Objective To construct and validate a predictive model for the risk of deep vein thrombosis (DVT) after lower extremity orthopedic surgery.

Methods Clinical records of hospitalized patients who underwent lower extremity orthopedic surgery in Wuxi Traditional Chinese Medicine Hospital from January 2017 to October 2019 were collected. The univariate and multivariate analysis with the backward stepwise method were applied to screen variables and build a nomogram prediction model, and the performance of the nomogram was evaluated with respect to its discriminant capability, calibration ability, and clinical utility.

Results A total of 5 773 hospitalized patients with orthopedic surgery of lower extremity were included in the study, with the incidence of DVT of 0.9%. Through single factor and multi-factor stepwise regression analysis, 5 variables were selected from 31 variables to construct the prediction model, including age, mean corpuscular hemoglobin concentration(MCHC), D-dimer, platelet distribution width(PDW), and thrombin time (TT). The receiver operating characteristic (ROC) curve showed that areas under the ROC curve in the training and validation cohort were 0.859 and 0.857, respectively. The model had good calibration ability and clinical practicability.

Conclusion The DVT risk prediction model constructed in this study has good differentiation ability, calibration ability and clinical practicability, which is helpful for doctors to classify DVT patients after lower extremity orthopedic surgery and formulate early treatment plan.