Construction and validation of a nomogram for venous thromboembolism risk in elderly inpatients: a multicenter retrospective study

Objective To identify independent risk factors for venous thromboembolism (VTE) in elderly inpatients, develop a clinically applicable nomogram, and evaluate its predictive performance. Methods We conducted a multicenter, retrospective 1:1 frequency-matched case–control study including elderly inpatients (aged ≥65 years) from one municipal and two county-level hospitals between January 1, 2022 and December 31, 2023. After exclusion, 248 VTE cases and 248 non-VTE controls were enrolled. Independent predictors were identified using multivariable logistic regression (with a priori variable selection complemented by univariate screening), and a nomogram was built in R. Temporal validation was performed in an independent cohort of 206 elderly inpatients (103 VTE cases, 103 matched controls) admitted in 2024. Results History of blood transfusion, Caprini score, C-reactive protein (CRP, analyzed as a continuous variable), coronary heart disease, and cerebrovascular disease were independently associated with in-hospital VTE. The nomogram showed strong discrimination with area under the receiver operating characteristic curve (AUC) values of 0.903 (95% CI 0.877–0.930) in the derivation cohort and 0.891 (95% CI 0.845–0.936) in the validation cohort, with good calibration confirmed by Hosmer-Lemeshow test (derivation cohort: χ²=8.24, P=0.413; validation cohort: χ²=7.56, P=0.477) and a Brier score of 0.12 in the derivation cohort. The model had a sensitivity of 0.86 and specificity of 0.82 at the optimal cutoff value. Conclusion The proposed nomogram integrates five clinically accessible predictors to enable relative VTE risk stratification among elderly inpatients, which may support early identification and targeted prophylaxis in clinical practice. Further prospective validation is required to confirm its absolute risk estimation and real-world utility. Clinical trial number Not applicable.