Basic Economic Analysis
David Epstein, Centre for
Health Economics, York
Contents
Introduction
Resource use and costs
Health Benefits
Economic analysis
Conclusions
Introduction
What is economics?
Choices under scarcity
In health care, to allocate available
resources to maximise health benefits
Why conduct an economic evaluation
alongside your clinical trial?
Informdecision making by quantifying
expected health benefits and costs and
the uncertainty around them
Example : RITA-3 trial
Randomised intervention for
treatment of angina
Patients with unstable angina or non-
ST-elevation MI
Routine early angiography with
myocardial revascularisation as
indicated versus a conservative
strategy
Nt = 895 ; Nc = 915 ; 5 years
Contents
Introduction
Resource use and costs
Health Benefits
Economic analysis
Conclusions
Resource use
Costs per patient are volume of resource x
unit cost of each resource
Which resource use? Identify ‘cost drivers’
Angiography, revascularisation procedure, days
in ward, ITU and CCU
Acute cardiac medication during admission
Long term cardiac medication
GP and other primary care
Hospitalisation for other events
What else? Non-cardiac related? Private costs?
days lost from work? (Perspective)
Collecting resource use
Patient specific
Trial case record forms
Patient questionnaires : Postal? Face-to-face
interview?
Hospital notes, GP notes
Administration system records
Resource use diaries
Other
Questionnaire completed by trial coordinator at
each centre
Collecting resource use on a sample of patients
Unit costs
Try and obtain local costs if possible
Hospital administration / finance dept
NHS reference costs (detail available on CD
from Quarry House)
Questionnaire
Expert opinion
National sources
Drugs – BNF and PPA website
Other trial reports, HTA reports and NICE
appraisals (adjust for inflation)
PSSRU website
Manufacturers list prices (rarely disclose
discounts!)
Contents
Introduction
Resource use and costs
Health Benefits
Economic analysis
Conclusions
Health Benefits
Disease-specific measures versus
generic measures versus utility
measures
Disease specific (eg blood pressure)
easier to collect but do not easily relate
to mortality or health-related quality of
life
Generic measures may be measured on
several dimensions eg SF36
Utility measures create a single index
number scaled between full health (1) and
death (0) , and can be worse than death
The EQ-5D
Values of
sample of 3400
members of the
general public
Expressing Health Benefit in QALYs
With programme
Health Related Quality
Without programme
of Life (weights)
0
Death 1 Death 2
Health state duration (yrs)
Contents
Introduction
Resource use and costs
Health Benefits
Economic analysis
Conclusions
Economic analysis
What not to do…
Don’t use cost minimisation analysis
– Costs and health benefits have a joint
distribution, so t-tests of health benefits
alone are not valid
Don’t use average cost-effectiveness ratio
c A - £1500 / QALY
o £6000
B - £2000 / QALY
s
t
s £3000 This only compares A with “do
nothing” and B with “do
nothing”. We want to compare
2 3 QALY A with B
Economic Evaluation Potential Results
Difference in Costs
New intervention New intervention
more costly and less more costly and more
beneficial beneficial
- 0 Difference in Benefits
+
New intervention less New intervention less
costly and less costly and more
beneficial beneficial
-
Economic Analysis
Use incremental cost effectiveness
ratio Societal
Difference in mean costs
Decision Rule < valuation of
Difference in mean benefits health outcome
In previous example ICER = (6000-
3000)/(3-2) = £3000 per QALY
Usually compared with other funded
treatments, benchmark around £20-
£40000 per QALY gained
RITA-3 Results at 4 years
Intervention Conservative
arm n=895 arm n=915
Mortality 60 deaths 80 deaths
Mean HRQol 0.08 0.06
(change)
£7593 £6000
Total costs
2.579 2.500
Total QALYs
Incremental cost effectiveness ratio =
£1593 / 0.079 = 20170
Results are not yet published, therefore illustrative values given instead
Other considerations
Discount health benefits and costs if>1 year
Do sensitivity analyses : test robustness of
conclusions to changes in assumptions made
Is the length of the trial ‘sufficient’ ?
Consider extrapolation.
If follow up time is of different lengths
between patients (censoring) special
analytical techniques are needed
Conclusions
Economics is not about saving money.
It is about trying to do the most good
within available resources.
We all make choices, economic
evaluation makes those choices
explicit.
