Behavioral Decision Making in Health
Introduction- March 29
HEPL/ HE: improving healthcare and health policy
- What challenges doe health policy makers face?
o How to measure people’s preferences for health?
o How to spend the money the most efficiently?
- How can economic thinking help understand the behavior of patients, providers, payers and
policy makers?
o We use (behavioral) economic models to describe individuals’ behavior
- How should empirical research in healthcare be conducted and interpreted?
o How can we incorporate empirical evidence of behavioral biases into policy making?
Aims of this course
- To help you get a sound understanding of the methodological difficulties surrounding
economic evaluation of healthcare and their possible solutions
- To encourage you to critically reflect on the tools used in economic evaluations of healthcare
- Introduction into the most recent evidence obtained in the study of decision theory in
health, using a behavioral economic approach
- Inform you of key insights derived from such experiments and discuss their implications for
health economic evaluations
Week 1
- Introduction
- Theoretical properties of the QALY model
- Empirical evidence on the underlying assumptions
Week 2
- Biases in health utility measurement
o In what way do people violate rationality axioms when measuring health
preferences?
o How does this affect economic evaluations
- Workgroup to practice with the material
Week 3
- Discounting
o How do we deal with delays in costs and effects
- Which method to use?
o What works best?
Week 4
- Workgroup on child health valuation
- Which method to use? Part 2
- From theory to practice: correcting methods for health state valuation
Week 5
- Equity weighing (part 1 and 2)
o How should we aggregate individual preferences
o Should we all get the same weight or not?
o What is we have both risk and inequality
Week 6
- Workgroups
o Child health valuation, part 2
o Exercises
1
, - Q&A sessions
Lecture 2- Theoretical properties of QALYs- March 29
Position in this course
- Topics Behavioural Decision Theory in Health:
o Quality-Adjusted Life Years (QALYs)
o Monetary valuation of health
o Time preference (discounting)
o Utility measurement (mainly in situations of risk)
o Equity
- In effect, we study QALYs in three dimensions:
o Under risk
o Over time (discounting)
o At the societal level (equity)
Total Healthcare expenditures per head in 2010
- A lot higher in the US than in the Netherlands
Criticism on putting a value on human life
- Cost-cutting device
- Human life is priceless
- Leave it to the politicians
Economic evaluation
- Comparison of costs and benefits
- Central question: how can we express the benefits of healthcare numerically?
Approaches to economic evaluations
- Ignore health benefits- costs minimizations: CMA
- Express benefits as life-years gained: CEA
- Express benefits as life-years gained and adjust for quality of life: CUA
How to express this numerically
1. Deriving a monetary value of a QALY
Example: road inuries
- U: unconscious followed by death
- You face an annual risk of 6 in 100,000 to end up in state U due to a road accident. How
much are you willing to pay for a safety device that reduces this risk with 1 in 100,000?
- The value of life
o 12 Dollars WTP
o Value of life: 12/0.00001= 1.2 M
Problem 1: sensitivity to irrelevant information
- Starting point bias
- Range effect
o The elicited WTP depends on if and, if so, which comparators are used in the
question. A separate evaluation may for instance laed to a higher value than when
the valuated item is compared to a better item
- WTP/ WTA disparity
2
,Problem 2: insensitivity to relevant information
- Scope effects
o WTP values are not sensitive to the amount of the good that is being valued
o Reduction of 3 in 100,000
o WTP= 18 dollar
o VoL= 18/0.00003= 0.6 million
Meta- analysis by Ryen and Svensson
- Overall mean WTP-Q of 118,839, median of 24,226
o Skewed: median lower than the mean
- 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
- WTP-Q not constant across QALY changes: larger QALY changes give lower WTP-Q estimates
(insensitivity to scope)
2. Introduction to the QALY model
Most common model: Quality-Adjusted Life-Years (QALYs)
- Additive model
- Let (q1,…,q2) be a health profile
- QALY model: U(q1,..,qt)= SUM H(qt)
- Chronic health: U(Q,T)= H(Q)* T
- Advantages
o Intuitively appealing
o Easy to use in practice
- Disadvantages
o May be too simple
Questions related to QALY model
- How restrictive is the QALY model?
- How can we determine the utilities H(Q)?
