Epidemiology of thrombotic risk factors: the difficulty in using clinical trials to develop a risk assessment model.

There is a paucity of data assessing the risks and prevention of venous thromboembolism in critical care settings. Consequently, it is difficult to estimate the risk of venous thromboembolism for a typical patient, on an individual basis, and this difficulty leads to underuse of thromboprophylaxis, especially in intensive care unit patients. Thrombosis is a multifactorial disease and patients may present multiple risk factors simultaneously. The problem in quantifying risk factors is to combine these risk factors even when they are not detected in the same multivariate analysis. A model for predicting the risk of venous thromboembolism in hospitalized medical patients has yet to be developed and validated. Meanwhile, other approaches have been proposed to replace the ideal study utilizing a large prospective cohort of hospitalized medical patients. In this context, several scoring systems based on risk assessment models have been proposed, some including the use of computerized electronic prompts, to help physicians prescribe appropriate prophylaxis. This article reviews evidence on the risk of venous thromboembolism associated with different medical conditions and risk factors, and presents a tentative risk-assessment model for risk stratification in hospitalized medical patients.