【Object】 We have developed some intranet systems for regional medical services with the aim of assisting information sharing and medical cooperation. The systems also support medical education after graduate for general practitioners. We rebuilt these systems which are safe in Internet and have high security to internal attack. 【Methods】 We use a ciphered password which cannot be input by a keyboard as recognition of a user. It is generated in a server computer and delivered to each person with offline media. We compared VPN and SSL for secured data communication and adopted SSL form. The most unique technology of our security system is a real-time ciphered database. Transferred data to the server computer are encoded and stored to the database immediately. 【Results】 We can make three security mechanisms without any other commercial products than Microsoft Windows Server and InterSystems Cache database. A copy of the database cannot be read. Each data will be decoded in real-time for the user who is recognized as the legal user. 【Conclusion】 Almost injustice accesses have been performed by internal staff in Japan. Our security system is very effective also in such situations.

View full abstract

 For the purpose of electronic processing of medical records, we have developed a pen-based interface to realize the flexibility of conventional hand-written patient records in a natural form. This pen-based interface has effective functions not only to support the data entry but also to survey the input data. In fact, however, we had a serious problem because intact hand-written characters input by a pen device are too hard to be utilized for secondary data use. In this paper, we have discovered the possibility of using secondary data from our pen-based interface by using advanced functions of a box-free character recognition and a hand-written query search. With a box-free character recognition function, we can easily transfer hand-written characters into text data whenever the need arises. Moreover, after inputting a series of hand-written data, the text data, which can be used on a secondary level, is also made possible by the process of conversion in bulk. Additionally, we can find the targeting characters in the whole of the hand-written characters with the function of the hand-written query search, even if it does not make a specific database structure.

View full abstract

 The purpose of this study is construction of knowledge discovery method from medical documents with free format. As the first step of this study, this paper discussed the keyword map generation method using Self Organizing Map(SOM) from incident report with free format to discover new trends or features. In particular, this paper proposed (1) the keyword extraction method using morphological analysis tools and heuristics rules from incident reports and (2) the coding method for SOM using the extracted keywords. In the proposed method, the some items in the incident reports (e.g. occurred time, category of occurrence, medical staff’s name etc.) were not employed to generate the keyword map but only free format sentences were did. By the proposed methods, this paper showed that appropriate keywords in the incident reports were extracted, and the effective keyword map was automatically generated. An example of generated keyword map showed some connections of keywords visually and gave the possibility for knowledge discovery(i.e. trends or features underlaid in reports and difficult to find from each document) using the proposed method. Moreover, this paper described the problems and the future works of the proposed method, and discussed the possibility of the proposed method to develop the risk management support system.

View full abstract

 Clinical data analysis requires well-designed clinical data retrieval systems that help users to efficiently obtain desired data without considering the details of complex database structures. In this paper, specialized data-retrieval methods for clinical data analysis are proposed: a condition-setting method suited for dividing retrieved data into two groups, and a method of selecting data items for user data analysis based on data mining. The condition-setting method provides three types of search conditions: precondition, classifier condition, and search items. For both types of condition, complex data-retrieval statements can be expressed by using logical operators to join multiple conditions. The method of selecting data items shows association rules which have high odds ratio as candidate at ributes by using association rule mining.
 Both methods have been implemented in an inpatient case database for cardiac disease. The good usability of the methods has been demonstrated.

View full abstract

 The percentage of prescription received and dispensed in community pharmacies increases every year so that the number of community pharmacies also tends to increase. The number of pharmacy’s stocks are becoming greater as the prescription of generic drugs gradually prevails and it is an important task to reduce the number of pharmacy’s stocks, especially pharmaceutical dead stocks. At present, a practical system for supporting transfer of drugs among different pharmacies has not been developed, in spite of the increase in the number of community pharmacies. Since it is not possible for a pharmacy to store all drug supplies, a practical and minimal system that is superior to conventional stocks-management systems in its cost-performance and data-sharing capability is desired for a multi-store deploying pharmacy chain. Therefore, in this study, we attempted to develop a practical, community pharmacy-oriented system by which we can promptly search all drug listings and dead stocks in every pharmacy and optimize the delivery schedule of drugs between different pharmacies. Utility of our system will make it possible to supply drugs very quickly and efficiently, thus may help pharmacists in community pharmacies contribute better to patients’ care.

View full abstract

 This paper shows several technical issues of applying ATSQL 2-based temporal databases to medical information processing. First, taking ATSQL 2 as an example, we give a brief explanation about temporal databases and their query languages in a general way. Second, using a currently available temporal database mediator called TimeDB, an implementation of ATSQL 2, we exemplify simple queries, then compare the expressive power of time-dealing operators of SQL-92 and ATSQL 2. We describe the usefulness of temporal databases and argue over 1) three semantics of ATSQL 2, 2) hidden columns of temporal tables, and 3) TimeDB. Using ATSQL 2 under mixed environment with SQL-92, we concluded that ATSQL 2 has critical problems to be solved, especially in the semantics and implementation.

View full abstract