OBJECTIVES: Great debate exists over appropriate cost/QALY thresholds. We assessed reporting of cost-effectiveness thresholds in the literature to examine prevailing judgements about society's willingness-to-pay per QALY. We hypothesized that stated thresholds were directly related to an author's home country GDP per capita and that they have increased over time to reflect economic growth. METHODS: We evaluated 573 cost-effectiveness analyses (CEAs) published between 1998-2003 and listed in the Tufts-NEMC CEA registry. We scrutinized authors' judgments regarding appropriate cost/QALY thresholds (e.g., that an incremental cost/QALY ratio under $50,000/QALY represents “good value for money”), and examined whether thresholds differed by author's home country GDP per capita (2000 values). RESULTS: CEAs most frequently originated from the US (n=320), UK (n=68), Canada (n=45), and the Netherlands (n=35). Of 573 studies, 325 CEAs (57%) stated an explicit threshold. The most frequent thresholds used in the US were $50,000/QALY (52%), followed by the classification proposed by Laupacis (CAN$20,000- CAN$100,000) (16%), and $100,000/QALY (11%). Most Canadian studies (65%) used the Laupacis classification, and most UK studies (52%) used GBP30,000/QALY. The prevailing US threshold ($50,000/QALY) corresponded to approximately 1.4 times US GDP per capita, whereas the UK threshold (GBP30,000/QALY) was approximately 1.9 times UK GDP per capita. The lower and upper limits of the Laupacis classification used in Canada corresponded to 0.6 to 2.9 times Canadian GDP per capita. Thresholds used by authors have not changed significantly over time and where not inflated to address inflation or economic growth. CONCLUSIONS: Thresholds used in the US, UK, and Canada do not appear to be related to GDP per capita, and have not increased over time. More research is needed to determine appropriate country-specific or disease-specific thresholds. Mechanisms for updating thresholds to reflect inflation rates or changes in GDP are warranted.