Objectives:  In cost-effectiveness analysis, we aim to account for all future costs and benefits for all patients who are currently eligible for a new health technology and who will become eligible in the future.Methods:  We adapt the fundamental concept from epidemiology of the incidence and prevalence of a disease to cost-effectiveness analysis.  We define the prevalent cohort as those patients eligible to switch from the comparator to the new technology at the time the new technology is introduced.  Next, we introduce the concept of multiple future incident cohorts.  The incident cohort starting t years in the future consists of those patients who first become eligible for the new technology t years in the future.  Currently cost-effectiveness analyses worldwide consider only either the prevalent cohort, the incident cohort in only the first year, or a mixture of the two.Results:  On average, patients in the prevalent cohort are older and at a more advanced stage of illness than patients in the incident cohort.  If the cost and benefit discount rates differ, we show mathematically that the cost-effectiveness of all technologies will be substantially affected by our method.  Otherwise, the incremental cost-effectiveness ratio will not change for acute conditions, but may change substantially for chronic conditions, particularly for chronic progressive conditions. Conclusions:  We suggest that analyses capture the costs and benefits arising from the prevalent cohort and all future incident cohorts.  If our method had been used in the past, some health technologies would have appeared substantially more cost-effective, others substantially less cost-effective.  If possible, parameter values (e.g. average age, disease severity) for both the incident and prevalent cohorts should be obtained from the literature.  Otherwise, we describe how such parameters can be estimated.