300 Section - Development and validation of a novel screening tool for deep vein thrombosis in patients with spinal cord injury: A Five-Year Cross-Sectional Study

Deep vein thrombosis (DVT) presents a significant risk of complication in patients with spinal cord injury (SCI), necessitating accurate screening methods. While the Caprini Risk Assessment Model (Caprini RAM) has seen extensive use for DVT screening, its efficacy remains under scrutiny.

Design:

We created and evaluated three nomograms for their effectiveness in DVT screening. Model 1 incorporated variables such as age, D-dimer level, red blood cell (RBC) count, platelet count, presence of type 2 diabetes mellitus, high blood pressure, mode and level of injury, degree of impairments, and Caprini scores. Model 2 was derived from Caprini scores alone, and Model 3 focused on independent risk factors. We assessed these models using the area under the curve (AUC) of the receiver operating characteristic (ROC), calibration curves, and decision curve analysis (DCA), employing bootstrap resampling tests (500 iterations) to determine their accuracy, discriminative ability, and clinical utility. Internal validation was performed on a separate cohort.

Results:

 Nomogram was established with well-fitted calibration curves for model 1, 2 and 3(AUC=0.808, 0.751 and 0.797; 95%CI=0.76-0.86, 0.70-0.80 and 0.75-0.84; respectively), indicating model 1 outperformed the others in prediction DVT risk, followed by model 3 and 2. These findings were consistent in the validation cohort, with DCA further corroborating our conclusions.

Conclusion:

A nomogram integrating clinical data with Caprini RAM provides a superior option for DVT screening in SCI patients within rehabilitation settings, outperforming Caprini RAM.