What is Economic Evaluation?
Set of scientific methods to assist decision-makers in
choosing between alternative interventions (“Value for
Money”) to achieve a desired health outcome
Key Concepts
¨ Opportunity Cost
¨ Efficiency, not just Effectiveness
¨ Marginal analysis
Insights from changes in costs and benefits at the
margin instead of ‘totals’ or ‘averages’.
EFN423 Health Economics 4
Evaluation Toolkit: ‘Marginal’ Analysis
¨ Efficiency is achieved by producing/ consuming something
to the point where the (opportunity) cost of the last unit is
no greater than the benefit derived from that unit.
¨ Requires assessment of relative costs and benefits of
each marginal addition or reduction in
production/consumption of health services and products
EFN423 Health Economics 5
The Margin and the Consumer – Diminishing
Marginal Utility
Utility/Benefit
Quantity of Medical
services
Total
Benefit
13.2
13
12
10
0 1 2 3 4
How does the Marginal Benefit look like?
EFN423 Health Economics 6
The Margin and the Producer – Cost Functions
Fixed Costs= Do not vary in short term. Cost
incurred at zero production
Variable Costs = Vary with output at constant rate
Total Cost = Sum of fixed and variable costs
Average Cost = Total cost no of units of output
Marginal Cost = Additional cost of producing one extra unit
of output
EFN423 Health Economics 7
The Margins of Cost and Benefit: Optimisation
Quantity of medical care
Cost,
Benefit
Q2 0
MC
AC
MB
Q* Q1
At Q1, MC > MB; at Q2, MB > MC
The optimum quantity is Q* where MB = MC
EFN423 Health Economics 8
Importance of Marginal Cost – Screening
Costs & Detection of Colon Cancer
EFN423 Health Economics 9
DYI: Find the optimal no. of transplants
Costs and Benefits of Lung Transplantation
No of
Transplants
Total
Cost
Total
Benefit
Average
Cost
Average
Benefit
Marginal
Cost
Marginal
Benefit
1 100 500 100 500 100 500
2 150 900 75 450 50 400
3 180 1200 60 400 30 300
4 230 1400 58 350 50 200
5 330 1500 66 300 100 100
6 530 1550 88 258 200 50
7 1030 1600 147 229 500 50
EFN423 Health Economics 10
Is effectiveness of interventions equal?
Cost minimization study
Can all outcomes be valued in monetary
terms (e. g. willingness to pay?) Cost benefit analysis
Can outcomes be measured as
quality adjusted life years?
Cost-effectiveness analysis
Cost-utility analysis
No
No Yes
Yes
Not always
Yes
Types of economic evaluation
EFN423 Health Economics 11
Method
How are
benefits
measured?
How are results
expressed?
What is the
decision making
rule?
Cost minimisation Proven equal $ Choose that which
costs least
Cost Benefit
Analysis
Health and non
health benefits
valued in
monetary term $
Net present value
(NPV) in $
Benefit cost ratio
NPV > 0;
1
Cost Effectiveness
Analysis
Natural units,
e.g. pain free days,
life years gained,
blood pressure
reduction
Incremental Cost
effectiveness ratio
(ICER)=
That with the lowest
ICER is best value for
money
Cost Utility Analysis Quality Adjusted
Life Years (QALYs)
Incremental Cost
effectiveness ratio:
ICER =
That with the lowest
ICER is best value for
money
EFN423 Health Economics 12
In this unit we will focus on
Cost Effectiveness Analysis
EFN423 Health Economics 13
How to estimate Costs and Benefits?
The ‘counterfactual’:
• The state without the intervention against which the
costs and benefits should be compared
• Problem: Counterfactuals cannot be observed once
intervention is implemented!
