Diagnostic accuracy of the Wells score versus duplex ultrasound for deep vein thrombosis in a Ghanaian cohort

Introduction: The Wells score, which is a 9-item clinical rule, is used to assess the diagnostic probability of deep vein thrombosis (DVT) in suspected patients. However, the accuracy of the Wells rule has not been investigated in Ghanaians. Methods: In a cross-sectional design, 138 suspected DVT patients underwent pretest probability of DVT using the Wells rule. Thereafter, a compressive duplex ultrasound study (CDUS) was performed on both legs to confirm the presence of DVT. Using the CDUS as the gold standard, the accuracy and reliability of the Wells rule were assessed as sensitivity, specificity, predictive values and accuracy diagnostic odds. Youden’s index was used to determine the optimal cutoff of Wells score to predict DVT. Results: The prevalence of DVT was 38.8% and it was associated with localized tenderness [OR (95% CI) = 3.4 (1.49–7.76), p = 0.003], calf swelling [3.72 (1.7–8.12), p = 0.001] and previous DVT [4.69 (1.74–12.61), p = 0.001]. The sensitivity, specificity, and accuracy diagnostic odds ratio for the 2-level Wells score were 38.3%, 93.3%, 73.9% and 8.44 respectively. From the receiver operator characteristic analysis, the area under curve for the 2-level Wells score was 0.76 and the optimal Wells score cut-point for diagnosing DVT based on Youden’s index was >1, with sensitivity, specificity, positive and negative predictive values of 69.4%, 72.1%, 58.6% and 80.5% respectively. Conclusion: The prevalence of DVT in suspected patients in our study population was 38.8% and associated with localized tenderness, calf swelling and previous DVT. Reducing the cut-point of the Wells score to 1 increased the sensitivity and accuracy of the Wells rule in our study population.