Email updates

Keep up to date with the latest news and content from BMC Cancer and BioMed Central.

Open Access Highly Accessed Correspondence

The randomised clinical trial and the hazard ratio – medical research’s Emperor’s New Clothes?

Richard Stephens1* and David Stewart2

Author Affiliations

1 Retired, previously MRC Clinical Trials Unit, London, UK

2 Division of Medical Oncology, The University of Ottawa, Ottawa, Canada

For all author emails, please log on.

BMC Cancer 2014, 14:260  doi:10.1186/1471-2407-14-260

Published: 14 April 2014


As the enthusiasm for individualized treatment and targeted therapies continues to gain momentum, it seems timely to re-assess whether our current research tools are fit for purpose. Randomized Clinical Trials compare groups of patients, the Hazard Ratio is a ‘group summary statistic’, and modeling shows that the same Hazard Ratio score could result from a number of scenarios. Thus the current tools do not provide definitive information as to how to treat an individual patient. We therefore need to concentrate on the use of predictive factor analyses to identify the characteristics of subgroups of patients who respond to specific treatments.

Randomised clinical trials; Hazard ratio; Individualized treatment; Targeted therapy; Predictive analyses