| Recent discoveries in molecular biology, pharmacology and medicine have resulted in a growing need for human tissue samples for biomedical research. However, it is also becoming clear that tissue samples must be associated with a large amount of clinical and molecular information. The annotation of tissue samples with clinical information has therefore become a major challenge for the biomedical community.; Characteristics of users, queries and data elements in relation to a tissue centric data warehouse design were investigated in this study. Sixty-two biomedical researchers were recruited from the University of Pittsburgh Cancer Institute (University of Pittsburgh Medical Center) and the University of Pittsburgh. Subjects were members of a tissue centric research group who require complex sets of health information. Relationships between users, queries, and the data elements required were also examined. Methodologies used include a self-reporting questionnaire, a follow-on interview, publication analysis, and dimensional modeling. Hierarchical clustering methods, publication analysis, and query analysis techniques were used to collect characteristics of users, queries, and data elements in relation to dimensional modeling for a tissue centric data warehouse design.; The study data indicated that tissue centric researchers are individuals with complex and diverse requirements for medical information. However cluster analysis identified distinct, statistically significant subgroups (clusters) on the basis of tissue use and research requirements. Significantly, the combination of query analysis and dimensional modeling (the creation of logical data warehouses based on research's information needs as defined by query analysis) was very effective in defining a core set of information needs across a wide range of diverse information users. Furthermore, it was able to identify two semantic representations of the users' information needs as well as two universally important dimensions, Time and Diseases, for data aggregation in a warehouse design.; Results from this study will provide a better understanding of a complex discipline and user information behavior in a wide range of institutions beyond its current application in a cancer center. Furthermore, a core set of tissue centric information requirements (or tissue annotation) exists that can form the basis for a tissue centric data warehouse, which can effectively serve biomedical researchers. |
