An Expert System on Detection, Evaluation and Treatment of Hypertension - Artificial Intelligence and Symbolic Computation - page 251

An Expert System on Detection, Evaluation
and Treatment of Hypertension
E. Roanes-Lozano 1 ,E.Lopez-Vidriero Jr. 2 ,L.M.Laita 3 ,
E. Lopez-Vidriero 4 , 5 ,V.Maojo 3 , and E. Roanes-Macıas 1
1 Universidad Complutense de Madrid, Algebra Dept.
c/ Rector Royo Villanova s/n, 28040-Madrid, Spain
{ eroanes,roanes } @mat.ucm.es
2 Hospital Universitario “Nuestra Senora de Valme”
Autovıa Sevilla-Cadiz s/n, 41004-Sevilla, Spain
[e-mail]
3 Universidad Politecnica de Madrid, Artificial Intelligence Dept.
Campus de Montegancedo, Boadilla del Monte, 28660-Madrid, Spain
{ laita,maojo } @fi.upm.es
4 Hospital “Gregorio Maranon”, Hypertension Unit
Doctor Castelo 49, 28009-Madrid, Spain
5 Universidad Complutense de Madrid, Internal Medicine Dept.
Facultad de Medicina, Ciudad Universitaria, 28040-Madrid, Spain
Abstract. This article presents the development of an expert system
on detection, evaluation and treatment of hypertension (high blood pres-
sure), together with an outline of the associated computational processes.
It has been implemented on the computer algebra system Maple .
The starting point is the knowledge about detection of major cardiovas-
cular disease (CVD) risk and about detection, evaluation and treatment
of hypertension, provided in table and algorithm format by experts from
some well known medical societies and committees. As the drug choices
for treating hypertension depends on whether the patient suffers, among
other diseases, from high CVD risk or not, the expert system consists
of two consecutive subsystems. The first one determines the CVD risk
level, meanwhile the second one uses the output of the first one to detect,
evaluate and, if necessary, suggest a treatment of hypertension.
The knowledge expressed by the experts was revised and reorganized by
the authors. Curiously, some errata were found in some of the classifica-
tions provided in table format.
The computational processes involved are simple because the cases are
already separated (disjoint), so they can be translated into IF...THEN...
rules that can be applied using classifications in simple procedures.
Therefore, verification is restricted to considering a few cases (the con-
junction of all diseases together with different CDV risks). Nevertheless,
we think this is a very useful piece of software for practitioners, because,
to reach, for instance, the treatment of hypertension, several messy con-
catenated steps have to be taken.
Keywords: Expert Systems, Hypertension, Computer Algebra Systems,
Medical Informatics.