|Budget Amount *help
¥3,500,000 (Direct Cost : ¥3,500,000)
Fiscal Year 2004 : ¥1,700,000 (Direct Cost : ¥1,700,000)
Fiscal Year 2003 : ¥1,800,000 (Direct Cost : ¥1,800,000)
In electronic patient record (EPR) systems, progress notes, summaries, operation reports, examination reports are recorded. These records includes significant data for clinical research, thus the precise data extraction method is expected. These records are usually entered in free text format because of its variety. But the data extraction from the records in free text format is difficult. In this study, we developed the method for structured data entry and retrieving of any data from the variety of records.
We developed the data entry module named "Dynamic Template". By using this template, data elements (item and value pair) are arranged in tree structure. These data are presented in XML.
For data analysis, data warehouse (DWH) solution is generally recommended, in which data are transferred from daily active database to the DWH. We had developed DWH for order entry system. In this study, we developed DWH for EPR.
We used relational database for the DWH. One data element is stored in one record. Patient ID, entry date, data recorder, item, path, value, suffix are the main fields. The items and the values selected from the list have their own ID. Each data element is uniquely identified by the path data.
In this DWH, data about blood pressure, heart rate, heart sound are stored by the templates. We searched the case with suspicion of mitral valve or tricuspid valve regurgitation, suspicion of aortic valve regurgitation, severe hypertension, and suspicion of pheochromocytoma from the DWH. We searched 28 cases, 4 cases, 34 cases and 40 cases, respectively by the relatively simple SQL.
By the method mentioned above, we could search precisely the case from the huge variety of records entered in EPR system.