Additional knowledge

Question 1

Torpedo diagram (sensitivity)

Answer 1

Low to high
Shows variation towards initial nominal values
Shows correlation
Wide bars need more attention

Question 2

Welfare theory approach

Answer 2

Objective: Maximize individual utility/health
Marginal rate for substition (health for money)
Should be measured in population
Assumes: Perfect divisibility/independence/no uncertainty

Question 3

Resource allocation (RA)

Answer 3

Allocate scarce resources
Threshold is the shadowprice of budget constraints
Assumes: All can be combined, perfectly divisible, constant return.
Not always possible due to constraints.

Decision rule: CE-rate < lambda, but lambda can change due to budget variation. So in practice, sort by ICER and add new ones till budget is gone.

Question 4

Knapsack problem

Answer 4

Individual interventions with fixed size which can be combined
No longer a simple optimim

Question 5

Advice in NL

Answer 5

Use the CE as a starting point for negotiations, so it has X% probability to be CE, thus price drops required of Y%.
–> Appropriate scope needed and dealing with uncertainty

Question 6

Average CE

Answer 6

Comparing slope i to threshold (care as usual)

Question 7

Incremental CE

Answer 7

Comparing slope between 2 treatments to the threshold

Question 8

Extended dominance

Answer 8

Combination of two might still be preferred over third option (easier to leave the third one out by the reimbursement officer)

Question 9

Rate

Answer 9

Potential for occurance of event (added/substracted)

Rate to probability = 1 - EXP(value)

Question 10

Probabilities

Answer 10

Likelihood for event to happen (0-1)

Question 11

EQ5D

Answer 11

5 subjects, few options
Answers have weight
Optimal health (1) - ….

Question 12

Standard gamble (decision tree)

Answer 12

Seen as timeless (1 healthy, 0 death)
Others assume tradeoff between quality and time.
Look at scenario A and then fillin the utility preferred over A
(40% probon death over A)

Question 13

Time trade-off

Answer 13

Example; 15 years in A or 12 years full health = 12/15 = 0.8 
Strong assumptions
1. Linear life duration
2. No loss aversion
3. No scale comparability 
Upwards bias

Question 14

VAS

Answer 14

Rating scales with intervals. Critics:
1. Achors not well defined
2. Biases
3. Results differ towards utility measure (SG)
Too much aversion towards extreme when using scales.

Question 15

Paired Comparison (Thurstone)

Answer 15

Compairing entities in pairs. Assumes:

1. Posses varying attributes
2. Preference exists
3. Unidimensional quality
4. Normal distribution

The greater the preference/likelihood, the bigger the weight

Question 16

Discremental process

Answer 16

Interacts with stimuli