Transcript Tony O'Hagan

Design of costeffectiveness studies
Some introductory remarks by
Tony O’Hagan
Welcome!
• Welcome to the fifth CHEBS dissemination
workshop
• This forms part of our Focus Fortnight on
“Design and sample size for cost-effectiveness
studies”
• Our format allows plenty of time for discussion
of the issues raised in each talk, so please feel
free to join in!
Trial design
• Design of clinical trials and other observational
studies is an area of research with many facets
› comparators, blinding, sequential/interim, outcome
measures, sample size, …
• One question that is routinely referred for
statistical evaluation is that of sample size
• This is itself a large and complex research topic
› depends on other design facets, distributions of
outcomes, objectives, …
C-E trials
• The increasing importance of economic
evaluation has led to an emphasis on gathering
economic information alongside a conventional
efficacy trial
• Sample size calculations have been devised
› Only assuming joint normality of costs and efficacy
• But sample size is in practice rarely determined
with economic evaluation as the primary
objective
C-E trials (2)
• Despite the growing number of such trials, they
will rarely provide all the evidence needed for an
economic evaluation
› The more realistic context is that they provide
evidence on some parameters in an economic model
› Given that other information must be sought from
outside the trial anyway, why bother to get economic
data in the trial?
Challenges
• Sample size calculations with more realistic
modelling of
› data
› decision problem
• Accepting that sample size will actually be
chosen on other criteria
• Quantifying the extra value of
› resource use data
› more pragmatic trials