Development and Pre-validated Multiple Variable Prediction Model Using Machine Learning for Early Functional Recovery After Joint Replacement Surgery

What this Trial Is About

Researchers aim to develop and test a machine learning algorithm that predicts early recovery of mobility and hospital length of stay in patients who have undergone elective hip or knee joint replacement surgery. The study includes both retrospective data from 2020 to 2023 and prospective data collected from March 2026 to December 2027. The primary question is whether the model can accurately classify patients with faster versus slower recovery of autonomous mobility shortly after surgery. The study uses clinical and perioperative data from patients who received post-operative physiotherapy. The machine learning model will be trained on retrospective data and then evaluated on prospectively collected data to assess its predictive accuracy. Several algorithms like logistic regression, random forest, and gradient boosting will be compared. The model's performance will be measured by statistical metrics including the area under the receiver operating characteristic curve (AUROC), precision, and calibration plots. Participants' recovery will be assessed by their ability to climb stairs within the first four days after surgery, as recorded in physiotherapy diaries and health records. Length of hospital stay is tracked, with discharge by day five considered regular. Data collection includes demographics, surgical details, pain levels, and other clinical factors. Predictive results are stored separately and do not affect clinical care. The study involves two phases with a total of 943 patients and will monitor model performance over approximately two years.

Development and Pre-validation of a Machine Learning-based Prediction Algorithm for Early Functional Recovery in Patients Undergoing Hip and Knee Replacement Surgery