Background: Cost-effectiveness analyses worldwide assume that the price of any single drug increases with inflation. New guidance from the Pharmaceutical Management Agency in New Zealand suggests that, when it is known that a generic drug will be available in the near future, a best estimate of the lower price of the generic should be included in the base-case cost-effectiveness analysis. Furthermore, in the sensitivity analysis, the real prices of the new and comparator drugs should be deflated by 2% per year as a proxy for inflation. Objectives: To challenge the widespread assumption that the price of any single drug increases with inflation in the UK, and to calculate the impact on the incremental cost-effectiveness ratio (ICER) of using a more realistic estimate for the future price of individual drugs. Methods: The change in the real price of 373 drugs in the UK over the period 1980-2006 was calculated. Only those drugs launched after 1984 and with more than 500 prescriptions per year were analysed. A linear model of the change in real price by drug was fitted as a function of launch year, number of prescriptions, and British National Formulary (BNF) section. Results: The mean annual decrease in the real price of individual drugs was 3.8% (95% CI 3.4, 4.2), with a standard deviation of 2.5%. Using this value, drugs would generally appear more cost effective than as presently calculated, i.e. the ICER would generally fall. The ICER would fall substantially for drugs for chronic conditions, e.g. by 15%, from 61 pound 900 to 52 pound 700 per QALY (year 2004 values) for cinacalcet for hyperparathyroidism. It is predicted that the ratio would fall even more for longer-term conditions such as multiple sclerosis. Conclusions: Most of the drugs previously appraised by the National Institute for Health and Clinical Excellence (NICE) are actually more cost effective than stated by NICE. Furthermore, most or all drugs for chronic conditions are actually far more cost effective than stated by NICE. Hence, it is likely that some of the previous negative decisions made by NICE concerning drugs for chronic conditions would instead have been positive if the methodology in this study had been implemented. It is recommended that, to capture the true cost of a drug, UK-based cost-effectiveness analyses should assume that the future real cost of a drug decreases over time, typically by 4% per annum, with a standard deviation of 2.5%. This change is very easy to implement in cost-effectiveness analyses. Similar conclusions may apply worldwide.