0
3
6
9
Incremental life expectancy 7
12
When is a health technology cost-effective?
Depends on societal WTP (threshold) for an additional unit of effectiveness
• • •
Cost-effective if it costs < WTP /additional unit of effectiveness Cost-effectiveness thresholds subjective but not arbitrary Human capital approach, Hypothetical valuations, Revealed preferences
WHO thresholds based on GDP/per capita
•
Very cost-effective if ICER < GDP /QALY; Cost-effective if 1 -3 times GDP /QALY, Cost ineffective > 3 times GDP /QALY
~ $90,000 to ~ $120,000 /QALY often used in the US
Thresholds will be different for other effectiveness measures 8
Examples when/how CEA may be useful Seattle Genetics Inc's blood cancer drug Adcetris recently approved by the FDA
Effective compared to no treatment
A course of treatment can cost > $100,000
CEA would determine the relative value of Adcetris. Can public health gains be higher if resources are invested elsewhere?
Example for when CEA may be useful
Example how existing CE studies can be used by policy makers to improve public health outcomes Intervention
Cost per (qualityadjusted) life year
Extended buprenorphine-naloxone Tx for opioiddependent youth (Polsky et al. 2010 Addiction)
$25,049
Screening every 5 years, vs. a one-time screening program for HIV (Sanders et al. 2005 NEJM)
$57,138
Airline security: having air marshals on planes
> $4 million
(Stewart and Mueller 2008)
($180 million / life saved)
More cost-effective interventions buy more QALY-s / each additional dollar
11
Outline
Why use health economics? CEA: definition and application Using CEA to improve policy making
Measurement & analysis of costs and effectiveness Proposing and designing a CEA
12
Measurement of costs
Costs quantify the resources used for an intervention Expressed in $ to allow within/cross sector comparisons Cost evaluations include: 1. Defining resources/costs to include in the analysis 2. Describing how cost data will be collected/measured 3. Describing the analyses of cost data
13
1. What costs should be included in analyses
Micro costing is recommended; each resource is • identified and listed • quantified in physical units and • valued using per unit costs
Aggregate costs into broader cost categories
14
1. What costs should be included in analyses
Common ways to categorize costs (resources) include:
• • • • • • • • •
Often: direct HC, direct non-HC and indirect, but that depends on what is useful for the specific evaluation Intervention vs. costs that change because of intervention Start up vs. ongoing Variable vs. fixed Direct vs. indirect Marginal vs. average By setting : inpatient, ER, outpatient hospital, outpatient care Productivity costs Intangible costs (physical/emotional suffering and distress) 15
2. Measurement of costs
Define the perspective: societal, payer, provider Analytic timeframe (for both costs and effects) Sources of cost data:
• • • • • •
Own cost data (e.g. CHB cost accounting system) Charges adjusted for cost -charge ratio ( pref. at dept. level) Payments (private plans, Medicaid, Medicare) Published studies (individual studies, meta analyses) National surveys (e.g. MEPS) or datasets Expert opinion
Above may not represent true societal economic costs
3. Analysis of cost data
Inflation adjustment
•
Consumer Price Index, published by Bureau of Labor Statistics
Discounting of future costs to present value
•
Often 3% used initially and 0% - 10% in sensitivity analyses
Statistical tests depend on research design
Issues, specific to the distribution of cost data:
• • •
Excess zeroes (Consider 2-part models) Skewed (log/other transformation, especially if sample is small) Presence of outliers which are valid data points (high cost pts) 17
Measuring effectiveness: quality-adjusted life years (QALYs)
Allow comparisons across HC interventions and across sectors of the economy
Combine life expectancy & health-related quality of life (HRQL)
HRQL measured on 0 (death) - 1 (perfect health) scale:
• •
HRQL among patients with breast cancer = 0.77 (Kwon, JS2010) 10 years of life expectancy is adjusted to 7.7 QALYs
18
Measuring health related quality of life (HRQL)
Direct methods
•
Standard Gamble, Time Trade Off
Preference based surveys
• •
Generic surveys: HUI, SF-36, Quality of Well-Being, EQ-5D Disease specific surveys
External sources
• •
Published articles Metaanalyses
19
HRQL surveys used in pediatric populations
• • • • • •
EQ-5D youth The Child Health Utility 9D (CHU9D) HUI Quality of Life Inventory (Peds-QL) Child Health and Illness Profile (CHIP) Child Health Questionnaire (CHQ)
Ideally, use preference-based surveys in CEA Often there is no data on HRQL for children/adolescents Useful reference: Solans, M, Pane, S et al. 2008. Health-Related Quality of Life Measurement in Children and Adolescents: A Systematic Review of Generic and Disease-Specific Instruments. Value in Health 11(4), 742-764.
Other measures of effectiveness
When final outcomes (QALYs) can not be measured
• • •
Example: Intervention to reduce blood pressure Causality b/n intermediate (blood pressure) and final outcomes (QALYs) is expected in theory CEA with short analytic time frame will not capture improvements in QALYs but will capture changes in blood pressure
In CEA of tests for conditions for which no Tx exists
•
E.g. cost per additional diagnosis (rather than cost/QALY)
21
Incorporating uncertainty in CE estimates
ICER usually relies on estimates of costs and effectiveness
1 way, 2 way, …, n-way sensitivity analyses
Probabilistic analyses (using random draws from the distribution of each uncertain parameter)
• •
Confidence intervals Cost-effectiveness acceptability curves
22
Incorporating uncertainty: ICER estimates
$900,000
Incremental costs
I
II
$400,000
-$100,000
-$600,000
-$1,100,000 -12
-9
-6
-3
0
3
6
9
Incremental life expectancy 23
12
Cost-effectiveness acceptability curves
Designing CEA
• • • • • • • • • • • • • •
Define interventions to be compared Define the perspective of analysis Determine the analytic time frame Sources of cost data: own, surveys, cost studies, insurance rates What cost categories will be included? Describe the statistical analysis of costs Inflation adjustment Discounting of future costs Justify the measure/s of effectiveness & how each will be obtained If HRQL: surveys, direct methods, or literature reviews? Calculate ICER Sensitivity analyses, 95% CI and CE acceptability curves Decision model to be used (if any), i.e. Markov models Summarize everything briefly in results/grant proposals
25
Criticisms of CEA
• CEA may not result in “fair” distribution of resources • Methodological challenges in the estimation of ICER • CEA results not always intuitive / easy to understand by the public • Practically impossible to evaluate all health technologies which creates bias towards the status quo
Useful reference Gluck ME. Incorporating Costs into Comparative Effectiveness Research AcademyHealth’s 2009. National Health Policy Conference.
26
CEA vs. other methods
• Cost analysis, cost minimization, cost-offset, • Cost-benefit analysis • Comparative effectiveness • Budget Impact Analysis • Cost-consequence analysis
27