OBJECTIVE: To determine the probability that patients with psoriatic arthritis (PsA) will require musculoskeletal surgery. To identify factors predictive of surgery in patients with PsA. To determine the clinical outcome of patients with PsA who underwent surgery compared to patients who did not. METHODS: The database of the Psoriatic Arthritis Clinic was searched to identify individuals who had undergone musculoskeletal surgery. Biological and clinical data such as erythrocyte sedimentation rate (ESR), rheumatoid factor, clinical pattern, nail changes, functional class, number of inflamed joints, and radiological damage, as well as health scores such as Arthritis Impact Measurement Scale 2 (AIMS-2), SF-36, and Health Assessment Questionnaire (HAQ) were available for these patients. Patients who had surgery were compared to those who did not. RESULTS: Out of 444 patients with confirmed PsA, 31 had musculoskeletal surgery (6.98%). This probability increased with the duration of PsA. Surgery patients had their first operation at an average of 13.9 years (range 1-46) after onset of joint disease. Age at onset of PsA, clinical pattern, constancy of clinical pattern over time, rheumatoid factor, functional class, symmetry, nail changes, and the time separating the onset of skin disease and the onset of joint disease were not significantly different at the 0.05 confidence level. Patients who had surgery had significantly more radiological damage (p < 0.001) and more actively inflamed joints (p < 0.02) at first assessment than patients who did not. AIMS2, SF-36, and HAQ scores at final review were not statistically different across both groups. CONCLUSION: In our patients the probability of having musculoskeletal surgery for PsA was 7%. It increased with disease duration. Patients had their first surgery after an average of 13 years. The number of actively inflamed joints and the extent of radiological damage at first assessment were highly predictive of subsequent surgery: patients with the highest numbers of severely affected joints, both clinically and radiologically, were more likely to have surgery. "Baseline characteristics" such as ESR, rheumatoid factor, functional class, clinical pattern, nail changes, and symmetry were not predictive of subsequent orthopedic surgery.