There is a large variability associated with hip stem designs, patient anatomy, bone mechanical property, surgical procedure, loading, etc. Designers and orthopaedists aim at improving the performance of hip stems and reducing their sensitivity to this variability. This study focuses on the primary stability of a cementless short stem across the spectrum of patient morphology using a total of 109 femoral reconstructions, based on segmentation of patient CT scan data. A statistical approach is proposed for assessing the variability in bone shape and density [Blanc, 2012]. For each gender, a thousand new femur geometries were generated using a subset of principal components required to capture 95% of the variance in both female and male training datasets [Bah, 2013]. A computational tool (Figure 1) is then developed that automatically selects and positions the most suitable implant (distal diameter 6–17 mm, low and high offset, 126° and 133° CCD angle) to best match each CT-based 3D femur model (75 males and 34 females), following detailed measurements of key anatomical parameters. Finite Element contact models of reconstructed hips, subjected to physiologically-based boundary constraints and peak loads of walking mode [Speirs, 2007] were simulated using a coefficient of fricition of 0.4 and an interference-fit of 50μm [Abdul-Kadir, 2008]. Results showed that the maximum and average implant micromotions across the subpopulation were 100±7μm and 7±5μm with ranges [15μm, 350μm] and [1μm, 25μm], respectively. The computed percentage of implant area with micromotions greater than reported critical values of 50μm, 100μm and 150μm never exceeded 14%, 8% and 7%, respectively. To explore the possible correlations between anatomy and implant performance, response surface models for micromotion metrics were constructed using the so-called Kriging regression methodology, based on Gaussian processes. A clear nonlinear decreasing trend was revealed between implant average micromotion and the metaphyseal canal flare indexes (MCFI) measured in the medial-lateral (ML), anterio-posterior (AP) and femoral neck-oriented directions but also the average bone density in each Gruen zone. In contrast, no clear influence of the remaining clinically important parameters (neck length and offsets, femoral anteversion and CCD angle, standard canal flares, patient BMI and weight or stem size) to implant average micromotion was found. In conclusion, the present study demonstrates that the primary stability and tolerance of the short stem to variability in patient anatomy were high, suggesting no need for patient stratification. The developed methodology, based on detailed morphological analysis, accurate implant selection and positioning, prediction of implant micromotion and primary stability, is a novel and valuable tool to support implant design and planning of femoral reconstructive surgery.