Noninvasive approach based on self organizing maps to classify the risk of diabetic foot

Abstract

The purpose of this study was to classify the risk of developing diabetic foot on patients with diabetes mellitus for nursing care. We performed a quantitative research using Self Organizing Maps (SOM), by non-invasive manner, and by means of questions about diabetic self-care, their present health condition and socioeconomic issues. Data analysis was carried out and, then, was developed a SOM which resulted in two different risk groups of developing diabetic foot, and the high risk comprised 40.8% of 250 surveyed participants. Formation of these two groups showed that, from data about care provided by the participants in relation to diabetes and feet, and their present health situation, it is possible identifying the risk for developing diabetic foot and performing the better attendance of those most susceptible to health complications. The risk groups obtained were validated through statistical analyzes of the dominant variables in each group and by using linear discriminant analysis. The study conducted also pointed out the variables that contributed most to the division of risk levels of developing diabetic foot, which helps in enhancing self-care.