A multi-level approach for data description and management of a large hierarchical database supporting a hospital patient information system

Abstract

The paper describes a systematic approach to satisfy the various — and partly conflicting — data processing requirements occurring in a large operational hospital database. The central patient data bank of the Medical System Hannover contains at the present time the data of about 124000 patients with a large number of medical and administrative data for accounting and reporting. This data is stored in several databases under the hierarchical database management system IMS and in standard OS-files. A number of parametric update and retrieval programs is in operation on this data bank and the main design criteria for the data structures had to be the support of the efficiency of these routine programs in addition to database administrator considerations such as storage space economy, security and stability. In the recent years, the need for more flexible tools to describe and process this large amount of data has become apparent. Therefore, a data description system handling the various views of data has been designed and implemented in a first version. The choice of the appropriate type of data processing language for a specific task depends on a number of criteria, such as efficiency, input/output requirements and the complexity of the database problem. Therefore, the following hierarchy of user-languages had to be developed: (1) The conventional PL/1-programming using the DBMS sublanguage DL/1 (2) A procedural database language (3) A descriptive query language. The main features of the languages (2) and (3) are described in the paper.