Objective Prioritisation of patients for total joint replacement (TJR) represents a challenge. The objective of this study was to construct a propensity score (PS) that both predicts and informs prioritisation of TJR among elderly men.

Methods Clinical data from 11 388 men were integrated with hospital morbidity data and mortality records. A PS quantifying each individual's probability of having TJR was calculated using multivariable competing risk regression models. The PS was then used to assess risk of incident in-hospital complications and mortality following TJR.

Results Younger and healthier patients were selected for TJR. Co-morbidities such as diabetes mellitus, peripheral vascular disorders, and cancer lowered the probability of having TJR. Among men who had TJR, 25% developed a major in-hospital complication. The PS independently predicted both major complications and short- and long-term mortality. Patients with low PSs who nonetheless underwent TJR were more likely to experience an adverse outcome. After adjusting for risk factors, patients who were in the lowest tertile of the distribution of the PS were 67% more likely to develop a major complication (p=0.023), and 2.2 times more likely to die within 10 years after TJR (p=0.035). Other predictors of major complications following TJR included weight, injury, and having a minor complication. In-hospital complications independently increased the risk of mortality after TJR.

Conclusion In the presence of clinical indications for TJR, this PS informs clinical decision making about selecting patients who are most likely to benefit and least likely to be harmed as a consequence of TJR.