Sistema de apoio a decisão medica em valvulopatias

AUTOR(ES)
DATA DE PUBLICAÇÃO

2000

RESUMO

One of the most complex fields of medical decision making for a medical doctor is the diagnosis, as it depends on the analysis of information of different complexities and characteristics, and it can be influenced by the physician s previous experience and intuition. It is difficult to formalize this knowledge and represent it in a computational software, because the mental mechanisms and the reasoning process used by the physicians to reach a diagnosis are still not well known. There is no standardization of medical terms and definitions mainly due to the fact that, frequently, there is no consensual opinion of the experts on an area of study, there is difficulties in deciding when facing conflicting evidence from many different scientific journals, books, issues or just from different minds. In this thesis, a new computer program for medical decision support was developed and presented. The software runs on a personal compute r and, by means of the answers to questions made to the patient by his physician and by the results of physical examination and critical exams like X-rays and ECG, it is possible to confirm a clinical expectation concerning heart diseases (cardiac valve insufficiency or stenosis) The developed program utilizes fuzzy sets to deal with the intrinsic imprecision of this processo. It has also been developed a computer program that serves as an interface between the compute r and the medical user using Visual Basic and C++ languages. In the interface with the user it is introduced ali the medical aspects needed to generate and visualize the results. After computer implementation, the program has been tested with the data from 178 patient records with confirmed diagnosis of different cardiac valve diseases. The main correct results achieved by the program are: 67% of totally correct diagnosis; 76% for type of valve; 81 % type of disease (cardiac valve insufficiency or stenosis); 85% of severity of the disease (light, moderate, or severe). In summary, the program has achieved 95% of correct valvular disease (type of valve, disease and severity of the disease) and 5% of totally incorrect diagnosis, making it an important toei for the medical decision aid

ASSUNTO(S)

computadores - doenças - diagnostico medicina - processo decisorio inteligencia artificial conjuntos difusos analise por conglomerados coração - valvulas - doenças

ACESSO AO ARTIGO

Documentos Relacionados