Μελέτη της εφαρμογής της τηλεϊατρικής στην Ελλάδα με χρήση στοχαστικών και αριθμητικών αναλυτικών μεθόδων και υπολογιστικών αλγορίθμων

Abstract

Discussing introduction of telemedicine should consider two key aspects of the project: the feasibility of implementing in technology terms and the feasibility of the application in economic terms. Regarding the second aspect of economic and technical study, the majority of these solutions have been econometric studies of implementation. Inevitably, the introduction of a telemedicine system will have consequences on the functioning and efficiency of hospital units. The studies focused on implementation and effectiveness study of hospital units before and after the introduction of a telemedicine system. The role of Data Envelopment Analysis (DEA), in these methods of efficiency evaluation, is dominant. This method is, by construction, a deterministic method that extracts point estimates of relative efficiency. This PhD thesis aims to make the study of stochastic DEA, studying and extending the stochastic nature of the method and extending results in the field of probability and statistics by using analytical methods (characteristic functions of stochastic variables) or, where it is impractical to thoughtfully approximate, stochastic or numerical algorithms. One of the tools in this study is Bootstrap techniques. These techniques are helpful for studying the use of internal and environmental variables. Monte Carlo techniques are used to examine the stochastic nature of efficiency when it is feasible to make assumptions about inputs and outputs and to test the performance of algorithms and the stability of solutions. Multivariate techniques (Factor, PCA) are used in DEA to increase the accuracy of estimation by increasing the information used and reducing dimensionality. The EM algorithm is used to estimate parameters that optimize the likelihood of the results of DEA. All above are implemented in code language R. The stochastic methods and algorithms developed are applied to study the efficiency of hospital units before and after the implementation of a proposed telemedicine system for the Greek area. The statistical rather than deterministic nature of exported results gives more opportunities to carry out conclusions and decisions on the application of telemedicine and measure the efficiency of hospital units and production units in general.