It's a good question and one that causes no end of problems.
My view is that when one is setting up an economic model, there are
actually very few things that are gained from drug trials, particularly if
one is trying to simulate the 'real world'. Basically, it comes down to
efficacy and side effects. on the efficacy side, one tries to get
the 'best evidence' - often this may be from ITT - not ideal obviously but
generally at least, the results should be on the conservative side. on the
side effect side, again, providing that non-specific questioning is used,
one should elicit more adverse events than would normally occur - the
question here is, which ones would lead to a change in management and
hence modify the cost.
This will always be an issue in trying to 'piggyback' an econmic study on
to a clinical trial - the study is done according to a protocol which may
not be that used in normal clincial practice, and on the other hand,
economic models are jsut that, simplified reflections of real-life.
So, in answer to the question, I would use the best available evidence but
recognise the problem!
regards
Richard
The Goffin Consultancy Ltd
Riding House,
Bossingham Road
Stelling Minnis
Canterbury, Kent CT4 6AZ
United Kingdom
Tel: +44 (0)1227 709220
Fax: +44 (0)1227 709721
Email: [log in to unmask]
Web: www.goffin-network.co.uk
Member of the PSC Group (http://psc-g.tripod.com/)
|