Performance of Risk Assessment Models for VTE in Patients Who Are Critically Ill Receiving Pharmacologic Thromboprophylaxis: A Post Hoc Analysis of the Pneumatic Compression for Preventing VTE Trial

Abstract

The diagnostic performance of the available risk assessment models for VTE in patients who are critically ill receiving pharmacologic thromboprophylaxis is unclear. For patients who are critically ill receiving pharmacologic thromboprophylaxis, do risk assessment models predict who would develop VTE or who could benefit from adjunctive pneumatic compression for thromboprophylaxis? In this post hoc analysis of the Pneumatic Compression for Preventing VTE (PREVENT) trial, different risk assessment models for VTE (ICU-VTE, Kucher, Intermountain, Caprini, Padua, and International Medical Prevention Registry on VTE [IMPROVE] models) were evaluated. Receiver-operating characteristic curves were constructed, and the sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios were calculated. In addition, subgroup analyses were performed evaluating the effect of adjunctive pneumatic compression vs none on the study primary outcome. Among 2,003 patients receiving phar