BEHAVIORAL DECISION MAKING
LECTURE 1 &2. INTRODUCTION AND THEORY
How can economic thinking help understand the behavior of patients, providers, payers and
policy makers?
- We use behavioral economic models to describe individual’s behavior.
Economic evaluation is a comparison of costs and benefits.
How can we express the benefits of health care numerically?
Ignore health benefits- costs minimization.
Express benefits as life-years gained
Express benefits as life-years gained and adjust for the quality of life.
Deriving a monetary value of a QALY.
In case of non-dominance (a new program produces better outcomes at additional costs),
CUA gives us price of achieving a goal in terms of incremental costs of a QALY gained.
But we don’t know yet if such a goal is worth achieving given the opportunity costs of the
resources consumers. You need to place a money value on health outcomes to make any
resource allocation decision using CUA data.
Allocative efficiency.
Choice of what health care to provide: which programs are worthwhile, given the alternative
uses of resources?
This is done by assigning relative values to health and non-health related goals.
How to provide health care➔ strive for minimum input for given output.
Principles of CBA
• There must be a common unit of measurement.
• All the benefits and costs of the program must be measured in terms of their
equivalent money value.
• The valuation of benefits and costs should reflect preferences revealed by choices.
• The benefits are usually measured by market choices.
• The marginal benefit should be equal to the marginal price.
• The gross benefits of consumption are an area under the demand curve.
Advantages of CBA
• It helps to allocate scarce resources to programs that maximize societal economic
benefit
• It studies the full economic impact of all potential outcomes of an intervention
• It enabled comparing different programs having different health outcomes, or health
programs to non-health programs.
• It allows analysts to examine its distributional aspects; who will receive these
benefits and who will bear the costs.
,Limited use of CBA’s
• CEA and CUA are more common
• Societal perspective is often not fully explored
• Reasons
o Practical measurement problems
o Nature of the commodity health care
Threshold value can be based on;
• Opportunity costs approach (supply side)
• WTP approach. (demand side)
Opportunity costs approach
• There is a fixed healthcare budget
• Each new technology will displace some existing services
• Cost per QALY of new intervention should be lower than these to be considered cost-
effective
• Also called the shadow price of the budget constraint
• Based in UK
WTP approach
• Links CUA with welfare economics
• Compares cost per QALY of intervention to WTP-based threshold value
• May tell something about welfare effects of new intervention
• Utility loss of foregone private consumption equivalent to utility gain due to an
increase of 1 additional QALY
• Flexible budget.
Welfare economics;
• Addresses normative questions because it embodies value judgements.
• Key assumptions
o Social welfare is made up from the welfare of each individual member of
society
o Individuals are the best judges of their own welfare.
Approaches to estimate WTP
• Review of past decisions
• Human capital approach
• Monetary value of preventing fatalities
• Contingent valuation
• Discrete choice experiments (DCEs) with a cost attribute
VSL
• Marginal rate of substitution between wealth mortality risk
• Commonly used in economic evaluation of life-saving policies in transport and
environmental economics.
• Estimation can be done by
, o SP:
o RP: wage-risk studies➔ wage premium for hobs with greater risks of death
▪ weaknesses
• Wide variation of estimates
• Context and job-specific
o WTP-Q: can be estimated from VSL by using present-discounted value
formula
Contingent Valuation
• Perspective: individual or societal (including altruism)
• Open-ended or closed-ended
• Sample: patient-specific or general population
• QALY change based on QoL or life expectancy improvement
• General health improvement or specific disease
• Payment vehicle (insurance premium, out of pocket, taxes)
Problem→ sensitivity to irrelevant information
1. Starting point bias
2. Range effect
3. WTP/WTA disparity
Problem→ insensitivity to relevant information
1. Scope effects
2. Reduction of 3 in 100,000
3. WTP is 18 dollar
4. VoL = 0.6 million
Pinto et al (2009). It seems to be necessary to relax several assumptions when estimating the
monetary value of a QALY:
- Non EU (expected utility)
- Nonlinear QALY model
- But also procedural invariance→ outside conventional economic theory, need to
better understand psychological processes.
Stated preference studies give lower estimates than VSL studies.
Higher values when risk of death is included than when pure quality of life changes are being
valued.
There is insensitivity to scope: larger QALY changes give lower WTP-Q estimates.
Introduction to the QALY model.
Most used utility model = Quality-Adjusted Life-Years (QALY).
QALY
Advantages:
• Intuitively appealing
• Easy to use in practice.
Disadvantages:
• May be too simple.
, How restrictive is the QALY model?
How can we determine the utilities H(Q)?
Assumptions:
- Health states are chronic
- Health states are preferred to death
- Expected utility holds.
Standard gamble
LECTURE 4 BIASES IN HEALTH UTILITY MEASUREMENT
2 QALY models:
- Additive
- Multiplicative models
Which QALY model do we want to prove= linear QALY model.
U(T,Q) = H(Q) * T ➔ Linear QALY model.
Difference general and linear QALY model =
Y(T,Q) = (Lt) * (H(Q)= general QALY model.
o Lt.
Main difference between the two models.
General model can be nonlinear in time
The linear QALY model presupposed the linear model is linear.
General model does not imply any functional form.
Expected utility is about risk and about gambles. It implies that you multiply utility of
outcomes by probability.
(Q1,T1)p. the probability of having 100% or having 0%.
Gamble for life.
40 years , 50% 0 years. OR 20 years for sure.
L(50) * p (50%) + 1-p *L(0).
Indifferent: risk neutral. They do not have a preference for either one option or another. Risk
neutral: what is the expected value of the gamble (in utility).
