Aims To develop prediction models using machine-learning (ML) algorithms for 90 -day and oneyear mortality prediction in femoral neck fracture (FNF) patients aged 50 years or older based on the Hip... Show moreAims To develop prediction models using machine-learning (ML) algorithms for 90 -day and oneyear mortality prediction in femoral neck fracture (FNF) patients aged 50 years or older based on the Hip fracture Evaluation with Alternatives of Total Hip arthroplasty versus Hemiarthroplasty (HEALTH) and Fixation using Alternative Implants for the Treatment of Hip fractures (FAITH) trials.Methods This study included 2,388 patients from the HEALTH and FAITH trials, with 90 -day and oneyear mortality proportions of 3.0% (71/2,388) and 6.4% (153/2,388), respectively. The mean age was 75.9 years (SD 10.8) and 65.9% of patients (1,574/2,388) were female. The algorithms included patient and injury characteristics. Six algorithms were developed, internally validated and evaluated across discrimination (c-statistic; discriminative ability between those with risk of mortality and those without), calibration (observed outcome compared to the predicted probability), and the Brier score (composite of discrimination and calibration).Results The developed algorithms distinguished between patients at high and low risk for 90 -day and one -year mortality. The penalized logistic regression algorithm had the best performance metrics for both 90 -day (c-statistic 0.80, calibration slope 0.95, calibration intercept-0.06, and Brier score 0.039) and one -year (c-statistic 0.76, calibration slope 0.86, calibration intercept-0.20, and Brier score 0.074) mortality prediction in the hold -out set.Conclusion Using high-quality data, the ML -based prediction models accurately predicted 90 -day and one -year mortality in patients aged 50 years or older with a FNF. The final models must be externally validated to assess generalizability to other populations, and prospectively evaluated in the process of shared decision-making. Show less
Objectives: In the SPRINT trial, 18% of patients with a tibial shaft fracture (TSF) treated with intramedullary nailing (IMN) had one or more unplanned subsequent surgical procedures. It is... Show moreObjectives: In the SPRINT trial, 18% of patients with a tibial shaft fracture (TSF) treated with intramedullary nailing (IMN) had one or more unplanned subsequent surgical procedures. It is clinically relevant for surgeon and patient to anticipate unplanned secondary procedures, other than operations that can be readily expected such as reconstructive procedures for soft tissue defects. Therefore, the objective of this study was to develop a machine learning (ML) prediction model using the SPRINT data that can give individual patients and their care team an estimate of their particular probability of an unplanned second surgery. Methods: Patients from the SPRINT trial with unilateral TSFs were randomly divided into a training set (80%) and test set (20%). Five ML algorithms were trained in recognizing patterns associated with subsequent surgery in the training set based on a subset of variables identified by random forest algorithms. Performance of each ML algorithm was evaluated and compared based on (1) area under the ROC curve, (2) calibration slope and intercept, and (3) the Brier score. Results: Total data set comprised 1198 patients, of whom 214 patients (18%) underwent subsequent surgery. Seven variables were used to train ML algorithms: (1) Gustilo-Anderson classification, (2) Tscherne classification, (3) fracture location, (4) fracture gap, (5) polytrauma, (6) injury mechanism, and (7) OTA/AO classification. The best-performing ML algorithm had an area under the ROC curve, calibration slope, calibration intercept, and the Brier score of 0.766, 0.954, -0.002, and 0.120 in the training set and 0.773, 0.922, 0, and 0.119 in the test set, respectively. Conclusions: An ML algorithm was developed to predict the probability of subsequent surgery after IMN for TSFs. This ML algorithm may assist surgeons to inform patients about the probability of subsequent surgery and might help to identify patients who need a different perioperative plan or a more intensive approach. Show less
