Periacetabular osteotomy (PAO) is a common treatment for prearthritic hip dysplasia. The goal of this investigation was to determine if computationally assessed hip contact mechanics are associated with joint failure at minimum 10-year follow-up. One hundred patients with hip dysplasia (125 hips) completed patient-reported outcomes an average of 13.8 years (range 10.0-18.0 years) after PAO. 63/125 hips were classified as having failed: 26 converted to total hip arthroplasty (THA) and 37 with significant disability indicated by modified Harris Hip Score (mHHS) ≤ 70. Differences in discrete element analysis-computed contact mechanics were compared between (1) preserved and failed hips, (2) preserved hips and hips that failed by THA, and (3) preserved hips and hips that failed by mHHS ≤ 70. Failed hips had significantly higher preoperative contact stress and exposure metrics (p < 0.001-0.009) than preserved hips. Failed hips also had significantly higher postoperative peak contact stress (p = 0.018), higher mean contact stress (p < 0.001), and smaller contact area (p = 0.044). When assessed based on type of failure, hips that failed by THA had significantly higher postoperative contact stress and exposure metrics than preserved hips (p < 0.001-0.020). In hips that failed by mHHS ≤ 70, mean postoperative contact stress exposure was significantly higher compared to preserved hips (p = 0.043). Despite improved radiographic measures of dysplasia after PAO, pathologic joint contact mechanics can persist and predict treatment failure at minimum 10 years after surgery. Operative and nonoperative techniques specifically intended to reduce harmful contact mechanics in dysplastic hips may have the potential to further improve clinical outcomes after PAO.