Risk factor analysis and establishment of a nomogram model to predict blood loss during total knee arthroplasty.

PURPOSE: The risk factors for excessive blood loss and transfusion during total knee arthroplasty (TKA) remain unclear. The present study aimed to determine the risk factors for excessive blood loss and establish a predictive model for postoperative blood transfusion. METHODS: This retrospective study included 329 patients received TKA, who were randomly assigned to a training set (n = 229) or a test set (n = 100). Univariate and multivariate linear regression analyses were used to determine risk factors for excessive blood loss. Univariate and multivariate logistic regression analyses were used to determine risk factors for blood transfusion. R software was used to establish the prediction model. The accuracy and stability of the models were evaluated using calibration curves, consistency indices, and receiver operating characteristic (ROC) curve analysis. RESULTS: Risk factors for excessive blood loss included timing of using a tourniquet, the use of drainage, preoperative ESR, fibrinogen, HCT, ALB, and free fatty acid levels. Predictors in the nomogram included timing of using a tourniquet, the use of drainage, the use of TXA, preoperative ESR, HCT, and albumin levels. The area under the ROC curve was 0.855 (95% CI, 0.800 to 0.910) for the training set and 0.824 (95% CI, 0.740 to 0.909) for the test set. The consistency index values for the training and test sets were 0.855 and 0.824, respectively. CONCLUSIONS: Risk factors for excessive blood loss during and after TKA were determined, and a satisfactory and reliable nomogram model was designed to predict the risk for postoperative blood transfusion.

A generalized fuzzy clustering framework for incomplete data by integrating feature weighted and kernel learning.

Missing data presents a challenge to clustering algorithms, as traditional methods tend to pad incomplete data first before clustering. To combine the two processes of padding and clustering and improve the clustering accuracy, a generalized fuzzy clustering framework is proposed based on optimal completion strategy (OCS) and nearest prototype strategy (NPS) with four improved algorithms developed. Feature weights are introduced to reduce outliers' influence on the cluster centers, and kernel functions are used to solve the linear indistinguishability problem. The proposed algorithms are evaluated regarding correct clustering rate, iteration number, and external evaluation indexes with nine datasets from the UCI (University of California, Irvine) Machine Learning Repository. The results of the experiment indicate that the clustering accuracy of the feature weighted kernel fuzzy C-means algorithm with NPS (NPS-WKFCM) and feature weighted kernel fuzzy C-means algorithm with OCS (OCS-WKFCM) under varying missing rates is superior to that of seven conventional algorithms. Experiments demonstrate that the enhanced algorithm proposed for clustering incomplete data is superior.