![](bg1.png)
DESIGN OF A MEDICAL DATABASE TRANSFORMATION
ALGORITHM
Karine Abbas
(1)
, Christine Verdier
(2)
, André Flory
(1)
Lyon Research Center for Images and Intelligent Information System
UMR 5205 CNRS/ INSA of Lyon
(1)
, University Claude Bernard Lyon1, University Lumière Lyon 2
(2)
,
Ecole Centrale of Lyon Bat. Nautibus, 43 bld du 11 Novembre 1918, F 69622 Villeurbanne cedex
Keywords: Medical record structure, database reverse engineering, database transformation algorithm.
Abstract: The aim of this article is to create a unique medical record structure from the metabase of any medical
record. The work proposes the design of transformation algorithms which consists in translating the legacy
relational database (RDB) into a unique medical record structure by analysing the correlation between the
legacy RDB keys and the classification of the relations into four types : base relation, dependent relation,
inheritance relation and composite relation.
1 INTRODUCTION
The medical record consists in collecting health
events corresponding to patient’s lifelong. Medical
data are of different types: laboratory test results,
radiology reports, diseases, etc. These data come
from many places: physician practices, hospitals,
nursing facilities, etc. Consequently, for each type
of data and care place, a medical information
system can have its own medical record structure.
Dealing with various record structures requires
much effort and time. Creating a unique medical
record structure becomes a primordial requirement.
In this context, our approach proposes to transform
any medical record towards a unique medical record
structure.
Thus, the aim of our research is the design of a
technique that ensures facility to extract hidden
semantics embedded within database relations and
efficiently to rebuild medical record’s structure.
This paper proposes the design of a medical
database transformation algorithm. This
methodology allows for
- Restructuring the physical schema by using the
data dictionary (métabase).
- Extracting a semantic description in each
relation.
- Analysing of the correlation between primary
keys and foreign keys.
This article is organized as follows. Section 2
presents related works. Section 3 describes the
principle of our reengineering technique by outlining
the database transformation algorithm. Finally,
section 5 concludes with our future work.
2 RELATED WORK
Many database reverse engineering works have been
proposed. (Chiang 1994) proposed an approach
which determines the regular and weak entities,
derives many-to-many and one-to-many
relationships, establishes generalisation hierarchies
and proceeds to the classification of relations,
attributes and their inclusion dependences.
Other works (Anderson 1994), (Petit 1995) based
on query/view analysis have been investigated. These
approaches analyse query language statements.
The method MERCI (Lammari 2001) suggests a
technique in three stages: 1) extract the physical
schema by using the data dictionary 2) design a
conceptual schema in the entity / relation form by
identifying entities, relationships, generalizations,
etc. 3) research hidden semantics embedded within
programs and data.
(Tari 1997) proposes a method which identifies
the relational schema in three categories: base,
dependent and composite relations, and translates
these relations into object-oriented schema by
analyzing the degree of correlation between the keys
of relations and the degree of correlation between the
tuples of the relations.
335
Abbas K., Verdier C. and Flory A. (2006).
DESIGN OF A MEDICAL DATABASE TRANSFORMATION ALGORITHM.
In Proceedings of the Eighth International Conference on Enterprise Information Systems - DISI, pages 335-339
DOI: 10.5220/0002466703350339
Copyright
c
SciTePress