Venous thromboembolism (VTE) risk stratification in general medical patients at an academic medical center

Hospital-acquired venous thromboembolism (VTE) is still a concern for general medical patients. Pharmacologic prophylaxis can reduce VTE incidence, but there is the potential for adverse effects. Therefore, determining which patients should receive VTE prophylaxis via risk scoring tools is essential. Limited evidence exists for the ideal venous thromboembolism risk assessment model (RAM) in hospitalized medical patients, as compared to other hospitalized patient subgroups such as surgical patients. The primary objective was to investigate the utilization and appropriateness of our institution-based VTE RAM and comparison to the Padua Prediction Score (PPS). This would allow for a gauge of provider risk assessment accuracy as well as appropriate predictive potential of the PPS or whether an alternative to the PPS should be considered. A total of 330 adult general medicine patients were included in this retrospective chart review. When compared to our institution-based VTE RAM, providers predominately stratified patients at a higher VTE risk than the institution-based VTE RAM. VTE incidence was 0.3%, which was lower than predicted. Significant discordance exists between providers' VTE risk assessment and that predicted by RAMs. Our institution-based VTE RAM appears comparable to PPS; however, it was not being utilized by providers, resulting in potentially unnecessary use of pharmacologic prophylaxis. The most appropriate venous thromboembolism risk assessment model for general medicine patients is undetermined. Our providers generally assess patients as moderate or high VTE risk, despite our institution-based RAM which typically recommends a lower risk category than provider selection. Because of provider risk assessment, more patients received pharmacologic VTE prophylaxis than would have been recommended by the RAM, which might correlate to the low incidence of VTE which was < 0.5%, although bleeding complications were not assessed in this study. A prospective study utilizing the Padua Prediction Score (or similar RAM) in general medicine patients is warranted in order to decipher the best method of predicting VTE risk.