| Abstract [eng] |
The dissertation “Survival with random effect and its transformation” has two parts: theoretical and practical. The goal of the theoretical part of the thesis is justification of random effect usage in survival analysis. In the theoretical part, it is proven that survival function (tail of distribution function) with random effect transformation is a new survival function (tail of distribution function). Additionally, it is shown that distribution function with random effect transformation belongs to the same regularity class as distribution function before the transformation. In the practical part, real mortality data of Baltic countries is used to determine which of the mortality models with random effect are most suitable to approximate empirical figures. Another goal of the practical part is to determine the effectiveness of Lithuanian prostate cancer screening program. For this purpose, data from the National Cancer registry and specific mortality model are used. |
