Primary stability of refixated fractures in case of shoulder hemiarthroplasty is a prerequisite to restore physiological glenohumeral joint function. Clinical observations often show a secondary dislocation and subsequent resorption of the bony anchor points like the greater and lesser tuberosity at the rotator cuff tendons. This failed integration leads to impaired glenohumeral load transmission and subsequent reduction of mobility. As a consequence, the optimisation of refixation methods is crucial for a better clinical outcome. To prove the stability of refixation techniques, a Finite Element fracture model was built. Resulting stresses at the bone surface and fragment migration relative to the prosthesis shaft were studied. The results of the calculations show that the isolated tuberosities show unstressed bone regions compared to the intact model. This circumstance may explain the clinically detected bone resorption due to the absence of mechanical stimuli. Furthermore, a cable guidance through lateral holes in the middle part of the proximal prosthesis results in a lower fragment displacement than a circumferential fixation method surrounding the entire proximal bone.