Shoulder arthroplasty outcomes have been reported in many case series. Typically, these series have followed either a single prosthesis used to treat a variety of arthritic disorders of the shoulder or experience in a single institution. In contrast, this report of a prospective study summarizes the experience of several surgeons with a single prosthetic design for treatment of primary osteoarthritis of the shoulder. A prospective, multicenter clinical outcome study evaluated 176 shoulders in 160 patients with primary osteoarthritis. This study evaluated a single prosthetic design (Global Shoulder) used by 19 contributing surgeons. Enrollment included 133 total shoulder replacements and 43 humeral head replacements (hemiarthroplasty) in 98 men and 62 women. Neither age nor sex affected whether hemiarthroplasty or total shoulder arthroplasty was performed. Patients with full-thickness cuff tears preferentially had hemiarthroplasty. The decision to perform total shoulder arthroplasty or hemiarthroplasty was based on the surgeon's preference. There were significant improvements (P <.001) in all evaluated and self-assessed outcome parameters from the preoperative baseline for both total shoulder arthroplasty and hemiarthroplasty. The results confirm that prosthetic arthroplasty leads to dramatic improvement in pain, function, and patient satisfaction. Intraoperative complications occurred in 5.4% of cases, and postoperative complications occurred in 7.8%. The most common intraoperative complications were intraoperative fractures, occurring in 9 cases. The most common postoperative complications were glenoid component loosening and humeral head subluxation. Almost all cases of humeral head instability were associated with rotator cuff tears or glenoid component loosening (or both). Seven shoulders underwent 9 additional surgeries during the 5-year study period. Thirteen shoulders in 11 patients were lost as a result of death unrelated to the procedure; 2 shoulders in 1 patient were lost within 3 days/3 months after the bilateral replacements as a result of death from pulmonary embolism. Nine percent of the shoulders (16/176) had full-thickness rotator cuff tears. Eight of the 16 shoulders with full-thickness supraspinatus cuff tears had hemiarthroplasty. All of these tears were isolated to the supraspinatus tendon, and all were repairable. There were no differences in postoperative pain, function, American Shoulder and Elbow Surgeons scores, or range of motion. There were no differences between total shoulder arthroplasty and hemiarthroplasty in those patients with a reparable rotator cuff tear. Total shoulder arthroplasty and hemiarthroplasty for treatment of primary osteoarthritis result in good or excellent pain relief, improvement in function, and patient satisfaction in 95% of cases. Avoiding intraoperative humeral shaft fractures through use of an uncemented, canal-filling prosthetic stem requires careful attention to reaming and component sizing. Postoperative humeral head subluxation is often associated with other factors including rotator cuff tears or glenoid component loosening.