Relevant counterfactuals might include:
• Best practice
• Current practice (the status quo)
EFN423 Health Economics 14
Benefit Categories
IInntteerrvveennttiioonn
DDiirreecctt BBeenneeffiittss IInnddiirreecctt BBeenneeffiittss
Savings in
productivity
Improved
patient health
status/utility
Reduced health
services
resource use
Family and
friends’ quality
of life
EFN423 Health Economics 15
How can Health be measured?
¨ Length of life
• Mortality (numbers, rates, SMRs)
• Life expectancy
• Life years lost
¨ Quality of life
• Numerous QoL measures (generic and specific)
• SF-36, Nottingham Health Profile, Guttman Scale,
Rotterdam Symptom Checklist, Hospital Anxiety and
Depression scale etc.
EFN423 Health Economics 16
Sources of Effectiveness Data
- Clinical trials – e.g., RCT’s.
- Epidemiological studies – e.g., cohort studies.
- Synthesis methods – e.g., meta-analyses.
- Use of modelling.
EFN423 Health Economics 17
Cost Effectiveness analysis (CEA)
Outcomes expressed in natural units
e.g., Morbidity reduction, life years saved
¨ Marginal incremental costs and benefits
¨ Compares alternatives to measure the primary
objective of a program
• Aspirin and blockers lengthen life in patients with
heart disease
• But there are more complex, more expensive and
more effective alternatives (angioplasty, stents,
bypass, etc.)
EFN423 Health Economics 18
A new strategy (that is more effective and more costly) is
compared with existing practices to calculate the
Incremental Cost effectiveness ratio (ICER):
¨ Example: Two drugs to treat cholesterol; different cancer
screening programs: is ‘cases detected’ a reliable
measure of effectiveness for cancer?
Cost Effectiveness analysis (CEA)
Effectiveness of (new – existing practice)
ICER Cost of (new – existing practice)
EFN423 Health Economics 19
CEA: Strong vs. weak dominance
More Same Less
More
7
4 2
Same
3
9 6
Less
1
5 8
Key:
Strong dominance for decision
1 = accept treatment
2 = reject treatment
Weak dominance for decision
3 = accept treatment
4 = reject treatment
5 = accept treatment
6 = reject treatment
Non-dominance: no obvious
decisions
7 = Is added effect worth added cost to
adopt treatment?
8 = Is reduced effect acceptable given
reduced cost to adopt treatment?
9 = Neutral on cost and effects. Other
reasons to adopt treatment?
Incremental effectiveness of
treatment compared
to control
Incremental cost of treatment
compared to control
EFN423 Health Economics 20
Cost effectiveness relative to alternatives
+ effect
+ cost
- effect
- cost
Intervention less
effective and more
costly (dominated)
Intervention more
effective and less
costly (dominates)
IV I
III II
EFN423 Health Economics 21
CEA: CER versus ICER
Correct procedure for comparison of alternative health procedures
when treatments are mutually exclusive (when each patient will
receive only one of the alternative treatments):
Example 1: screening program with sequential tests (also see colon
cancer screen test above)
¨ Average (CERs) could be much lower than the Marginals (ICERs).
Total Cost $ Correct Diagnosis (Effect)
Option A 100,000 10
B* 180,000 12
*Assume B includes A plus a follow up treatment (mutually exclusive).
EFN423 Health Economics 22
CEA: CER versus ICER
Example 2: Surgery (A) vs Drug (B) as alternative treatments
¨ ICERs are more informative than CERs
Total Cost $ Life Expectancy (Effect)
Option Surgery 100,000 5
Drug 210,000 7
Note: Surgery and Drug are mutually exclusive and independent
treatments.
In Example 1, the second screen (option B) is not cost-effective.
In Example 2, the Drug is not cost-effective.
EFN423 Health Economics 23
Evaluation of alternative programs using CEA (ICER)
- Rank the options in order of increasing cost.
- Eliminate options for which another more effective but cheaper option
exists (strong dominance). - Calculate the ICER (incremental cost and incremental effect) of each
remaining option with respect to the previous option. - Eliminate options for which the next more effective option has a lower
ICER (weak or extended dominance). - Recalculate the ICERS.
