Αξιολόγηση της αύξησης της συγκέντρωσης ασβεστίου και οστικής αλκαλικής φωσφάτασης στον ορό ασθενών με καρκίνο ως προγνωστικοί παράγοντες για την ύπαρξη οστικών μεταστάσεων
Γκίνη, Ελένη Ι.
Longitudinal data are frequently present in almost all scientific fields where statistical modeling is used. Especially for clinical trails where the effectiveness of a treatment is of concern as well as monitoring and controlling its side effects or symptoms the analysis of repeated measurements is a necessary tool. In this thesis in order to prove the prognostic value of calcium and alkaline phosphatase as biochemical markers for bone metastases in patients with cancer it is analyzed a sample of 37 patients by repeated measurements analysis methods. Initially the problem of bone metastases is defined and a short review of biochemical markers and prognostic tests is presented. Furthermore, basic terms and definitions for repeated measurements are illustrated. Fields of application are mentioned as well as advantages and disadvantages of longitudinal designs. It is provided methods for calculating the appropriate sample size and power. Also it is examined techniques for treating missing and dropout values. Univariate and multivariate methods for analyzing repeated measurements when continuous response variables are of interest and linear mixed models of random or fixed effects are presented. It is provided a brief review of statistical software. Finally, evaluations and conclusions from the statistical analysis of our sample using multivariate analysis MANOVA and linear mixed models are provided.