[Skip to Navigation]
Sign In
Comment & Response
December 2017

Overestimating the Benefit of Cancer Drugs—Reply

Author Affiliations
  • 1London School of Economics and Political Science, London, England
  • 2Institute of Global Health Innovation, Imperial College London, London, England
JAMA Oncol. 2017;3(12):1738-1739. doi:10.1001/jamaoncol.2017.1976

In Reply For reference, we consider all initial, anticancer drug authorizations, and report mean treatment effects. Fojo et al1 consider initial and extension drug authorizations for only solid tumors, and instead report medians. We present drug overall survival (OS) benefits individually. Unlike Fojo et al, we then estimate the total impact on patient OS between 2003 and 2013. Where appropriate, OS benefits were combined sequentially to reflect drug development—we intentionally did not divide by 2. For example, within this period, lapatinib’s approval was based on a comparison against capecitabine monotherapy; T-DM1 was later compared against lapatinib + capecitabine. Drug OS gains were 0.3 to 2.4 and 5.8 months, respectively. Eligible patients would have therefore seen an expected OS increase of 6.1 to 8.2 months between 2003 and 2013. Dividing by 2 would redundantly estimate average OS benefits per new drug.

Add or change institution