Abstract
PURPOSE: Pre-operative scores based on patient characteristics are commonly used to predict hip fracture outcomes. Mobility, an indicator of pre-operative function, has been neglected as a potential predictor. We assessed the ability of pre-fracture mobility to predict post-operative outcomes following hip fracture. METHODS: We analysed prospectively collected data from hip fracture surgery patients at a large-volume trauma unit. Mobility was classified into four groups. Post-operative outcomes studied were mortality and residence at 30 days, medical complications within 30- or 60-days post-operatively, and prolonged length of stay (LOS, ≥ 28 days). We performed multivariate regression analyses adjusting for age and sex to assess the discriminative ability of the Nottingham Hip Fracture Score (NHFS), with and without mobility, for predicting outcomes using the area under the receiver operating characteristic curve (AUROC). RESULTS: 1919 patients were included, mean age 82.6 (SD 8.2); 1357 (70.7%) were women. Multivariate analysis demonstrated patients with worse mobility had a 1.7-5.5-fold higher 30-day mortality (p ≤ 0.001), and 1.9-3.2-fold higher likelihood of prolonged LOS (p ≤ 0.001). Worse mobility was associated with a 2.3-3.8-fold higher likelihood of living in a care home at 30-days post-operatively (p < 0.001) and a 1.3-2.0-fold higher likelihood of complications within 30 days (p ≤ 0.001). Addition of mobility improved NHFS discrimination for discharge location, AUROC NHFS 0.755 [0.733-0.777] to NHFS + mobility 0.808 [0.789-0.828], and LOS, AUROC NHFS 0.584 [0.557-0.611] to NHFS + mobility 0.616 [0.590-0.643]. CONCLUSION: Incorporating mobility assessment into risk scores may improve casemix adjustment, prognostication following hip fracture, and identify high-risk patient groups requiring enhanced post-operative care at admission.