Patient Care

Risk-Assessment Models Are Unreliable Predictors of Venous Thromboembolism


 

Clinical Question: Do risk-assessment models (RAMs) accurately predict which hospitalized medical patients are at risk for venous thromboembolism (VTE)?

Background: Predicting which patients are at high risk for VTE is important. Several models exist, but limited data support their generalizability and accuracy in medical inpatients.

Study Design: Retrospective cohort.

Setting: Hospitals participating in the Michigan Hospital Medicine Safety Consortium (MHMSC).

Synopsis: Data collected through MHMSC for selected medical patients were used in the Kucher, Padua, predictive IMPROVE, and Intermountain DVT risk-assessment models. Patients were classified as “low risk” or “at risk” based on each RAM. Follow-up data came from chart extraction (100% of patients) and 90-day post-discharge telephone calls (58% of patients). The primary outcome was image-confirmed hospital associated VTE, including proximal upper- or proximal lower-extremity DVT or pulmonary embolism. These RAMs classified less than 20% of patients as “at risk.” The incidence of VTE was less than 1%. In this external validation study, the Kucher RAM was the least discriminate and the Intermountain was the best, but none yielded results equivalent to the original studies.

This study was limited by the retrospective design, subjectivity of some risk factors (such as immobility), and inability to obtain 90-day telephone follow-up in all patients. Lastly, the binary approach (“at risk” versus “low risk”) may not align with the original derivation studies in which each factor was evaluated independently.

Bottom Line: The incidence of VTE is low in medical inpatients, and current RAMs may not accurately identify at-risk patients.

Citation: Greene MT, Spyropoulos AC, Chopra V, et al. Validation of risk assessment models of venous thromboembolism in hospitalized medical patients. Am J Med. 2016;129(9):1001.e9-1001.e18. doi:10.1016/j.amjmed.2016.03.031.

Next Article:

   Comments ()