BACKGROUND: Shoulder hemiarthroplasty (HA) has been the standard treatment for complex proximal humerus fractures in the elderly requiring surgery but not amenable to fixation. Reverse total shoulder arthroplasty (RTSA) has also emerged as a costly albeit highly effective alternative. The purpose of this study was to compare the cost-effectiveness of nonoperative fracture care, HA, and RTSA for complex proximal humerus fractures from the perspective of both U.S. payors and hospitals. METHODS: A Markov model was constructed for the treatment alternatives. Costs were expressed in 2013 U.S. dollars and effectiveness in quality-adjusted life-years (QALYs). The principal outcome measure was incremental cost-effectiveness ratio (ICER). Sensitivity analyses were performed to evaluate model assumptions. RESULTS: In the base case, from the payor perspective, RTSA was associated with an ICER of $8100/QALY; HA was eliminated from payor analysis as a cost-ineffective strategy. From the hospital perspective, however, HA was not cost-ineffective and the ICER for HA was $36,700/QALY, with RTSA providing incremental effectiveness at $57,400/QALY. RTSA was the optimal strategy in 61% and 54% of payor and hospital probabilistic sensitivity analyses, respectively. The preferred strategy was dependent on associated QALY gains, primary RTSA cost, and failure rates for RTSA. CONCLUSIONS: RTSA can be a cost-effective intervention in the surgical treatment of complex proximal humerus fractures. HA can also be a cost-effective intervention, depending on the cost perspective (cost-ineffective for payor but cost-effective for the hospital). This analysis highlights the opportunities for increased cost-sharing strategies to alleviate the cost burden on hospitals.