A Software Architecture-Based Framework Supporting Suggestion of Medical Surveillance Level from Classification of Electronic Patient Records

Surveillance Level (SL) is an attempt to accompany human development and promote health in primary attention. SL supports patient classification through analysis of his/her risk and protective factors. Considering SLs, health care professionals are able to adopt specific educative, therapeutic or specialized therapeutic actions. This paper proposes a software architecture for manipulation of patient information aiming to suggest her/his SL. The architecture contains a classification layer composed by linguistic and classification software modules. The classification modules implement three different classification methods and the linguistic artifact exploits medical thesauri. The instantiated architecture run experiments in pediatric health care. These experiments showed the feasibility of our proposal and indicated better performance of automatic definition of SL when linguistic processing occurs before classification.