Abstract
OBJECTIVES: To assess and compare the Caprini and Padua risk assessment models (RAMs) for predicting venous thromboembolism (VTE) in hospitalized patients. METHODS: We retrospectively reviewed 28 VTE and 450 non-VTE patients hospitalized at King Abdulaziz University Hospital, Jeddah, Saudi Arabia, in 2019. Their baseline medical, demographic, and radiological reports were recorded. We compared Caprini scores (defined at admission) and Padua scores (calculated retrospectively) for their ability to predict VTE. A cumulative risk score was created by adding the individual scores for each risk factor. We also analyzed the sensitivity, specificity, and diagnostic accuracy of the RAM scores. RESULTS: Major differences in risk factors were shown between patients with and without VTE. Previous VTE was significantly associated with a higher risk of VTE (28.6%), as was reduced mobility (57.1%), acute infection (25%), high Caprini score (50%), and high Padua score (64.3%, p<0.05). The sensitivity of the Caprini score (96%) was higher than that of the Padua score (64.3%), as was the specificity (92.1% vs. 46.9%), positive predictive value (93% vs. 7%), and accuracy (94.1% vs. 47.9%). The specificity of the Caprini score was higher than that of the Padua score in Critical Care, Gynecology/Obstetrics, and Surgical departments. The Caprini RAM showed the lowest level of specificity in the medical department. CONCLUSION: The Caprini RAM demonstrated higher sensitivity, specificity, and predictive accuracy than did the Padua RAM and thus distinguished low and high VTE risk in hospitalized patients.