Non-invasive method to analyse the risk of developing diabetic foot

Abstract

Foot complications (diabetic foot) are among the most serious and costly complications of diabetes mellitus. Amputation of all or part of a lower extremity is usually preceded by a foot ulcer. To prevent diabetic foot, an automatic non-invasive method to identify patients with diabetes who have a high risk of developing diabetic foot is proposed. To design the proposed method, information concerning social scope and self-care of 153 diabetic patients was presented to the K-means clustering algorithm, which divided the data into two groups: high risk and low risk of developing diabetic foot. In the operational stage, the Euclidian distance from the information vector to the centroids of each group of risk is used as criterion for classification. Both real and simulated data were used to evaluate the method in which promising results were achieved with accuracy of 0.97 ± 0.06 for simulated data and 0.68 ± 0.16 considering the classification of specialists as the gold standard for real data. The method requires a simple computational processing and can be useful for basic health units to triage diabetic patients helping the health-care team to reduce the number of cases of diabetic foot.

Keywords: Euclidian distance; K-means clustering algorithm; automatic noninvasive method; computational processing; data classification; diabetes mellitus; diabetic foot; diabetic patients; diseases; foot complications; foot ulcer; health care; information vector; medical diagnostic computing; operational stage; patient diagnosis; pattern classification; pattern clustering; risk analysis; self-care; simulated data; social scope.

Figure 1

Proposed method a Design stage b Operation stage

Figure 2

Silhouette values a Two clusters b Three clusters

Figure 3

Silhouette values for the artificial dataset