0:00
Hello, I'm Tony O'Hagan.
Welcome to the fifth and
last talk in my series,
Bayesian Methods in
Health Economics.
This one is entitled
Formulating Input Uncertainty.
0:14
Let's just quickly review one more
time the outline of the whole course.
The first two talks were about
Bayesian methods in general,
and then the final three have been about the use
of Bayesian methods in health economic evaluation.
In Part 3, we introduced the ideas
of health economic evaluation
and issues around uncertainty.
We talked a bit about
measuring uncertainty
when the analysis is of cost-effectiveness
data from clinical trials
but pointed out that the main way
that health economists do evaluations
is through economic modelling.
The fourth talk introduced
economic modelling,
and talked about the principles of
probabilistic sensitivity analysis
and dealt with a large
part of how you do that,
and this final talk
concerns one more issue
about probabilistic
sensitivity analysis (PSA),
which is how do we express the uncertainty
on the inputs to the economic model?
1:12
This talk is
organised as follows:
we begin by looking
at the general ideas
about expressing
uncertainty on inputs,
and look again at the vaccine
example from the previous talk.
We then move on to three more technical
issues about uncertainty and inputs.
The first is data gaps,
which is where the
information we have
doesn't relate quite to
the parameter that we want
to express uncertainty
about for our model.
Then we'll go into correlation,
where two or more inputs
are related to each other,
and finally, we'll talk about how we
express uncertainty about utilities.