Risk of death derived from Hazard Ratio should not be communicated as Relative Risk reductions for death in cancer clinical trials. Intentional or inadvertent?

Abstract

In the last four decades, median survival in advanced cervical cancer has increased from 7.1 months with single-agent cisplatin to 24 months with doublet chemotherapy plus bevacizumab and pembrolizumab. Coinciding with the emergence of targeted therapy, we observe higher pricing of novel cancer drugs and results presentation aimed to positively impact the audience. The first could result in poor drug affordability and financial toxicity, while the second can give patients a magnified sense of progress. In this review, we briefly comment on study designs that may favor obtaining positive results but, most notably, how results are presented in the latest randomized clinical trials in advanced cervical cancer. We found that survival results are expressed in Hazard Ratio (HR) reductions but communicated as a Relative Risk (RR) reduction for death. The HR of the control to the experimental regimen is given by exp[β]. An HR of 0.7 means an HR reduction of 30%. Risk reduction derived from HR may lead to a belief that the intervention can eliminate the chance of the event occurring, but a decreased HR means a reduction in the speed of the event happening, not the chances of it occurring. On the contrary, the “risk reduction” based on RR means that patients have less chance of having the event because RR is a binary measure (alive or dead). Not only Relative Risk, but Absolute Risks and Number-To-Treat (NNT) as well are omited. The issue of how the results of cancer clinical trials are presented deserves open discussion. After all, it is the patient’s right to make an informed decision before embarking on any cancer treatment.