Purpose QALY model
- Represent preferences
- If X > Y then V(X) < V(Y)
- If V(X)> V(Y) then X< Y
Assumptions for the next slides
- Health states are chronic
- Health states are preferred to death
- Expected utility holds
Expected utility
- EU ((Q1,T1),p,(Q2,T2))= pU(Q1,T1) + (1-p) U(Q2,T2)
o P is probability
o U is utility
o Q1 is what you get: full health for example
o T, between brackets is life years
- Two points of U can be chosen freely
3
, Standard gamble
- ( Back pain , 30 y . ) ( ( Full health ,30 y . ) , p , Death )
- Apply Expected Utility:
o U ( Back pain , 30 y . ) =p∗U ( Full health ,30 y . ) + ( 1− p )∗U ( Death)
- Apply scaling:
o U ( Full health, 30 y . )=1
o U ( Death)=0
o U ( Back pain , 30 y . ) =p
- What about the 30 years?
3. Original derivation of the QALY model
First characterization linear QALY model
- Pliskin, Shepard & Weinstein (1980)
- 3 conditions
o (mutual) utility independence
o constant proportional trade-off
o risk neutrality wrt life duration
Mutual utility independence
Quality of life utility independent of life duration
- (Back Pain, rest of life) ((FH, rest of life), 2/3, (Death, rest of life))
- Then also
- (Back Pain, 10y.) ((FH, 10y.), 2/3, (Death, 10y.))
Here life durations changed from rest of life to 10 years
Life duration UI of Quality of life
- (Back pain, 20y.) ((Back pain, 40y.), 2/3, (Back pain, 10y.))
- Then also
- (Full health, 20y.) ((Full health, 40y.), 2/3, (Full health, 10y.))
Here back pain changed to full health
Intermediate result
- The following statements are equivalent
o Utility independence holds
o U is either additive, U(Q,T)= H(Q)+ L(T), or multiplicative U(Q,T)= H(Q) * L(T)
Hence
- To arrive at the QALY model must
o 1) exclude the additive model and
o 2) ensure linearity of L(T)
Standard gamble, part 2
- ( Back pain , 30 y . ) ( ( Full health ,30 y . ) , p , Death )
- Apply Expected Utility:
o U ( Back pain , 30 y . ) =p∗U ( Full health ,30 y . ) + ( 1− p )∗U ( Death , 30 y .)
- Apply utility independence:
o Additive: H ¿
o Multiplicative: H ¿)¿ L(30 y .)¿
- Apply scaling:
4
Introduction- March 29
HEPL/ HE: improving healthcare and health policy
- What challenges doe health policy makers face?
o How to measure people’s preferences for health?
o How to spend the money the most efficiently?
- How can economic thinking help understand the behavior of patients, providers, payers and
policy makers?
o We use (behavioral) economic models to describe individuals’ behavior
- How should empirical research in healthcare be conducted and interpreted?
o How can we incorporate empirical evidence of behavioral biases into policy making?
Aims of this course
- To help you get a sound understanding of the methodological difficulties surrounding
economic evaluation of healthcare and their possible solutions
- To encourage you to critically reflect on the tools used in economic evaluations of healthcare
- Introduction into the most recent evidence obtained in the study of decision theory in
health, using a behavioral economic approach
- Inform you of key insights derived from such experiments and discuss their implications for
health economic evaluations
Week 1
- Introduction
- Theoretical properties of the QALY model
- Empirical evidence on the underlying assumptions
Week 2
- Biases in health utility measurement
o In what way do people violate rationality axioms when measuring health
preferences?
o How does this affect economic evaluations
- Workgroup to practice with the material
Week 3
- Discounting
o How do we deal with delays in costs and effects
- Which method to use?
o What works best?
Week 4
- Workgroup on child health valuation
- Which method to use? Part 2
- From theory to practice: correcting methods for health state valuation
Week 5
- Equity weighing (part 1 and 2)
o How should we aggregate individual preferences
o Should we all get the same weight or not?
o What is we have both risk and inequality
Week 6
- Workgroups
o Child health valuation, part 2
o Exercises
1
, - Q&A sessions
Lecture 2- Theoretical properties of QALYs- March 29
Position in this course
- Topics Behavioural Decision Theory in Health:
o Quality-Adjusted Life Years (QALYs)
o Monetary valuation of health
o Time preference (discounting)
o Utility measurement (mainly in situations of risk)
o Equity
- In effect, we study QALYs in three dimensions:
o Under risk
o Over time (discounting)
o At the societal level (equity)
Total Healthcare expenditures per head in 2010
- A lot higher in the US than in the Netherlands
Criticism on putting a value on human life
- Cost-cutting device
- Human life is priceless
- Leave it to the politicians
Economic evaluation
- Comparison of costs and benefits
- Central question: how can we express the benefits of healthcare numerically?
