DESIGN AND IMPLEMENTATION OF AN ACTIVE DATABASE FOR MANAGING HYPERTENSION

CHAPTER ONE
INTRODUCTION
An increasing number of institutions are approaching the problem of distributing and improving the use of guidelines in patient care. Electronic medical record systems have the potential of making a great impact in the increased utilization of guidelines since these systems are often used by clinicians at the time when care is delivered. In some cases guidelines are put into the systems in narrative form; in others, several independent logic modules react when actions prescribed by guidelines do not take place. Systems integrating both data-maintenance and decision support functions have requirements that span from efficient data-manipulation, typical of databases, to inferential elaboration, typical of expert systems. Active databases8 may be considered as a bridge between these two kinds of systems: they have both the efficiency of databases in handling data storage and retrieval and the power of a complete rule language in expressing complex inference mechanisms. Moreover, in active databases the knowledge is collected in schema and rules enforced by the database manager. The same knowledge is thus available and transparent to all applications accessing the database. This property, known as knowledge independence, allows knowledge to be easily maintainable. The inferential complexity of the guidelines for the management of hypertension and the necessity of archiving the huge amount of data acquired in periodic follow-up visits make active databases particularly interesting when developing decision support systems for hypertension guideline compliance. HyperCare is a prototype of a decision support system for compliance with essential hypertension management guidelines,1,2 implemented using the active database language Chimera. The choice of this language is justified by the rich object-oriented data definition language and the expressive syntax for active rule definition offered by Chimera. Even if the system that we propose is not the first support system for hypertension management, as far as we know this is the first application of active databases to medical guideline compliance.