Authors: Elena Martín Pérez, Quintín Martín Martín
Purpose: Study of Patient Information within informed consent, in particular the "Not" category of the "Patient Information" variable.
Methods: This study collects data from hospitals in the University Hospital of Burgos, Spain, for two years, configuring a file with data with 647 cases and 23 variables, 21 of them referred to the attitude towards informed consent, Sex and Age. We will previously carry out a descriptive-exploratory and comparative analysis to have information on the variables that make up the classification/prediction model (Artificial Neural Network), how the data are distributed by category ("Yes" and "Not") of the variable "Patient Information". The study using the three-layer perceptron (input, hidden and output) will be carried out in three phases: Phase I, variables that have two categories; Phase II, variables that have three categories; Phase III, all variables (two and three categories).
Results: Tables 3 show the results of cross-referencing the variable "Patient Information" with the rest of the qualitative variables. The study on the variable "Age", the study of the difference in mean age, generated by the variables that have two categories (Table 4) and three categories (Table 5) leads us to know which difference in means is significant for a level of significance of 5%.
The most efficient artificial neural network structure found in the classification of the categories of the variable "Patient Information" ("Yes" and "Not" categories) is the binomial hidden layer-output layer: hyperbolic tangent- softmax (Dependent variable: "Patient Information"; Partition: Training 60%, Testing 20% and Holdout 20%). Qualifying results are very low for the "Not" category.
Conclusions: The information process, in order to obtain informed consent, has an essentially particular character for each patient, it must be away from any situation of overcrowding, bureaucratization and dehumanization and must be based on their self-determination and freedom.
The study of the variable "Patient Information" using the artificial neural network, perceptron, offers us a low classification/prediction of the "Not" category. One of the factors why the classification of the "Not" category is very low in the variable "Patient Information" is mainly due to the limited data available for this category in the three phases.
View/Download pdf