OBJECTIVES: To demonstrate that differential discounting can confuse comparisons of cost-effectiveness between interventions. In particular, to show that directly comparing cost-effectiveness ratios of different interventions from conventional cost-effectiveness models can result in misleading cost-effectiveness rankings under differential discounting.  METHODS: A simple example of a comparison of the cost-effectiveness of two hypothetical interventions is used. The first intervention is a once-off vaccination, which imposes costs in one year alone per patient-cohort. The second intervention is a therapy for a chronic condition, which imposes costs for many consecutive years per patient-cohort. Both interventions are assessed on the conventional basis of a single cohort of patients, resulting in cost-effectiveness estimates for each intervention, which are then compared. An alternative comparison of cost-effectiveness is then made; it considers the opportunity cost of adopting the therapy for one cohort of patients in terms of the vaccination that could be provided for many cohorts over the same period. This alternative comparison yields a second pair of cost-effectiveness ratios which are compared to those from the conventional per cohort comparison. RESULTS: Comparing the interventions on the basis of the actual opportunity cost with the same distribution of spending over time shows the vaccination to be relatively more cost-effective than under the conventional per cohort analysis. Furthermore, the alternative comparison shows the two interventions to have the same relative cost-effectiveness as if the conventional per cohort comparison had been undertaken with discount rates unadjusted for growth in the cost-effectiveness threshold (equal discounting). CONCLUSIONS: The debate over differential discounting has not considered the practical implications for decision making. Our analysis shows that accounting for threshold growth using differential discounting can make comparisons from conventional forms of cost-effectiveness analysis misleading as to which interventions are most cost-effective. Accounting for threshold growth using differential discounting is less relevant for decision making than previously understood.