Biomedical translational research can be facilitated by integrating clinical and research data. In particular, study cohort identification and hypothesis generation is enabled by the mining of integrated clinical observations and research resources. The "informatics for integrating biology and the bedside, " or i2b2, framework is widely used for this biomedical data mining. The i2b2 "star schema" data model using entity-attribute-value (EA V) formatted concepts is a very efficient strategy for querying large amounts of data. However, until the most recent i2b2 release, the utility of the platform was somewhat constrained by the limitations on being able to express "facts about facts" - i.e., modify the observations about the patients. We have found that exploiting the new modifier functionality has significantly and favorably impacted the design of i2b2 ontologies, leading to easier and more meaningful query results.