Assessing value for money:
principles, methods and issues
Professor Nancy Devlin
City Health Economics Centre
[Link]@[Link]
1. The role of value for money arguments in
purchasing decisions
• NHS (and private) health care providers make strategic
decisions based largely on the business case
HRG tariff price vs. average cost
Risk assessment
• The demand for providers’ services is determined by PCT
spending decisions
Exercising some degree of discretion,
Within the constraints of their funding and various requirements
and obligations
PCT decision making will increasingly be informed by value for
money considerations
As are decisions by, for example, NICE
Reflects a growing emphasis, in NHS policy, on outcomes
(rather than activity)
2. Principles
• Economic assessments of value for money
have two distinctive characteristics:
Opportunity costs
A focus on marginal analysis
• Focussing on changes in costs (and
benefits) at the margin gives important
insights that can be obscured by average
or total costs (and benefits)
3. Insights from Marginal Analysis
Example 1: Cost savings from early discharge
Source: Drummond et al. (2006)
Methods for the Economic Evaluation of Health Care Programmes
4. A decision making exercise
The Government says that it will earmark a sum for
the prevention of two diseases (Disease A and
Disease B) that are prevalent in your PCT. These
diseases are sometimes fatal, but can be prevented by
suitable procedures.
You are asked to advise on how to spend the money
to maximise the number of premature deaths averted.
The Government hints that the sum will be £1 million.
You ask public health experts, who tell you that the
number of premature deaths averted by spending
£1 million would be:
49 for disease A or 101 for disease B
What would you advise?
The Government now tells you that, at the insistence
of the Treasury, the sum will actually be £500,000.
Again you ask public health experts, who tell you that
the number of premature deaths averted by spending
£500,000 would be
39 for disease A or 81 for disease B
What would you now advise?
Government documents on this decision, including
your advice, are leaked before a crucial by-election in
your region.
The Government announces publicly that they will,
after all, make £1 million available.
What would you now advise?
Deaths averted Average Cost
A B Total A B
£1 m 49 101 101 £20,408 £9,901
£0.5 m 39 81 81 £12,821 £6,173
£1m 39 81 120 £12,821 £6,173
Disease A Disease B
Deaths Cost per Deaths Cost per
Total cost (£)
averted death averted averted death averted
100 000 10 10 000 26 3 846
200 000 19 10 526 43 4 651
300 000 27 11 111 58 5 172
400 000 34 11 765 70 5 714
500 000 39 12 821 81 6 173
600 000 43 13 953 87 6 897
700 000 46 15 217 92 7 609
800 000 48 16 667 96 8 333
900 000 49 18 367 99 9 091
1 000 000 49 20 408 101 9 901
Cost A B
Total Marginal Total Marginal MC Total Marginal MC
£0.5m £0.5m 39 39 £12,821 81 81 £6,173
£1m £0.5m 49 10 £50,000 101 20 £25,000
MC = Marginal cost per death averted
Disease A Disease B
Marginal cost Marginal cost
Deaths Deaths
Total cost (£) per death per death
averted averted
averted averted
100 000 10 10 000 26 3 846
200 000 19 11 111 43 5 882
300 000 27 12 500 58 6 667
400 000 34 14 286 70 8 333
500 000 39 20 000 81 9 091
600 000 43 25 000 87 16 667
700 000 46 33 333 92 20 000
800 000 48 50 000 96 25 000
900 000 49 100 000 99 33 333
1 000 000 49 101 50 000
Amount spent on disease B
900,000 800,000 700,000 600,000 500,000 400,000 300,000 200,000 100,000
100,000 100,000
90,000 90,000
Marginal cost per death averted
Marginal cost per death averted
80,000 80,000
70,000 70,000
60,000 60,000
50,000 50,000
40,000 40,000
30,000 30,000
20,000 20,000
10,000 10,000
0 0
100,000 200,000 300,000 400,000 500,000 600,000 700,000 800,000 900,000
Amount spent on disease A
Disease A Disease B
5. Optimisation rules: the equimarginal principle
Marginal Benefit,
Marginal Cost £
MC
MB
Quantity of medical
care
Q2 Q* Q1
The optimum quantity is Q* where MB = MC
At Q1, MC > MB
At Q2, MB > MC
6. Principal Methods
Economic evaluation is:
‘The comparative analysis of alternative courses of
action in terms of both their costs and consequences’
• Cost minimisation (CM)
• Cost-benefit analysis (CBA)
• Cost-effectiveness analysis (CEA)
• Cost-consequences analysis (CCA)
• Cost-utility analysis (CUA)
What is the
How are benefits How are results
Method decision making
measured? expressed?
rule?