Approaches to economic evaluations
- Ignore health benefits- costs minimizations: CMA
- Express benefits as life-years gained: CEA
- Express benefits as life-years gained and adjust for quality of life: CUA
How to express this numerically
1. Deriving a monetary value of a QALY
Example: road inuries
- U: unconscious followed by death
- You face an annual risk of 6 in 100,000 to end up in state U due to a road accident. How
much are you willing to pay for a safety device that reduces this risk with 1 in 100,000?
- The value of life
o 12 Dollars WTP
o Value of life: 12/0.00001= 1.2 M
Problem 1: sensitivity to irrelevant information
- Starting point bias
- Range effect
o The elicited WTP depends on if and, if so, which comparators are used in the
question. A separate evaluation may for instance laed to a higher value than when
the valuated item is compared to a better item
- WTP/ WTA disparity
2
,Problem 2: insensitivity to relevant information
- Scope effects
o WTP values are not sensitive to the amount of the good that is being valued
o Reduction of 3 in 100,000
o WTP= 18 dollar
o VoL= 18/0.00003= 0.6 million
Meta- analysis by Ryen and Svensson
- Overall mean WTP-Q of 118,839, median of 24,226
o Skewed: median lower than the mean
- 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
- WTP-Q not constant across QALY changes: larger QALY changes give lower WTP-Q estimates
(insensitivity to scope)
2. Introduction to the QALY model
Most common model: Quality-Adjusted Life-Years (QALYs)
- Additive model
- Let (q1,…,q2) be a health profile
- QALY model: U(q1,..,qt)= SUM H(qt)
- Chronic health: U(Q,T)= H(Q)* T
- Advantages
o Intuitively appealing
o Easy to use in practice
- Disadvantages
o May be too simple
Questions related to QALY model
- How restrictive is the QALY model?
- How can we determine the utilities H(Q)?
Purpose QALY model
- Represent preferences
- If X > Y then V(X) < V(Y)
- If V(X)> V(Y) then X< Y
Assumptions for the next slides
- Health states are chronic
- Health states are preferred to death
- Expected utility holds
Expected utility
- EU ((Q1,T1),p,(Q2,T2))= pU(Q1,T1) + (1-p) U(Q2,T2)
o P is probability
o U is utility
o Q1 is what you get: full health for example
o T, between brackets is life years
- Two points of U can be chosen freely
3
, Standard gamble
- ( Back pain , 30 y . ) ( ( Full health ,30 y . ) , p , Death )
- Apply Expected Utility:
o U ( Back pain , 30 y . ) =p∗U ( Full health ,30 y . ) + ( 1− p )∗U ( Death)
- Apply scaling:
o U ( Full health, 30 y . )=1
o U ( Death)=0
o U ( Back pain , 30 y . ) =p
- What about the 30 years?
3. Original derivation of the QALY model
First characterization linear QALY model
- Pliskin, Shepard & Weinstein (1980)
- 3 conditions
o (mutual) utility independence
o constant proportional trade-off
o risk neutrality wrt life duration
Mutual utility independence
Quality of life utility independent of life duration
- (Back Pain, rest of life) ((FH, rest of life), 2/3, (Death, rest of life))
- Then also
- (Back Pain, 10y.) ((FH, 10y.), 2/3, (Death, 10y.))
Here life durations changed from rest of life to 10 years
Life duration UI of Quality of life
- (Back pain, 20y.) ((Back pain, 40y.), 2/3, (Back pain, 10y.))
- Then also
- (Full health, 20y.) ((Full health, 40y.), 2/3, (Full health, 10y.))
Here back pain changed to full health
Intermediate result
- The following statements are equivalent
o Utility independence holds
o U is either additive, U(Q,T)= H(Q)+ L(T), or multiplicative U(Q,T)= H(Q) * L(T)
Hence
- To arrive at the QALY model must
o 1) exclude the additive model and
o 2) ensure linearity of L(T)
Standard gamble, part 2
- ( Back pain , 30 y . ) ( ( Full health ,30 y . ) , p , Death )
- Apply Expected Utility:
o U ( Back pain , 30 y . ) =p∗U ( Full health ,30 y . ) + ( 1− p )∗U ( Death , 30 y .)
- Apply utility independence:
o Additive: H ¿
o Multiplicative: H ¿)¿ L(30 y .)¿
- Apply scaling:
4
