Comparing the time required to complete risk assessment models for venous thromboembolism

BACKGROUND: Venous thromboembolism (VTE) is one of the most preventable causes of hospital acquired mortality. Recent guidelines by the American College of Chest Physicians suggest using risk assessment models to help determine which non-surgical hospitalized patients should receive pharmacologic VTE prophylaxis. However, some risk assessment models tend to be cumbersome and time consuming and may not be practical to implement, while others may be more pragmatic but are not as widely recognized. This study sought to assess and compare the time needed by admitting physicians to complete two different VTE risk models, individually and in combination with a bleeding risk model in non-surgical patients. METHODS: Adult patients newly admitted for non-surgical conditions were assessed by one of four medical interns who were randomly assigned to complete either a VTE risk assessment tool (Padua or Maynard), a bleeding risk tool (IMPROVE), or a combination of a VTE risk tool and bleeding risk tool. The clinical data necessary to determine the risks were obtained in a manner that closely approximated what happens during routine admissions, which included brief reviews of the electronic medical record and direct patient interviews. Data elements required for risk assessment were clearly defined at the beginning of the study by adopting descriptions from the derivation and validation studies, and by corresponding with the primary authors when necessary. The amount of time to complete the risk assessment tools individually and in combination were compared using the two sample t-test. RESULTS: In the assessment of VTE risk alone, 25 patients were evaluated using the Maynard tool with a mean time of 56.6 s (95 % CI: 46.2-67.1), and in another 25 patients the Padua score was completed in an average time of 237.3 s (95 % CI: 200.2-274.5), p<0.001. When the risks for both VTE and bleeding were both assessed in 25 patients using the combination of Maynard and IMPROVE tools, it took a mean time of 211.5 s (95 %: 178.0-245 .0); but when both Padua and IMPROVE scores were calculated in another 25 patients, it took an average of 336.5 s (95 % CI: 293.9-379.2) to complete, p<0.001. CONCLUSIONS: The Padua VTE risk assessment model requires significantly more time to complete compared to the simplified Maynard tool, and thus may not be practical to implement unless the collection of some of its elements can be automated through an electronic medical record system. The inclusion of bleeding risk assessment incurs additional time to the admission process..