ABSTRACT
Number needed to treat (NNT) is a useful measure to meaningfully translate the therapeutic effect of a drug to clinical practice. Time-to-event endpoints are commonly used in clinical trials. NNT has been extended to time-to-event endpoints, but these extensions have some limitations. In addition, NNT and number needed to harm (NNH) can be applied to provide useful information in support of benefit–risk assessment of a drug. In this article, we will first propose a hazard-based NNT using an additive hazards model. Estimation and inference procedures will be provided. Graphical methods and formal statistical tests will be proposed to evaluate the assumption of additivity. We will also discuss using NNT and NNH to facilitate benefit–risk decision making. Inference on the NNH and NNT ratio will be explored as well. The proposed methods will be illustrated with a real clinical trial dataset.
Acknowledgments
We thank the editor, the associate editor, and the two anonymous reviewers for their careful review and constructive comments which greatly improved the content and presentation of this article.