In studies of technologies with a potential for life-long consequences, the time horizon of cost-effectiveness analysis would ideally be over the remaining lifetime. However, the relevant time horizon will also depend on the availability of appropriate data, as well as on the availability of appropriate statistical or mathematical models, in order to estimate costs and outcomes. In the cost-effectiveness studies located, the majority use a time period over which most of the significant economic and health consequences of the intervention could become apparent. In order to do this, they invariably link intermediate endpoints to final endpoints and/or extrapolate evidence on effectiveness. One approach when undertaking an economic evaluation with a time horizon of the remaining lifetime of the modelled cohort, is that costs and effects can be measured alongside a trial and then, when appropriate, be extrapolated using modelling (e.g., by synthesizing trial-based information with information on costs and effects from other studies).
Although some of these models may not describe reality with full face validity, they may be able to offer a useful method of making relative cost-effectiveness comparisons between different screening modalities.
The main differences and similarities between the studies with respect to the stated time horizons and discount rate(s) can be found from the Appendix (see table “Study designs and results”).