Title : Comparison of the risk assessment models in predicting the risk of venous thrombosis in hospitalized patients: A meta-analysis
Abstract:
Objective: To systematically compare the predictive accuracy of the Caprini and the Padua RAMs in VTE in hospitalized patients. Providing a basis for nurses to select a more appropriate assessment tool when performing inpatient venous thrombosis risk assessment.
Methods: Relevant studies were searched in PubMed, Embase, the Cochrane Library, and Web of Science from inception to March 2023 using a combination of Medical Subject Headings and entry terms. Study selection and data extraction were performed independently by two or more researchers. Sensitivity and specificity of the Caprini and the Padua RAMs were pooled using Stata. The quality of the included studies was assessed using the Quality Assessment of Studies for Diagnostic Accuracy (QUADA-2).
Results: The meta-analysis included 15 studies involving 25,947 patients, of which 22,035 were medical patients and 1,872 were surgical patients. For the Caprini RAM, the sensitivity of the Caprini RAM was 0.88 [95%CI (0.72, 0.95)], the specificity was 0.36 [95%CI (0.18, 0.59)], and the diagnostic odds ratio (DOR) was 4.00 [95% CI (3.00, 6.00)] in hospitalized patients; The sensitivity of the Padua RAM was 0.68 [95%CI (0.50, 0.81)], the specificity was 0.53 [95% CI (0.29, 0.75)], and the DOR was 2.00 [95% CI (1.00, 4.00)]; The direct comparison of these two RAMs indicated that the Caprini RAM was more accurate than the Padua RAM for predicting VTE (DOR, Caprini 2.80 [95% CI (2.67, 0.93)] vs. Padua 2.01 [95% CI (1.93, 2.09)]). Furthermore, the Caprini RAM was again better than the Padua RAM in both medical and surgical patients in the subgroup analysis (medical DOR, Caprini 7.00 [95 % CI (2.00,25.00)] vs. Padua 4.00 [95 % CI (2.00,9.00)]; surgical DOR, Caprini 4.00 [95 % CI (3.00,6.00)] vs. Padua 2.00 [95 % CI (1.00,5.00)]).
Conclusion: The Caprini RAM has superior predictive sensitivity and diagnostic odds ratio than the Padua RAM in both medical and surgical patients. However, the Padua RAM was superior to Caprini RAM in specificity. Nonetheless, both RAMs had low positive likelihood ratio and negative likelihood ratio as well as unsatisfactory diagnostic performance. Based on current findings, the Caprini RAM has considerable diagnostic value in VTE risk in hospitalized patients, but further studies are warranted to explore the influencing factors for different disease characteristics and to optimize and improve the diagnostic efficacy of the risk assessment model.
