OBJECTIVES: Effective and robust health economic models are built on a foundation of reliable, accurate data sets.  With the availability of various real-world data sets, the objective of this study was to examine several real-world data sets and rate their overall utility for use in health economic models in the United States. METHODS: Real-world data sets included in the assessment were one claims dataset, one electronic medical record (EMR) dataset, 11 state-level all payer claims databases (APCDs) and data from the Healthcare Cost and Utilization Project (HCUP).  Factors assessed included:  coverage of national versus regional population, inclusion of various sites of care, free/public access, and availability of longitudinal patient-level data on utilization and outcomes.  Each factor was rated by three independent reviewers on a scale from 1 to 5, with 1 being limited use in economic modeling and 5 being highly useful for health economic modeling. Ratings were summed for each reviewer and averaged to produce a score out of 20 possible points. RESULTS: Of the 14 real-world data sets assessed for utility in health economic modeling, claims data rated highest (15 out of 20) due to the availability of patient-level data from multiple sites of care, but lacked a link between utilization and patient outcomes.  Despite the availability of patient outcomes associated with utilization with EMR data, the data set was rated the lowest (8 out of 20) due to its lack of nationally representative data and proprietary access. CONCLUSIONS: Of the currently available real-world data sets, claims data was viewed as most useful for health economic models because they include patient-level utilization data from multiple sites of care.   Data mined from EMR represents a significant opportunity to link healthcare utilization with patient outcomes, but currently available EMR data lacks national representation and is expensive to obtain.