This work attempts to predict the long-term outcome of total hip arthroplasty based on available patient-specific information and possible installation positions of the prosthesis. For this purpose, a holistic modeling approach for the numerical simulation of osseointegration and long-term stability of endoprostheses, including possible prosthesis positions, is developed. In addition, new, efficient, and reliable methods for the numerical description of adaptive bone remodeling and osseointegration are proposed:The adaptive bone remodeling is described as a geometric-linear, material-nonlinear finite element model, following thermodynamically consistent material modeling guidelines. The resulting constitutive equations are expanded to describe osseointegration and transferred into a contact interface between bone and prosthesis. Finally, the results are projected to an imaging format that is easier to interpret for medical professionals, using a newly developed simulation for X-ray images.The inclusion of possible prosthesis positions spans an infinite-dimensional event space. Therefore, the model is reduced to a finite-dimensional surrogate model sampled with an adaptive sparse-grid collocation method.Without clinical validation, reliable statements cannot be made, and therefore the numerical examples given in this thesis can be regarded as proof of correct implementation and feasibility studies. This dissertation thus provides an answer to how much computational effort is required to provide a real digital decision aid in orthopedic surgery.