A Decision Tree Analysis of Diabetic Foot Amputation Risk in Indian Patients.
A Decision Tree Analysis of Diabetic Foot Amputation Risk in Indian Patients.
复制标题
DOI：
10.3389/fendo.2017.00025
复制
发表时间：
2017
影响因子：
5.2
通讯作者：
Jadhav SP
中科院分区：
文献类型：
--
作者：
Kasbekar PU;Goel P;Jadhav SP
本研究的目的是创建一个循证工具，指导糖尿病足患者截肢的风险。对301例糖尿病足患者的住院记录进行回顾性分析，以寻找足截肢决定的解释变量。该研究包括所有患有下肢溃疡且已知有糖尿病史的患者或入院后诊断的患者。对数据集进行分析，并使用决策树算法C5.0构建风险评分系统。两个分类器，一个简单的和另一个复杂的，被构造用于预测截肢结果。根据我们的评估，决定截肢的最有影响力的预测因素是多普勒血流测量和溃疡的瓦格纳分级。简单分类器仅使用这两个参数来确定风险。所获得的结果表明，在初级组中的准确度为96.4%，在测试组中的准确度为94%。第二个分类器是一个更复杂的计算机衍生结构，在主组中显示出100%的准确性，在测试期间显示出96%的准确性。在当今的精准医学时代，这两个分类器可以作为糖尿病足患者肢体预后的准确指南，并可以预测未来截肢的风险。
The aim of this study is to create an evidence-based tool that guides the risk of amputation in diabetic foot patients. Hospital records of 301 diabetic foot patients were examined retrospectively for explanatory variables of foot amputation decisions. The study included all patients with a lower limb ulcer with a known history of diabetes mellitus or those diagnosed post-admission. The dataset was analyzed, and a risk scoring system was constructed using the decision tree algorithm, C5.0. Two classifiers, one simple and another complex, were constructed for predicting amputation outcome. Based on our evaluation, the most influential predictors for a decision to amputate are Doppler flow measurements and the Wagner grading of the ulceration. The simple classifier uses just these two parameters in determining risk. The results obtained show an accuracy of 96.4% in the primary group and an accuracy of 94% in the test group. The second classifier is a more complex computer-derived construct that showed 100% accuracy in the principle group and an accuracy of 96% during testing. In the present era of precision medicine, these two classifiers act as an accurate guide to the prognosis of the limb in patients with diabetic foot and can predict the risk of future amputation.