OBJECTIVES: Economic evaluation requires reliable estimation techniques for hospital costing in multinational trials. We have developed methods to assign costs to all hospitalizations by multiplying country-specific cost estimates for a 'base' diagnosis by a relative DRG weight (U.S. DRG weight for the diagnosis in question/DRG weight for the 'base' diagnosis). Our objective was to assess the validity of calculating country-specific cost estimates computed with U.S.-based DRG weights. METHODS: Unit costs collected alongside Val-HeFT, a multinational trial in heart failure, were used to compare cost estimates provided by local economists with estimates computed using U.S.-based DRG weights. Unit cost estimates for 8 diagnoses from 14 countries were evaluated. We calculated the correlation between the cost estimates and performed regression analysis to examine the relationship between them. RESULTS: When hospitalization for heart failure was used as the 'base' diagnosis, DRG-based cost estimates were within 35% of the survey estimates for 86% of the countries for unstable angina, 82% for acute MI, stroke, and colon cancer. Costs were less well predicted for pulmonary embolism, coronary stenting, PTCA, and heart transplant, where only 61.5%, 54.5%, 33.3% and 30.0% of predicted costs were within 35% of survey estimates. The DRG-based estimates appeared to overestimate costs for procedure-based diagnoses as the costs were overestimated by 40.3% for stenting, 34.9% for PTCA, and 77.8% for heart transplant. The Pearson correlation coefficient between the two sets of estimates was 0.682. When excluding heart failure, the regression of DRG-derived costs on survey-derived costs indicated that the DRG-based methodology predicted cost estimates fairly well (adj-R2=0.44) (adj-R2=0.571 when adjusting for country). CONCLUSIONS: Estimation of hospital costs based on U.S. DRG weights appears to be a reasonable solution to costing problems in the multinational trial setting. However, care is required in applying this technique in studies with high frequencies of surgical procedures.