Semantic models in medical record data-bases.

Abstract

A great effort has been recently made in the area of data-base design in a number of application fields (banking, insurance, travel, etc.). Yet, it is the current experience of computer scientists in the medical field that medical record information-processing requires less rigid and more complete definition of data-base specifications for a much more heterogeneous set of data, for different users who have different aims. Hence, it is important to state that the data-base in the medical field ought to be a model of the environment for which it was created, rather than just a collection of data. New more powerful and more flexible data-base models are being now designed, particularly in the USA, where the current trend in medicine is to implement, in the same structure, the connection among more different and specific users and the data-base (for administrative aims, medical care control, treatments, statistical and epidemiological results, etc.). In such a way the single users are able to talk with the data-base without interfering with one another. The present paper outlines that this multi-purpose flexibility can be achieved by improving mainly the capabilities of the data-base model. This concept allows the creation of procedures of semantic integrity control which will certainly have in the future a dramatic impact on important management features, starting from data-quality checking and non-physiological state detections, as far as more medical-oriented procedures like drug interactions, record surveillance and medical care review. That is especially true when a large amount of data are to be processed and the classical hierarchical and network data models are no longer sufficient for developing satisfactory and reliable automatic procedures. In this regard, particular emphasis will be dedicated to the relational model and, at the highest level, to the same semantic data model.