The objective of this review was to analyse the methods of paediatric and adult cost-utility analyses (CUAs) conducted in US populations, and to compare the cost-utility ratios of health interventions by classifying them by disease prevention stage, intervention category and primary disease type. A CUA database developed by the Tufts-New England Medical Center, Boston, MA, USA was used to compare the descriptive and methodological characteristics of paediatric and adult studies. The final dataset included 35 paediatric and 491 adult studies, which generated a total of 91 paediatric and 1498 adult cost-utility ratios. In paediatric studies, the most common intervention types were immunisations and pharmaceutical interventions, which each accounted for 17% of studies. Pharmaceutical interventions accounted for the plurality of adult studies (36%). In studies that used a single source of preferences to determine quality-of-life weights, preferences most frequently came from the author in paediatric studies (29%) and the patient in adult studies (14%). Almost all studies with available discount rate data used the same rate (most commonly 3%) for costs and benefits. Few studied used generic health-state classification instruments; preferences for health states were most often based on author and community preferences in paediatric studies, and author and patient preferences in adult studies. The overall median cost-utility ratio was $US7300/QALY (year 2002 values) in child studies and $US26,000/QALY in adult studies; child studies tended to have lower published cost-utility ratios than adult studies, even when categorised according to intervention or disease. In conclusion, CUAs of paediatric and adult health interventions vary across descriptive characteristics, but are largely similar methodologically. Further, cost-utility ratios for interventions evaluated in the literature tend to be lower for paediatric interventions than for adult ones. When allocating resources, policy makers who use economic analysis as a decision-making aid can take some comfort in the methodological similarities between paediatric and adult studies, but more work is required to standardise methods in both groups.