Predictive modeling of multistep clinical pathways: application to infertility treatment process

The predictive model was made from data that included complete information about the infertility treatment process. The data included the diagnostic stage, the laboratory stage, and all subsequent stages, such as embryo transfer or cryopreservation. The research focuses on the stages of the treatment process and considers external aspects and medical decisions. The influence of a particular physician on treatment outcome is taken into account through this approach. To achieve this goal, the feature set was divided into groups characterizing the stages of the treatment process. These stages include the initial patient consultation, the results of the patient examination, the results of materials from the laboratory, and the results of treatment. The groups of features from each stage were sequentially added to the model, and the experiment identified the most significant features from each group. In addition, the work focuses on the study of the doctor's role in the treatment process.