HEALTH ECONOMICS
Sponsored by a Grant TÁMOP-4.1.2-08/2/A/KMR-2009-0041 Course Material Developed by Department of Economics,
Faculty of Social Sciences, Eötvös Loránd University Budapest (ELTE) Department of Economics, Eötvös Loránd University Budapest
Institute of Economics, Hungarian Academy of Sciences Balassi Kiadó, Budapest
2
Authors: Éva Orosz, Zoltán Kaló and Balázs Nagy Supervised by Éva Orosz
June 2011
Week 11
Steps of economic evaluation
Authors: Zoltán Kaló and Balázs Nagy Supervised by Éva Orosz
Steps of economic evaluation
Definition of programme
• A comprehensive description of the competing alternatives
• Who did what to whom, where and how often?
Calculation of net health benefits
• Estimation of potential types of health status changes
• Incremental life years, quality of life etc.
• Consideration of time preference
• Calculation of net health benefit Calculation of net costs
• Calculation of programme costs
• Calculation of monetary savings
• Discounting costs and savings
• Calculation of net costs
3 Application of decision rules
• Selection of the appropriate decision rule
• Application of decision rule Sensitivity analysis
• Calculation of costs and consequences according to plausible values of uncertain variables
• Assessment effect on decisions
Definition of programme Selection of the comparator
• Most critical question in economic evaluations – incremental cost-effectiveness compared to which technology?
• What is the appropriate comparator?
– new drug therapy vs.?
– screening programme vs.?
– diagnostic procedure vs.?
Rule of thumb in selection of comparator
• Select that (current, old or even future) technology, which competes with the new technology
• Consider completed therapy (not only partial intervention)
4
Calculation of net health benefits Opinion of experts
The greatest danger to a man with high blood pressure lies in its discovery, because then some fool is certain to try and reduce it.
Hay JH, BMJ 1931, 2:432-47
Hypertension may be an important compensatory mechanism which should not be tampered with, even were it certain that we could control it.
White PD, Heart Disease, New York: McMillan, 1931:400
Source: Moser M (2006). Historical Perspectives on the Management of Hypertension.
The Journal of Clinical Hypertension 8 (s8), 15–20.
Hierarchy of evidence
5
Most crucial factor of evidence gathering
• Systematic review of the literature – unbiased collection of evidence
• What makes a review “systematic”?
– Based on a clearly formulated question – Identifies relevant studies
– Appraises quality of studies
– Summarizes evidence by use of explicit methodology – Comments based on evidence gathered
Need for comparative effectiveness research (indirect comparison)
• Control therapy in the pivotal clinical trials is not relevant locally
• Potential reasons:
• control therapy is out-of-date
− not used in clinical practice anymore
− registration is withdrawn (e.g. side-effects)
• therapeutic guidelines
− control therapy is used first-line, new therapy is only second-line due to its high price
• local reimbursement status
− control therapy is not reimbursed
6
Meaningful health benefit for patients
• Life years (not only 5 year survival)
• Quality of life
• Combinations
– QALY (Quality Adjusted Life Years) – HYE (Healthy Year Equivalents) – DALY (Disablity Adjusted Life Years)
Efficacy vs. effectiveness
• Efficacy & safety: registration criteria (for pharmaceuticals)
• New medicines may result in different health gain in real world – rare but clinically significant adverse events
– less ideal conditions, including copayment: reduced persistence – improved tolerability or easier use: compliance, adherence
• Importance for payers: effectiveness can be measured in naturalistic, observational studies or registries
7
Transferability of health outcomes
Categories Transferability
Epidemiology limited
Efficacy & Effectiveness yes (with limitations)
Life expectancy no (or with serious limitations)
QoL / Utility yes (with limitations)
Transition probabilities yes (with limitations)
8
Calculation of net costs
Relevant costs from different perspectives
Societal • all medical and non medical costs
• productivity loss Third party payer /
health insurance
• total health care costs
• charges linked with reimbursement of providers Health care provider • variable costs that influence the expenses of provider
Patient
• out-of-pocket expenses
• lost wages
• family cost of caring
Employer
• productivity loss
• replacement costs
• risks on routine operation and business objectives
9
Main types of costs
Health care costs Non health care costs
Direct costs
– Direct costs of health care provision (drug, primary care, hospital care, nursing)
– Direct costs of patients (copayment, fee-for non reimbursed and private services, under-the-table payment)
– Travelling costs
– Home nursing
– Sick leave compensation
Indirect costs
– Costs during additional life years
– Lost wages of patients or family supporters
– Productivity costs
– Intangible costs (e.g. pain suffering, anxiety)
– Social value of premature death
10
Transferability of cost vectors among countries
• Production functions of health care services are country specific.
