How much Detail is needed in Cost Estimation in an Economic Evaluation alongside a Clinical Trial to Optimise Evidence for Decisions?

Wilson, Edward (2014) How much Detail is needed in Cost Estimation in an Economic Evaluation alongside a Clinical Trial to Optimise Evidence for Decisions? Doctoral thesis, University of East Anglia.

Preview

PDF Download (7MB) | Preview

Abstract

Acquiring evidence to support decision making is expensive. Collecting resource use data
alongside a randomised controlled clinical trial is particularly so due to the multidimensional
nature of costs: different costs are incurred by different agencies with varying
methods and systems to account for these.
Trialists are faced with decisions over how to collect such data, in particular different ‘levels’
of detail are possible. For example, hospitalisations can be costed (1) on a top-down, per
admission basis multiplied by a representative unit cost, (2) a bottom-up basis measuring
every component of care such as nursing and medic time, investigations and other
procedures and drugs used which are each multiplied by relevant unit costs, or (3) some
intermediate level of aggregation. The top-down data will be less expensive to obtain but
may be less accurate (biased and/or over- or under-estimation of uncertainty) compared
with the bottom-up. I refer to these alternative methods as ‘data collection processes’.
Currently such decisions are based on the judgement of the trialist(s). However, formal
quantification of the added value of one data collection process versus another compared
with the added cost would inform the efficient allocation of research resources.
In this thesis I extend the use of value of information analysis to compare the incremental
cost and benefit of one data process with another, further extending this to estimate the
optimal mix of observations between two processes.
Using an example dataset I find that the method is workable, requiring prior information on
the relationship between the two processes which can be obtained from either a pilot or
feasibility study or expert opinion.
When incorporated with other concurrent developments in value of information analysis,
the method has the potential to provide a decision analytic approach to the complete design
of clinical trials.

Item Type:	Thesis (Doctoral)
Faculty \ School:	Faculty of Medicine and Health Sciences > Norwich Medical School
Depositing User:	Users 2259 not found.
Date Deposited:	09 Jul 2014 12:13
Last Modified:	09 Jul 2014 12:13
URI:	https://ueaeprints.uea.ac.uk/id/eprint/49474
DOI:

Downloads

Actions (login required)

View Item