Intelligent Querying, Visualization and Exploration of Multiple Time-Oriented Patient Records Denis Klimov and Yuval Shahar Medical Informatics Research Center Ben Gurion University of the Negev In a medical world with a large volume of time-stamped information, the clinicians and medical researchers need useful, intuitive intelligent tools to process the multiple time-oriented patient data. Standard means, such as tables, statistical tools, and even more advanced temporal data mining techniques, are often insufficient, can help only in particular cases, or require special experience. To solve the computational aspect of this problem, we have been using the knowledge-based temporal abstraction (KBTA) method for automated derivation of meaningful interpretations and conclusions, called temporal abstractions, from the raw time-oriented patient data, using a domain-specific knowledge-base (KB).The input of the KBTA method includes a set of time-stamped parameters (e.g., platelet counts) and interventions (e.g., bone-marrow transplantation (BMT)); and the output is a set of interval-based, context- specific parameters at the higher level of abstraction (e.g., a period of nearly 3 months of grade II bone- marrow toxicity). Then, these abstractions can be visualized and explored. To analyze clinical trials, or for quality assessment purposes, an aggregated view of a group of patients is more effective than exploration of each individual record separately. In addition, certain patterns can only be discovered through the analysis of multiple patients. Therefore, we have designed and developed a new system called VISITORS (VISualizatIon of Time- Oriented RecordS) which combines the intelligent temporal analysis and information visualization techniques. The VISITORS system includes several tools for intelligent visualization and exploration of raw data and derived abstracted concepts for multiple patient records. We, also, have been developing the ontology-based query language, and the graphical query-construction interfaces which enable the construction of the three types of queries: Select Patients Query, Select Time Intervals Query and Get Patients Data Query, based on the aggregation query-language semantics. These queries retrieve the list of patients, list of relevant time intervals and time-oriented patients' data correspondingly. The query language enables population querying, i.e. the Select Patients Query, using an expressive set of constraints. For example, the typical Select Patients Query is: ”Select all male patients, either younger than 20 or older than 70, whose hemoglobin state was conducted as “moderately low” or lower for at least two days, during the first month after bone-marrow transplantation”. Figure 1 shows the main interface of the VISITORS system. The top panel (A) is used for the patients and time intervals selection tasks. The user can select previously retrieved groups of patients/time intervals from the table, input the necessary patients/time intervals by himself, or construct a new Select Patients/ Select Time Intervals Queries. Figure 1. The main VISITORS interface