Choose that which
Cost minimisation Proven equal £
costs least
Net present value
Cost Benefit NPV > 0
£ (NPV) in £
Analysis B:C ratio > 1
Benefit cost ratio
Natural units, Cost effectiveness That with the lowest
Cost Effectiveness
e.g. pain free days ratio (CER)= CER is best value for
Analysis
life years gained Costs/outcome money*
In a variety of CERs for each That with the lowest
Cost Consequences
different natural alternative measure CER is best value for
Analysis
units. of effectiveness money*
Cost effectiveness That with the lowest
Quality Adjusted
Cost Utility Analysis ratio= CER is best value for
Life Years (QALYs)
Costs/QALYs money*
* and those with a CER lower than society’s ‘threshold’ CER are desirable
7. Quality adjusted Life Years
• A measure of outcome which incorporates both
quality and length of life.
• Can capture changes in quality of life, length of life
or both
• Facilitates comparisons between health care
services with very different effects upon health
• Estimating QALYs, changes in QALYs and cost per
QALY gained
8. What are costs and benefits
estimated against?
The ‘counterfactual’ is:
• The position against which costs and
consequences are compared
• The position to which costs and consequences are
incremental (or marginal)
Relevant counterfactuals might include:
• Best practice
• Current practice (the status quo)
• ‘Do nothing’ (e.g. best supportive care)
The importance of the choice of comparator
Table 1. cost effectiveness analysis including Drug B
Cost Cost/QALY
Intervention QALYs ΔQALYs Δcost
(£000) gained
Palliative care 0.4 0.4 1 1 2.5
Drug A 0.6 0.2 2 1 5
Drug B 0.7 0.1 10 8 80
Drug C 0.8 0.1 13 3 30
Table 2: Cost effectiveness analysis excluding Drug B
Cost Cost/QALY
Intervention QALYs ΔQALYs Δcost
(£000) gained
Palliative care 0.4 0.4 1 1 2.5
Drug A 0.6 0.2 2 1 5
Drug C 0.8 0.2 13 11 55
Using Drug B as the comparator means attributing spurious cost
effectiveness to Drug C
9. How is quality of life measured?
Measuring health on a
generic HR-QoL instrument:
the EQ-5D
10. How is quality of life valued?
Example of a ‘tariff’ of social
values (a value set) for the
EQ-5D
11. Methods for eliciting values:
The time trade-off method (TTO)
Offer choice between:
x years of full health
t years at health state i
At the point of indifference, the value of health state i is hi = x/t
VALUE OF HEALTH
Healthy = 1
State i = hi
Dead = 0
x t YEARS
12. The cost effectiveness plane
Difference in effect and cost of an option relative to its comparator
+ cost
IV I
Intervention less
effective and more
costly
- effect + effect
Intervention more
effective and less
costly
III - cost II
13. NICE decision making: the £30,000 question
• What is the basis for NICE’s cost effectiveness
threshold?
• Are NICE recommendations out of keeping with real
budget constraints in the NHS?
• What are the opportunity costs of services displaced by
new technologies?
• Is NICE efficiency increasing?
14. Investment and
disinvestment in the NHS
• The cost effectiveness thresholds in local NHS decision
making (The ‘Williams Project’: Appleby, Devlin, Parkin,
Buxton and Chalkidou, 2007)
• Services ‘at the margin’ identified: investment, disinvestment,
deferred investment
• Exclude: invest-to-save; decisions which are dominant on CE
grounds.
• CUA performed on remainder
• A preview of results
• Implications for value for money judgements.
• What health services are ‘at the margin’ in your
organisation?
15. Outcomes-based management
• BUPA’s experience with using the SF-12 to manage
the performance of clinical staff/teams
• Use of routine health outcome measures in the NHS
(Appleby and Devlin 2004)
