ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology (IJARCET) Volume 2, Issue 1, January 2013 218 All Rights Reserved © 2013 IJARCET Data Mining in Clinical Decision Support Systems for Diagnosis, Prediction and Treatment of Heart Disease Syed Umar Amin, Kavita Agarwal, Dr. Rizwan Beg Abstract – Medical errors are both costly and harmful. Medical errors cause thousands of deaths worldwide each year. A clinical decision support system (CDSS) offers opportunities to reduce medical errors as well as to improve patient safety. One of the most important applications of such systems is in diagnosis and treatment of heart diseases (HD) because statistics have shown that heart disease is one of the leading causes of deaths all over the world. Data mining techniques have been very effective in designing clinical support systems because o f its ability discover hidden patterns and relationships in medical data. This paper compares the performance and working of six CDSS systems which use different data mining techniques for heart disease prediction and diagnosis. This paper also finds out that there is no system to identify treatment options for HD patients. Keywords- data mining; heart disease prediction and diagnosis systems. I. INTRODUCTION Medical errors are both costly and harmful [1]. Medical errors cause tens of thousands of deaths in U.S. hospitals each year, more than from highway accidents, breast cancer, and AIDS combined [2]. Based on a study of 37 million patient records, an average of 195,000 people in the U.S. died due to potentially preventable, in-hospital medical errors [3]. Statistics also show that cardiovascular disease is one of the leading causes of death all over the world [4]. Hence reliable and powerful clinical decision support systems (CDSSs) are required to reduce the time of diagnosis and increase diagnosis accuracy especially for heart disease diagnosis [5]. Clinical decision support systems have evolved from statistical algorithms to complex artificial neural networks. The early decision support systems, also, were based on Bayesian statistical theory [6], probability diagnoses based on essential variables [7]. Syed Umar Amin, Department of Computer Science & Engg, Integral University, Lucknow, India. Kavita Agarwal, Departmentt of Computer Science & Engg, Integral University, Lucknow, India. Dr. Rizwan Beg, Departmentt of Computer Science & Engg, Integral University, Lucknow, India. The use of data mining tools has become widely used in clinical applications for disease diagnosis more effectively. Various data mining techniques such as decision trees, artificial neural networks, Bayesian networks, support vector machines kernel density, bagging algorithm have been actively used in clinical support systems for diagnosis of heart disease [8-10]. Although there have been promising results in applying data mining techniques in heart disease diagnosis and treatment, the study done in finding out treatment options for patients and particularly heart patients is comparatively elemental. It has been suggested by researchers that application of data mining techniques for proposing suitable treatments options for patients would not only improve patient care but would also reduce investigation time, errors and would also improve the performance of medical practitioners [11]. There has been a lot of investigation for applying different data mining techniques in the diagnosis of heart disease to find out the most accurate technique but there is no study to find out the data mining technique which can increase reliability and accuracy in finding out effective treatment for heart disease patients. The remainder of this paper is divided as follows: Various clinical decision systems for heart disease are presented in section 2 followed by comparison and analysis of the presented systems in section 3 and section 4 has conclusion. II. CLINICAL DECISION SUPPORT SYSTEMS FOR HEART DISEASE USING DATA MINING Heart disease refers to various ailments that affect the heart and the blood vessels in the heart. Heart attack Coronary artery disease, heart failure, Angina are some examples, which have different symptoms and causes [12]. The detection of heart disease is a complex procedure because of availability of incomplete data and its dependence on several diverse factors. Therefore, intelligent systems using data mining techniques are required for increasing the accuracy of diagnosis. A large number of clinical