Validation of a risk score for cancer-associated thrombosis using USA-wide electronic health record data

Venous thromboembolism (VTR) is associated with increased mortality and morbidity in patients with cancer. Existing risk prediction models are typically validated within individual sites, a fragmented approach that limits clinical adoption. A prognostic study using a retrospective cohort of 732,594 patients with cancer receiving systemic therapy between 2018 and 2023 from 184 US health systems was carried out to validate the electronic health record cancer-associated thrombosis (HER-CAT) score compared with the benchmark Khorana score. HER-CAT significantly outperformed the Khorana score and had 20% improved accuracy. The model had consistent calibration by demographic subgroups, health system sites, and cohorts stratified by bleeding risk. The authors conclude that these results suggest that standardized structured HER data from different health systems can support scalable validation and implementation of CAT risk assessment.