The role of arthroscopy in the management of glenohumeral osteoarthritis: a Markov decision model.

PURPOSE The purposes of this study were (1) to construct a theoretical Markov decision model to compare the total remaining quality-adjusted life-years following either arthroscopic management (AM) or total shoulder arthroplasty (TSA) for the treatment of glenohumeral osteoarthritis and (2) to determine the possible effects of age on the preferred treatment strategy. METHODS A Markov decision model was constructed to compare AM and TSA in patients with glenohumeral osteoarthritis. The rates of surgical complications, revision surgery, and death were derived from the literature and analyzed. The principal outcome measure was the mean total remaining quality-adjusted life-years after each treatment strategy. Sensitivity analyses were performed for age at the initial procedure, utilities, and transition probabilities. RESULTS This theoretical decision model showed that AM was the preferred strategy for patients younger than 47 years, TSA was the preferred strategy for patients older than 66 years, and both treatment strategies were reasonable for patients aged between 47 and 66 years. The model was highly sensitive to age at the index surgery, utilities of wellness states, survivorship, and the probability of failure after either AM or TSA. CONCLUSIONS According to this theoretical decision model, AM was the preferred treatment strategy for patients younger than 47 years, primary TSA was the preferred treatment strategy for patients older than 66 years, and both treatment options were reasonable for patients aged between 47 and 66 years. LEVEL OF EVIDENCE Level II, economic and decision analysis.