Questions of fact and questions of value?
The cost-effectiveness threshold:
what it is, what its not and
how to estimate it?
• When costs only displace health (∆ch)
c
c
h
 h
 c 0
k
v
Health
gained
Health
forgone
v


 v.h  ch  0 
k


Consumption
forgone
• When costs only displace consumption (∆cc)
c
c
h
 h
 c 0
k
v
Karl Claxton
Health
forgone
Department of Economics and Related Studies,
• Costs displace both
c
h
 h
k
Centre for Health Economics,
University of York.
www.york.ac.uk/inst/che
 v.h  cc  0 
Consumption
forgone

cc
v

v

 v.h  ch  cc  0 
k


0
• k = health displaced by increased costs
• v = how much consumption give up for health
• They are not the same thing (k=v is an empirical question)
Why does k matter?
What about the ‘going rate’?
Health
Threshold £30,000 per QALY
Cost
Price = P3
Threshold £20,000 per QALY
£60,000
C3/H1
Threshold £10,000 per QALY
Price = P2 £40,000
£20,000 per QALY
C1/H1
H1
Price = P1 £20,000
v or k ?
1
2
Net Health Benefit
2/3 QALY
What it is and what its not
3
4
Health gained
C2
How does the threshold change?
An efficient HCS
Health
1/k1
Health
1/k1
1/k1
H2
Increased productivity?
H1
1/k11
1/k
Expenditure Increased expenditure?
1/k1
1/k1
B1
Costs with a
‘clinical area’
1/k1
Current HCS
H1
C1
C3
Net Health Benefit
‐2 QALY
B1
Increased prices?
B2
Expenditure 1
What about waste?
Summary
Increase productivity
1/k1
Health
• Need k what ever view of social value
• What its not
Only eliminate waste
1/k1
1/k1
Current NHS
1/k1
H1
– Consumption value of health (v)
– ‘Going rate’ in a clinical area
– Marginal productivity of ideal NHS
• No simple relationship to changes in budget and prices
• Discretionary expenditure
– Most growth on things not easily displaced
– Prices of displaceable activities grown more slowly
– Innovation in technologies, medicine and service delivery
• Heath production outside health care
B1
Cameron’s cash
B2
Budget How can we estimate it?
Relationship between expenditure and outcomes
• Informed judgement of the cost-effectiveness of things the
HCS does and doesn’t do
• Infer a threshold from past decisions
• Find out what gets displaced and estimate its value
• Estimate the relationship between changes in expenditure
andd outcomes
t
• Martin et al (2008, 2009)
– Variations in expenditure and outcomes within programmes
– Reflects what actually happens in the NHS
– Estimates the marginal productivity (on average) across the NHS
Cancer
Circulation
04/05 per LY
per Qol
£13,137
(£19,070)
£7,979
(£11,960)
05/06 per LY
£13,931
£8,426
Respiratory
Gastro‐int
Diabetes
£7,397
£18,999
£26,453
• Need to estimate:
–
–
–
–
–
–
Expenditure and outcomes in the NHS?
Is this only relevant to the NHS?
ΔB, variation in overall expenditure
• Not about administrative budgets
Expenditure equations, elasticity of programme expenditure (%ΔE/%ΔB)
ΔE Programme 1
ICD..
ICD..
ICD..
ΔE Programme 2
ICD..
ICD..
ΔE Programme ..
ICD..
ICD..
ICD..
ICD..
O t
Outcome
equations,
ti
elasticity
l ti it off outcome
t
(%ΔM/%ΔE)
ΔMortality
ICD..
ICD..
ICD..
ΔMortality
ICD..
ICD..
ICD..
ICD..
ICD..
ICD..
R id l
Residual
ΔMortality
ICD..
– Where do the opportunity costs fall?
ΔE Programme 23
ICD..
ICD..
Life years gained
Life years gained
Life years gained
QALYs gained
QALY/LYs loss
QALYs gained
QALY/LYs loss
QALYs gained
QALY/LYs loss
More programmes
How changes in overall expenditure gets allocated across all the programmes
How changes in mortality might translate into QALYs gained
How uncertain any overall estimate will be
How it changes with scale of expenditure change
How it changes over time (panel data)
?
Prior or scenarios
• Any restriction on growth in expenditure
– Opportunity costs on health (k) and consumption (v)
– e.g., Amendments to basic package etc
• Any impact of costs on health
– Restricted benefits withdrawal of insurance
– Health impact of increased co-payments
k
2