LECTURE 1 &2. INTRODUCTION AND THEORY
How can economic thinking help understand the behavior of patients, providers, payers and
policy makers?
- We use behavioral economic models to describe individual’s behavior.
Economic evaluation is a comparison of costs and benefits.
How can we express the benefits of health care numerically?
Ignore health benefits- costs minimization.
Express benefits as life-years gained
Express benefits as life-years gained and adjust for the quality of life.
Deriving a monetary value of a QALY.
In case of non-dominance (a new program produces better outcomes at additional costs),
CUA gives us price of achieving a goal in terms of incremental costs of a QALY gained.
But we don’t know yet if such a goal is worth achieving given the opportunity costs of the
resources consumers. You need to place a money value on health outcomes to make any
resource allocation decision using CUA data.
Allocative efficiency.
Choice of what health care to provide: which programs are worthwhile, given the alternative
uses of resources?
This is done by assigning relative values to health and non-health related goals.
How to provide health care➔ strive for minimum input for given output.
Principles of CBA
• There must be a common unit of measurement.
• All the benefits and costs of the program must be measured in terms of their
equivalent money value.
• The valuation of benefits and costs should reflect preferences revealed by choices.
• The benefits are usually measured by market choices.
• The marginal benefit should be equal to the marginal price.
• The gross benefits of consumption are an area under the demand curve.
Advantages of CBA
• It helps to allocate scarce resources to programs that maximize societal economic
benefit
• It studies the full economic impact of all potential outcomes of an intervention
• It enabled comparing different programs having different health outcomes, or health
programs to non-health programs.
• It allows analysts to examine its distributional aspects; who will receive these
benefits and who will bear the costs.
,Limited use of CBA’s
• CEA and CUA are more common
• Societal perspective is often not fully explored
• Reasons
o Practical measurement problems
o Nature of the commodity health care
Threshold value can be based on;
• Opportunity costs approach (supply side)
• WTP approach. (demand side)
Opportunity costs approach
• There is a fixed healthcare budget
• Each new technology will displace some existing services
• Cost per QALY of new intervention should be lower than these to be considered cost-
effective
• Also called the shadow price of the budget constraint
• Based in UK
WTP approach
• Links CUA with welfare economics
• Compares cost per QALY of intervention to WTP-based threshold value
• May tell something about welfare effects of new intervention
• Utility loss of foregone private consumption equivalent to utility gain due to an
increase of 1 additional QALY
• Flexible budget.
Welfare economics;
• Addresses normative questions because it embodies value judgements.
• Key assumptions
o Social welfare is made up from the welfare of each individual member of
society
o Individuals are the best judges of their own welfare.
Approaches to estimate WTP
• Review of past decisions
• Human capital approach
• Monetary value of preventing fatalities
• Contingent valuation
• Discrete choice experiments (DCEs) with a cost attribute
VSL
• Marginal rate of substitution between wealth mortality risk
• Commonly used in economic evaluation of life-saving policies in transport and
environmental economics.
• Estimation can be done by
, o SP:
o RP: wage-risk studies➔ wage premium for hobs with greater risks of death
▪ weaknesses
• Wide variation of estimates
• Context and job-specific
o WTP-Q: can be estimated from VSL by using present-discounted value
formula
Contingent Valuation
• Perspective: individual or societal (including altruism)
• Open-ended or closed-ended
• Sample: patient-specific or general population
• QALY change based on QoL or life expectancy improvement
• General health improvement or specific disease
• Payment vehicle (insurance premium, out of pocket, taxes)
Problem→ sensitivity to irrelevant information
1. Starting point bias
2. Range effect
3. WTP/WTA disparity
Problem→ insensitivity to relevant information
1. Scope effects
2. Reduction of 3 in 100,000
3. WTP is 18 dollar
4. VoL = 0.6 million
Pinto et al (2009). It seems to be necessary to relax several assumptions when estimating the
monetary value of a QALY:
- Non EU (expected utility)
- Nonlinear QALY model
- But also procedural invariance→ outside conventional economic theory, need to
better understand psychological processes.
Stated preference studies give lower estimates than VSL studies.
Higher values when risk of death is included than when pure quality of life changes are being
valued.
There is insensitivity to scope: larger QALY changes give lower WTP-Q estimates.
Introduction to the QALY model.
Most used utility model = Quality-Adjusted Life-Years (QALY).
QALY
Advantages:
• Intuitively appealing
• Easy to use in practice.
Disadvantages:
• May be too simple.
, How restrictive is the QALY model?
How can we determine the utilities H(Q)?
Assumptions:
- Health states are chronic
- Health states are preferred to death
- Expected utility holds.
Standard gamble
LECTURE 4 BIASES IN HEALTH UTILITY MEASUREMENT
2 QALY models:
- Additive
- Multiplicative models
Which QALY model do we want to prove= linear QALY model.
U(T,Q) = H(Q) * T ➔ Linear QALY model.
Difference general and linear QALY model =
Y(T,Q) = (Lt) * (H(Q)= general QALY model.
o Lt.
Main difference between the two models.
General model can be nonlinear in time
The linear QALY model presupposed the linear model is linear.
General model does not imply any functional form.
Expected utility is about risk and about gambles. It implies that you multiply utility of
outcomes by probability.
(Q1,T1)p. the probability of having 100% or having 0%.
Gamble for life.
40 years , 50% 0 years. OR 20 years for sure.
L(50) * p (50%) + 1-p *L(0).
Indifferent: risk neutral. They do not have a preference for either one option or another. Risk
neutral: what is the expected value of the gamble (in utility).
