Validation of risk assessment models of deep vein thrombosis among hospitalized medical and obstetrics patients at two tertiary referral hospitals in rwanda

INTRODUCTION: Deep Vein Thrombosis (DVT) is a significant cause of morbidity and mortality for hospitalized patients and pregnant women. Different risk assessment models (RAMs) were developed to identify at-risk patients but not validated in either patients’ settings. Aim of the study: To evaluate and compare the performance of the Padua, Geneva, and Wells score in terms of risk stratification and pre-test clinical probability of DVT among hospitalized medical and obstetrics patients at two tertiary hospitals in Rwanda. METHODS: We conducted cross-sectional, descriptive monthly surveys. We included all consenting patients that were admitted to medical and obstetrics wards during the study surveys. We collected demographics, risk factors, signs, and symptoms of DVT. The DVT risk was calculated using each one of the above scores, followed by DVT screening of both lower extremities by compression ultrasound scans. RESULTS: Out of 807 participants; 571 (70.7%) were women, 43 (7.5%) were pregnant, and 173 (30.3%) were in postpartum period. DVT was detected in 46/807 (5.7%) patients and was statistically significantly more frequent in high versus low-risk categories when applying each score. By the Wells score, DVT was nearly twenty times greater among high versus low-risk patients (50%, vs. 2.6%, p < 0.0001), and six times by the Geneva score. CONCLUSION: The Wells and Geneva scores were significant predictors of DVT among hospitalized medical and obstetrics patients in Rwanda. More studies are needed to validate these scores in different clinical settings.