Background: Total knee (TKR) and hip (THR) replacement (arthroplasty) are effective surgical procedures that relieve pain, improve patients' quality of life and increase functional capacity. Studies on variations in medical practice usually place the indications for performing these procedures to be highly variable, because surgeons appear to follow different criteria when recommending surgery in patients with different severity levels. We therefore proposed a study to evaluate inter-hospital variability in arthroplasty indication. Methods: The pre-surgical condition of 1603 patients included was compared by their personal characteristics, clinical situation and self-perceived health status. Patients were asked to complete two health-related quality of life questionnaires: the generic SF-12 (Short Form) and the specific WOMAC (Western Ontario and Mcmaster Universities) scale. The type of patient undergoing primary arthroplasty was similar in the 15 different hospitals evaluated. The variability in baseline WOMAC score between hospitals in THR and TKR indication was described by range, mean and standard deviation (SD), mean and standard deviation weighted by the number of procedures at each hospital, high/low ratio or extremal quotient (EQ_5-95), variation coefficient (CV_5-95) and weighted variation coefficient (WCV_5-95) for 5-95 percentile range. The variability in subjective and objective signs was evaluated using median, range and WCV_5-95. The appropriateness of the procedures performed was calculated using a specific threshold proposed by Quintana et al for assessing pain and functional capacity. Results: The variability expressed as WCV_5-95 was very low, between 0.05 and 0.11 for all three dimensions on WOMAC scale for both types of procedure in all participating hospitals. The variability in the physical and mental SF-12 components was very low for both types of procedure (0.08 and 0.07 for hip and 0.03 and 0.07 for knee surgery patients). However, a moderate-high variability was detected in subjective-objective signs. Among all the surgeries performed, approximately a quarter of them could be considered to be inappropriate. Conclusions: A greater inter-hospital variability was observed for objective than for subjective signs for both procedures, suggesting that the differences in clinical criteria followed by surgeons when indicating arthroplasty are the main responsible factors for the variation in surgery rates.

Patient expectations regarding surgery may be related to outcomes in total joint replacement (TJR). The aim of this study was to determine the association of patient expectations with health related quality of life (HRQoL) outcomes measured by Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and Short Form 12 (SF-12) and satisfaction with current symptoms measured on a 4-point Likert scale, one year after surgery, adjusting for Body Mass Index (BMI), age, gender, joint, education, previous intervention and baseline scores.