For Example 1 above: - Graphs the 2 options (A & B) with Cost on vertical and Effect on
horizontal. - Calculate CERs. Draw rays through the origin to each. Interpret
slopes. - Calculate the ICER (slope of line from A to B).
EFN423 Health Economics 24
Patient group I Patient group II Patient group III
treatment C E C/E treatment C E C/E treatment C E C/E
A 100 10 10 F 200 12 17 K 100 5 20
B 200 14 14 G 400 16 25 L 200 8 25
C 300 16 19 H 550 18 31 M 300 12 25
D 400 19 21
E 500 20 25
Example: Evaluation of alternative programs
Cost per patient (C) and effectiveness per patient (E) for the hypothetical
available treatment alternatives in 3 different patient groups. There are 1000
patients in each patient group. The cost and effectiveness of the treatments
are calculated relative to a baseline strategy.
EFN423 Health Economics 25
Patient group I Patient group II Patient group III
treatment ΔC ΔE ICER
(ΔC/ΔE)
treatment ΔC ΔE ICER
(ΔC/ΔE)
treatment ΔC ΔE ICER
(ΔC/ΔE)
A 100 10 10 F 200 12 17 K 100 5 20
B 100 4 25 G 200 4 50 L 100 3 33
C 100 2 50 H 150 2 75 M 100 4 25
D 100 3 33
E 100 1 100
Incremental cost (ΔC), incremental effectiveness (ΔE), and ICER for the
hypothetical available treatment alternatives in 3 different patient groups.
There are 1000 patients in each patient group.
Example (contd.)
EFN423 Health Economics 26 - Rank the options in order of increasing cost.
- Eliminate options for which another more effective but cheaper option
exists (strong dominance). - Calculate the ICER (incremental cost and incremental effect) of each
remaining option with respect to the previous option. - Eliminate options for which the next more effective option has a lower
ICER (weak or extended dominance). - Recalculate the ICERS.
- Use a Budget or Pricing rule to determine allocation of resources to
across treatments.
Recall the steps to evaluating alternative programs
EFN423 Health Economics 27
¨ So no strong dominance
¨ C and L are weakly dominated
¨ Recalculate the ICERs after eliminating weakly
dominated options.
Example (contd.)
EFN423 Health Economics 28
Patient group I Patient group II Patient group III
treatment ΔC ΔE ΔC/ΔE treatment ΔC ΔE ΔC/ΔE treatment ΔC ΔE ΔC/ΔE
A 100 10 10 F 200 12 17 K 100 5 20
B 100 4 25 G 200 4 50 M 200 7 29
D 200 5 40 H 150 2 75
E 100 1 100
Example (contd.)
Incremental cost (ΔC), incremental effectiveness (ΔE), and ICER for the
hypothetical available treatment alternatives in 3 different patient
groups, after removing the weakly dominated treatments. There are
1000 patients in each patient group.
EFN423 Health Economics 29
Example (contd.)
EFN423 Health Economics 30
Example (contd.) – Marginal Cost of producing
Effectiveness: Rank according to ICER
EFN423 Health Economics 31
Issues in cost effectiveness
• How are resources/costs categorised?
• Which perspective?
• Discounting
• Trial period and beyond…
EFN423 Health Economics 32
How are resources/costs categorised?
Total Costs
Healthcare Sector
Health professional time
Admin support
Future hospitalisations
Capital costs
Light, Heat, Cleaning
Other Sectors
Productivity
Crime costs
Housing
Education
Social Services
Patient & Family
Travel costs
Treatment time
Waiting time
Out-of-pocket payments
Home modifications
Nursing support
EFN423 Health Economics 33
¨Ideally, we should take a societal perspective- i.e.
measure all the costs (and benefits)
¨Economists measure costs in terms of opportunity
cost- i.e. the cost of using resources is their value
in their next best alternative use.
