Bj. Obrien et Mf. Drummond, STATISTICAL VERSUS QUANTITATIVE SIGNIFICANCE IN THE SOCIOECONOMIC EVALUATION OF MEDICINES, PharmacoEconomics, 5(5), 1994, pp. 389-398
This article examines the twin concepts of the statistical significanc
e and quantitative importance of observed differences in studies compa
ring medicines in terms of economic parameters such as cost-effectiven
ess and measures of health-related quality of life (HRQOL). Central to
the design and interpretation of any comparative study, such as a ran
domised controlled trial, is some prior judgement about the order of m
agnitude of a difference that would make one switch from one therapy t
o another. Starting with current definitions of clinically important d
ifferences we argue by analogy that the importance of differences in H
RQOL require a shift of focus from the physician to the patient for pr
eferences and judgements concerning what is important to them. Whether
an intervention offers sufficient value for money (cost effectiveness
or cost utility) to warrant resources being reallocated to it is a co
llective decision requiring the input of public preferences about the
relative importance of alternative therapies and health benefits. Ulti
mately, the importance of the health benefits offered by a new drug is
revealed by societal willingness to-pay. This may be stated implicitl
y through committees using cost-effectiveness 'league tables' for deci
sion making, or explicitly by consumer surveys of willingness-to-pay i
n the context of cost-benefit analysis and stemming from the theoretic
al foundation of welfare economics.