PO49 | PROPHYLACTIC ANTICOAGULATION DECISIONS IN HIGH-RISK PATIENTS RECEIVING CANCER-DIRECTED THERAPY: ANALYSIS OF THE VERMONT METHOD

J. Barker¹, K. Libby¹, C. Holmes², K. Martin² | ¹University of Vermont Health System; ²Department of Medicine, University of Vermont, Burlington, VT, USA

Introduction. Validated scoring tools, such as the Khorana risk score and Protecht score, identify patients starting cancer-directed therapy with the highest VTE risk. Despite guideline recommendations for prophylactic anticoagulation for patients at high risk of VTE, less than 5% of patients receive VTE education and guideline-directed VTE prophylaxis. The Vermont Model is a clinical program dedicated to addressing VTE prevention in ambulatory cancer patients starting systemic therapy, including nurse-driven risk assessment for all patients and referral to consult a thrombosis specialist for anticoagulation discussion for high-risk patients. While the Vermont Model successfully improved rates of anticoagulation for high-risk patients, we still do not understand why high-risk patients did not start anticoagulation.

Aim. We aimed to determine the proportion of high-risk patients who received prophylactic anticoagulation and the reasons why we did not start anticoagulation.

Methods. In this retrospective observational study, we manually reviewed charts of high-risk patients, defined as having a Khorana or Protecht Score of 3 or higher seen in the Vermont Model between 2018-2023. We categorized reasons that anticoagulation was not recommended into 8 distinct categories. We used descriptive statistics and logistic regression to test associations between no anticoagulation and clinical characteristics.

Results. Of 284 high-risk patients, clinicians started 144 (50.7%) on anticoagulation and did not start 140 (49.1%). Of the 140 patients not started on AC, 98 (72%) consulted a healthcare provider other than a thrombosis specialist. Out of those 98, 31 (22.3%) had no referral to a thrombosis specialist, 23 (16.6%) were already on anticoagulation, and 20 (14.4%) clinically declined and either transitioned to hospice or died prior to consultation. Of the 42 out of 140 who consulted thrombosis specialists and were not started on anticoagulation, 16 (39.0%) had concomitant/interacting medications and 15 (36.6%) were deemed to have a high bleeding risk. Neither tumor type (p=0.09) nor type of treatment (0.56) was associated with starting anticoagulation.

Conclusions. When patients in the Vermont Model consulted a thrombosis specialist, 78.7% started anticoagulation. Ensuring expedited access to discuss primary prophylaxis may increase anticoagulation for high-risk patients starting cancer-directed therapy.