Poster | 6th Internet World Congress for Biomedical Sciences |
Antonio J. Serrano López(1), Gustavo Camps i Valls(2), EMILIO SORIA OLIVAS(3), NICOLÁS VICTOR JIMÉNEZ TORRÉS(4), José David Martín Guerrero(5)
(1)Dpto. Electrónica. Universidad de Valencia - Burjasot. Spain
(2)Universitat de València - Burjassot, Valencia. Spain
(3)DPTO INGENIERÍA ELECTRONICA. FACULTAD DE FISICAS - BURJASSOT/VALENCIA. Spain
(4)DPTO FARMACIA Y TECNOLOGÍA FARMACEÚTICA. FACULTAD DE FARMACIA - BURJASSOT/VALENCIA. Spain
(5)G.P.D.S. Departament d´Enginyeria Electrònica. Universitat de València - Burjassot. Spain
Contact address: |
Antonio J. Serrano López Dpto. Electrónica Universidad de Valencia Lab. GPDS Burjasot Valencia 46100 Spain antonio.j.serrano@uv.es |
[Health Informatics] |
||||
[New Technology] |
[Cardiolovascular Diseases] |
Decision making systems, play an important role in problems derived of pharmacological treatment where intoxication problems arise. Digoxin is a drug used for cardiac ischemia and auricular fibrilation, and it is usually associated to intoxications due to its narrow therapeutic scope.
Another crucial fact is that the concentration of digoxin in plasma, not only depends in the administred dose, but also on multiple factors, making intoxication prediction a complex problem.
An aid system able to endorse medical decisions based in neural networks (multilayer perceptron) has been developed. This method allows the use of complex relations to model the behaviour of the human body, based in information from real cases.
The data used to develop the Multilayer Perceptron has been acquired from 257 patients monitored by the Servicio de Farmacia of Hospital Universitario Doctor Peset of Valencia. The best result obtained, has been a sensitivity and specifity greater than 80 % for validation.
This neural network has 9 input variables (which include physiological and treatement variables), 7 neurons in the hidden layer and 1 output.
[Health Informatics] |
||||
[New Technology] |
[Cardiolovascular Diseases] |