• Transferability of cost vectors – unit cost: non-transferable
– resource utilisation: limited transferability
– relative reduction in resource utilisation: limited transferability
Health care costs
=
Σ resource utilisation x
unit costs
Cost vectors: crucial points
• Consistency of the same cost vectors in different economic evaluations – costing guide
– standard cost vectors published by Public HTA Office (cost dictionary)
• Cost of data collection should be small – payer’s database
– other secondary data sources
11
Sources of local data
• Secondary data collection – publications
– public databases (payer, hospital, registry) – private databases (IMS)
– published tariffs / fees – annual reports
– previous submissions
• Primary data collection
– clinical trials (e.g. subgroup analysis of local patients in Phase III) – non-interventional studies
– survey of health care providers
– patient resource utilisation survey (via clinicians or patient organisations) – market research/Delphi panel (signed by credible clinicians)
Application of decision rules
Economic evaluation does not replace decision!
Incremental cost-effectiveness ratio
12
Decision rule: incremental cost-
effectiveness ratio
13
Determining explicit cost/QALY threshold
• How to establish threshold – ask taxpayers (WTP)
– experienced thresholds (analysis of previous decisions) – benchmark (e.g. dialysis)
– League Table Oregon experiment – GDP (e.g. 3x GDP/capita)
• How many thresholds – single
– two (lower and upper)
– moving threshold relating to disease severity
Moving threshold related to disease severity
As the severity of disease increases, so does the threshold for cost-effectiveness!
0 20000 40000 60000 80000
1 0,8 0,6 0,4 0,2 0
Severity of disease (proportional loss of QALYs)
Cost per QALY
EURO 20,000 per QALY
Increase threshold
Source: Redekop K, 2007
14
Pros and cons of an explicit threshold
Advantages:
• Encourages consistency in decision-making.
• More transparent.
• An implicit threshold would be inferred anyway.
Disadvantages:
• Hard to determine what the threshold should be.
• Does not tell us the opportunity cost of adopting a new technology.
• Other considerations may come into play
• Caveat: new technologies are priced up to threshold
Sensitivity analysis
• Calculation of cost and health outcomes in case of different values of uncertain variables
• Impact analysis on decisions
Sources of uncertainty and solutions
• Uncertainty in the methodology (different results with different methodological approach?) guidelines for economic evaluations, reference data
• Uncertainty from input variables (robustness of results) sensitivity analysis
• Uncertainty from extrapolation (surrogate outcomes to hard endpoints) standard modelling approaches
• Generalizability of specific study results (conclusion valid for total patient population or only for study population?) sensitivity analysis
15
Types of sensitivity analysis
1. Deterministic: input variables with certain numbers – univariate
– bivariate (e.g. confidence box) – several variables
2. Probabilistic (stochastic) – the model calculates outcomes based on random values of variables, therefore the variables are selected by chance based upon their
distribution
– bootstrap method – Monte Carlo simulation
One way deterministic sensitivity analysis:
tornado diagram
16
Two-way sensitivity analysis
Confidence box
17
Probabilistic sensitivity analysis
18
Confidence ellipse
19
