A proposed Mining System for Recognition of Deep Venous Thrombosis and Analyzing the Predisposing Factors

Abstract

     DVT is a very common problem with a very serious complication like pulmonary embolism (PE) which carries a high mortality, and many other chronic and annoying complications ( like chronic DVT, post-phlebitic syndrome, and chronic venous insufficiency) ,and it has many risk factors that affect its course, severity ,and response to treatment. Most of those risk factors are modifiable, and a better understanding of the relationships between them can be beneficial for better assessment for liable patients, prevention of disease, and the effectiveness of our treatment modalities. Male to female ratio was nearly equal, so we didn’t discuss the gender among other risk factors. Data taken from 200 patients with DVT were recognized and analyzed by proposed mining system which contains two stages the first one was NN(back-propagation) trained on patients records with DVT, the second stage was to apply the association rule program to extract the relations among the Predisposing Factors. Immobility was the most important risk factor. Alcoholism and Smoking add more risk to immobile of post operative patient. Age per se has no effect. 100% of patients with long bone fracture were immobile. Fever occurred in one third of post operative patients who develop DVT.           

     The proposed miming system allows better and faster recognition and analysis of more data, which saves time and effort, and discovers the relations among many factors to one or more than one factors. So, we get the above mentioned relations, which are important for the future management of DVT.