Analyzing costs

Question 1

Mean/average

Answer 1

Unbiased
Sensitive to outliers
- Precision (SE) looks at the standard deviation
- Confidence Intervals (CI) looks at the range of values that are likely to cover the true (unknown) population mean value

Question 2

Median costs

Answer 2

Less influences by outliers

Question 3

Net monetary benefit (NMB)

Answer 3

NMB > CEAC
- Only shows if ICER is acceptable at lambda, has no info on if its actually preferred.

= (lambda x difference E) - difference C

Question 4

Key takeaways when analyzing costs (4)

Answer 4

1. Collect the right data
2. Pay attention to uncertainty in results
3. Report results of relevant subgroups
4. Interpretation is key

Question 5

Consistency argument

Answer 5

Equal CE ratios at different points in time should receive equal priority (same discount)
- Highly questionable (do not consider changing costs etc.)

Question 6

Paralyzing paradox

Answer 6

When discount health < money, the nominator shoul be delayed.

• Endless prosponing
• Experts say that lower discount rate health < costs is preferred.

Question 7

Allocative efficiency

Answer 7

Cost-benefit analysis

How much goes to HC? Look at different sectors

Question 8

Productive efficiency

Answer 8

Cost-effect analsysis

Productivity of each procedure?

Question 9

Maximum acceptable ICERS (4)

Answer 9

1. Rule of thumb
2. League table
3. Revealed preference
4. Stated preference

Question 10

Rule of thumb

Answer 10

Historical data/knowledge 
Create categories (WTP)

Question 11

League table

Answer 11

Rank and calculate cumulative costs

If info is available

Question 12

Revealed preference

Answer 12

Examine the WTP-decision makers

Microlevel too possible

Question 13

Stated preference

Answer 13

Estimate values (median) 
Choice modelling (preference of characteristics) 
Contingent valuation (valuation of goods/services)

Question 14

Prevalence

Answer 14

Cases/1000 with a given disease

Question 15

Incidence

Answer 15

New cases/1000 with disease at given time. E.g.,

28 in 2 years, then 14/1000 person-year

Question 16

Risk

Answer 16

Relative risk/risk difference

R1/Ro = exposure

Question 17

Odds ratio

Answer 17

Case control studies

P/1-P)/(Q/1-Q

Question 18

Hazard ratio

Answer 18

Difference in survival
Assumed to be constant
= Risk/cases

Question 19

CEAC

Answer 19

% CE bepaald door threshold.
CEAC < NMB, since NMB adds analysis of patient level data which is easier to itnerpret the ICERs and ensures a closer normal distribution

Question 20

Monte Carlo

Answer 20

Random values
Probabilistic for modelling risk and uncertainty
Better than deterministic since it measures not only expected outcomes but also the probability of the occurrence of the outcome.