In this study, we describe 'bootstrap' methodology for placing statist
ical confidence limits around an incremental cost effectiveness ratio
(ICER). This approach was applied to a retrospective study of annual c
harges for patients undergoing pharmacotherapy for depression. We used
MarketScan(SM) (service mark) data from 1990 to 1992, which includes
medical and pharmacy claims for a privately insured group of employed
individuals and their families in the US. Our primary effectiveness me
asure was the proportion of patients who remained stable on their init
ial antidepressant medication for at least 6 consecutive months. Our p
rimary cost measure was the total annual charge incurred by patients t
aking the selective serotonin reuptake inhibitor fluoxetine, a tricycl
ic antidepressant or a heterocyclic antidepressant. On average, fluoxe
tine pharmacotherapy tended to decrease annual charges by $US16.48 per
patient for each percentage increase in depressed patients remaining
stable on initial pharmacotherapy for 6 months, resulting in a negativ
e ICER point-estimate. However, the upper ICER confidence limit is pos
itive, which means that fluoxetine treatment may possibly increase ann
ual per patient charges. With 95% confidence, any such increase was no
more than $US130 per patient for each percentage increase in patients
remaining stable on initial pharmacotherapy for at least 6 months. On
e advantage of using a bootstrap approach to ICER analysis is that it
does not require restrictive distributional assumptions about cost and
outcome measures. Bootstrapping also yields a dramatic graphical disp
lay of the variability in cost and effectiveness outcomes that result
when a study is literally 'redone' hundreds of times. This graphic als
o displays the ICER confidence interval as a 'wedge-shaped' region on
the cost-effectiveness plane. In fact, bootstrapping is easier to expl
ain and appreciate than the elaborate calculations and approximations
otherwise involved in ICER estimation. Our discussion addresses key te
chnical questions, such as the role of logarithmic transformation in s
ymmetrising highly skewed cost distributions. We hope that our discuss
ion contributes to a dialogue, leading ultimately to a consensus on an
alysis of ICERs.