The use of factor analysis approach in the medical data: An application for the field of cardiology

The Use of Factor Analysis Approach in the Medical Data: An Application for the Field of Cardiology Factor analysis is a multivariate method of analysis used to reduce the number of data to a smaller number of dimensions in order to facilitate understanding and interpretation of a large number of variables that are considered to be highly correlated. Factor analysis method is preferred by contemporary researchers especially in the social sciences when it comes to scale development, grouping and studying the features that cause a situation. In studies conducted in the field of health, the aim is usually to show the relationship between a result and several causes. For this reason, it is possible to make this multiple relation more easily understandable by using dimension reduction methods. In this thesis study, it is aimed to show an alternative application of factor analysis approach in medical data obtained from health sciences. For this purpose, the markers used to determine hospital mortality in patients with cardiac insufficiency were examined. Using factors generated by factor analysis, it has been shown that the ability to identify mortality (death) of these markers can be increased. The data used in the thesis consist of 196 patients and 567 treatment patients admitted to the cardiology services of Adana Numune Training and Research Hospital Seyhan Application Center between 2014 and 2016 with cardiac insufficiency diagnosed and died during hospitalization. Of these patients, 35 markers were recorded that had an impact on hospital deaths. The successes of measuring the new factors obtained by factor analysis using markers have been measured. It has been found that the success of some factor constructs to determine mortality is higher than the individual success of the markers. In the different factor analyzes, the LDH and AST variables and also QRS interval, QT-interval value and EF values are usually together at the same factors. Since in these factors, the classification is higher, it is concluded that the use of these factors would be more appropriate to determine mortality instead of individual use. In this thesis, an alternative application of factor analysis approach in the field of health have been investigated. It has been shown that by using the factors generated by this method, the classification performance of some markers can be increased. With this thesis, it is shown that the multicollinearity problem in medical data can be partially reduced by the use of the factor analysis approach and this method can increase the amount of information obtained as a result of analysis. Key words: Factor analysis, heart failure, mortality, area under ROC curve