Design and implementation of integrated clinical record systems : a multidisciplinary approach.
In this work new approaches to the design and implementation
of clinical record systems are examined. Although information
technology has long been successfully used in specific areas of
medicine, very few situations exist where information systems
are routinely used to support the medical record. The underlying
thesis of this work is that the major reasons for this failure
are the complexity and vastness of the medical field and the
limitations of traditional methodologies and models for information
systems development. I contend that there is need of an
interdisciplinary basis for information systems development
methodologies, which account for the multiple characteristics
of medical care and for the related information systems.
The research has been done from the perspective of a real
hospital where the present goal of computerisation has seen the
introduction of information systems in routine clinical practice.
First, the problem of developing information systems for
clinical laboratories is addressed, and a proposed, entity-based
methodology developed and implemented. Then, a different entitybased
approach is devised for the area of clinical records. This
has been successfuly implemented in several clinical applications.
However, due to limitations of this approach the natural
language paradigm was selected as a basis for a different methodology.
A multi-functional information model and system is devised,
where information is represented and manipulated by means
of different models and representations.
These models correspond to three semiotic functions which
clinical record systems should support. First, there is the Atomic Object Model which manipulates 'atomic' predications.
This model is used primarily for the recording of simple facts
(both knowledge and data). Second, there is the Medical Record
Model which encompasses mostly structural and temporal properties
of information and its major semiotic function is communication.
It utilises abstraction principles such as 'generalisation'
and 'aggregation'. The third model, the Clinical Model,
is designed to incorporate different roles'that information can
play in reasoning for clinical problem-solving. An information
system was developed in which special care was given to problems
of man-machine interaction, both in regard to information modelling
and to manipulation of patient information.
An integrated information system was developed gradually
using different database management systems. A dozen different
clinical applications have been developed and implemented and
hundreds of physicians and nurses utilise the system in routine