¨Often the market price will be a useful
approximation
Which perspective?
EFN423 Health Economics 34
¨Future costs are just as valid as current ones
and should be taken into account.
¨If you had to pay a bill of $100 today or $100 in a
year’s time what would you prefer?
¨There is a natural tendency to value things NOW
more highly than in the future.
¨One obvious reason is that you could put the
money in the bank and it would grow to (1+r)*100
in a year at an annual interest rate of r%.
Discounting
EFN423 Health Economics 35
¨E.g. if interest rate is 10% will grow to $110 by
year end.
¨Another way of saying the same thing is that you
would be indifferent between $100 in 1 year and
$100/(1+r) now. i.e. $100/1.1= $90.91
¨Calculating total costs by discounting future costs
over time- is called calculating the present value.
Discounting – example
EFN423 Health Economics 36
¨ Most people with diabetes are afflicted with some type of vascular disease
¨ Traditional method targets BP threshold
¨ Time consuming, multiple trips to GP, tablets to lower blood pressure
¨ Ignores those with diabetes with ‘less than threshold’ BP – despite the fact that BP
still a major determinant of vascular disease.
¨ Alternative is to prescribe BP lowering drugs to ALL type 2 diabetes
¨ An ADVANCE trial-based study looked at the impact of prescribing drugs to lower
blood pressure versus placebo regardless of level of hypertension. This
prescription was on top of any other tablets used in treatment.
¨ Placebo was the control; additional prescription to lower BP was the treatment
¨ Random assignment to control and treatment.
¨ Recruitment 2001-2003
¨ BP treatment ended in 2007
¨ 20 country study including an Australian set of 978 (11140 in total).
¨ The study restricted to only direct health care costs.
Case study – Example:
Cost effectiveness of lowering blood pressure in type 2 diabetes patients (Glasziou et al MJA 2010)
EFN423 Health Economics 37
Key points of analysis:
¨ Survival estimated from survival curves
• Derived life expectancy if survived trial
¨ Measurement of quality of life
• Quality of life EQ 5D survey instrument
¨ Resource usage and unit cost
• Resource usage – diagnosis related group DRG
categories, Medicare claims data
Cost effectiveness and cost utility analysis
EFN423 Health Economics 38
EFN423 Health Economics 39
EFN423 Health Economics 40
EFN423 Health Economics 41
¨ Hospitalisation cost almost twice the average (Box 3)
¨ Hospitalisation cost saving for treated patient group
at $410 per patient
¨ Cost effectiveness:
• Net difference in cost between treatment and control = $555…
What drives this??
• Cost of therapy offset partly (60%) due to lower hospital costs,
outpatient visit costs, other drug therapies)
¨ Extra life years: 0.1
¨ Cost per life year saved approx. $6200.
Australian data:
EFN423 Health Economics 42
At a 5% discount rate:
• Increment in cost = $502 per patient
• Increment in benefits: 0.05 life years saved
(Use Present Discounted Value PDV to calculate the above two)
• Cost-effectiveness ratio = $10,040 per discounted life year saved
EFN423 Health Economics 43
Shortcomings of CEA
Remember: CEA is an aid to decision making, not a procedure for making such
decisions in healthcare
¨ Good CEAs require appropriate funding, enough to cover the extended
timescale and level of detail required
¨ Cost estimates may vary across region, across population. For example,
reasonable values of opportunity costs of time for population groups outside of
labour market (children, retired persons, those unable to work) are difficult to
obtain. Standard cost estimates should be established.
¨ How benefits should be valued over time is of critical importance when the
benefit accrue over a long time (ICER can be altered by the rate at which
benefits are discounted over time). However, discount rate is quite subjective.
¨ It is not clear whether we should incorporate the cost of unrelated future
diseases.
The post Week 11: Economic Evaluation of Health Intervention appeared first on My Assignment Online.