Background – With the growing costs in healthcare and the growing need for care, prediction models offer great opportunities for improving efficiency and patient outcome. With these models, one can manage the different outcomes as efficient as possible and next to that, understand which variables influence the patient’s pathway. The Erasmus Medical Centre in Rotterdam provided the data of the kidney transplantation patients (donors and acceptors) to examine the effects of different preoperative variables, such as age, body mass index and gender, on the operating time, the surgical complexity, the occurrence of peri- and postoperative complications and the length of stay. Prediction models using this data can give insight into the patients and the processes, it can improve the efficiency, and make it possible to steer the process when necessary. Goal – The goal of this retrospective thesis study is to gain insight into the patient groups and processes and to create a model able to predict the patients pathway through analyzing preoperative patient data. This model should lead to higher efficiency and better patient outcomes and give insight into the variables influencing the patients pathway for both the kidney donors and acceptors. Additionally, this study should provide a practical guideline for prediction model use to achieve the best results in efficiency and patient outcome. Materials and methods - Two sets of clinical histories were used for both the kidney donors and the acceptors: one for deriving the predictive equations and another for validation. For deriving the predictive equations, preoperative patient data was collected from 282 donors and 269 acceptors admitted to the Erasmus Medical Center (EMC) in the period of 2011-2012. For the validation of the equations, patient data was collected from 141 donor patients and 135 acceptor patients admitted in the year 2013. The data includes variables such as gender, age, patient’s body mass index (BMI) and medical history. The variables were selected with the consultation of previous research and experts in the kidney transplantation field. Regression analysis was used to identify the risk factors for each outcome and to estimate the predictive equations used for predicting the patients pathway. Results – For the 80% best predictions, the validation shows that the operating time can be predicted with an average deviation of 11,9% (21,4 minutes) from the actual operating time for the donors and 12,8% (17,0 minutes) for the acceptors. Relevant features for predicting the operating time include gender, BMI, number of veins and arteries, surgeon and surgical technique for the donors and gender, ethnicity and previous transplantations for the acceptors. The length of stay can be predicted with an average deviation of 17,9% (0,6 days) from the actual length of stay for the donors and 16,6% (2,3 days) for the acceptors. The relevant features for predicting the length of stay include gender and pulmonary history for the donors and age of the donor, BMI of the donor and number of months of dialysis for the acceptors. The other outcomes do not show good predictive powers in the regression analysis and in the validation. Conclusions and discussion – Although the predictive equations for operating time and length of stay showed good results for the 80% best predictions, the results for all data were less promising. It is thus recommended to first improve the models before using them in practice. Therefore this study provides a set of adaptations that could improve the model performance. When the models show better results, it is advised to start with a pilot study to find out the benefits of the prediction equations in daily use. Different requirements are created that should be considered when implementing the model for daily use.