Rapid development of technology and data analysis techniques, such as data mining and machine learning, provide opportunities in the field of telemedicine. One of the difficulties of application of data mining techniques in telemedicine information systems is impossibility to select a single data mining algorithm for every problem and for every data set. In this thesis, a method for implementation of medical data analysis with respect to telemedicine system goals and data properties is described and validated. The approach provides system flexibility, that is, ability to add new data sets and data analysis functionality. Along with the method, infrastructure and a prototype of a telemedicine system with the use of data mining techniques was developed.