We have read with interest the manuscript by Barber and Clarke-Pearson because we share the interest on a topic of large impact. We have comments, however, regarding some of their conclusions.

Although the title claims that the available risk assessment tools have limited utility, this is not an accurate generalization. The studied patients in the study by Barber and Clarke-Pearson had a 1 month, 1.8% incidence of symptomatic venous thromboembolism (VTE) despite at least 95% compliance with prophylaxis. Consistent with prior literature on bariatric, orthopedic, oncologic surgery and current guideline opinion, this rate may be considered as high. Thus, it is not surprising that the Caprini risk score (CRS) classified most gynecological cancer patients as individuals in need of pharmacological prophylaxis.

Given that there was a presumed high thromboprophylaxis rate done with no stratification based on either CRS or Rogers score, one cannot conclude that the scores and consequently risk-based prophylaxis are inadequate. There was no appropriate standard for comparison or an organized strategy by risk tier. Moreover, although the authors interpret that the scores do not stratify the thrombosis risk among patients with gynecological malignancy, their findings show that higher Caprini and Rogers scores exhibited a matching higher probability of VTE with good statistical linearity. They actually found, in concordance with other authors, a subgroup of adequately classified patients with high VTE risk despite conventional prophylaxis.

It is concerning that the authors used 2 different stratifications of the CRS (Table 2 vs Table 3) but did not clarify the rationale for using them interchangeably in the conclusion. In addition, many of the CRS variables were not available, which limits the interpretation, including the paradoxical VTE incidence in a misclassified CRS risk group. Indeed, contrary to the findings by Barber and Clarke-Pearson, Stroud et al have validated the CRS in a gynecology-oncology population. Among 1123 patients, the 3 month rate of VTE was 3.3% and the CRS accurately predicted all VTE events.

What the authors have successfully presented is that both the Rogers and Caprini scores have a demonstrable linearity with respect to the occurrence VTE among patients with gynecological malignancies. What needs to be urgently defined is which intensity and duration of thomboprophylaxis shall be offered to patients with very high scores, which despite conventional prevention had a potentially fatal 1 month VTE >2% in the present study. The idea of personalized duration of prophylaxis is not a new concept and has been successfully implemented in other surgical entities.