The problem of variability in computed cost-effectiveness ratios (CERs
) is usually addressed by performing sensitivity analyses to determine
the effects on these ratios of plausible ranges of values of input pa
rameters. However, the sampling variation that exists in these estimat
ed parameters can be utilized to obtain confidence intervals for cost-
effectiveness ratios. As cost-effectiveness analysis becomes more wide
ly used, new techniques need to be developed for establishing when a d
ifference in strategies evaluated is meaningful. A first step is to es
tablish the precision of the CER itself. The authors estimate the prec
ision of a CER in the context of a statistical model in which the prim
ary outcome is survival, with cost and effectiveness defined in terms
of the underlying survival distribution (S). Effectiveness (alpha) is
measured by life expectancy, restricted to a finite time horizon and d
iscounted at a fixed rate r, alpha = integral e(-rt)S(t)dt. Cumulative
cost (beta) per patient is regarded as resource utilization and incur
red randomly over time depending on the survival experience of the pat
ient, beta = integral e(-rt)S(t)dC(t), where C(t) is the total potenti
al resources utilized up to time t. Average cost-effectiveness (ACE) o
f a single strategy is beta/alpha, and when comparing two strategies,
the CER is Delta beta/Delta alpha, the ratio of the incremental cost t
o the difference in mean survival. Utilizing the sampling distribution
of the Kaplan-Meier estimate of S yields standard errors and confiden
ce intervals for ACE and CER. The technique is applied to survival dat
a from 218 previously studied patients to assess 95% confidence interv
als for the CER and ACE of the implantable cardioverter defibrillator
as compared with electrophysiology-guided therapy.