This article is only available in the PDF format. Download the PDF to view the article, as well as its associated figures and tables.
—I agree with Dr Burt's discussion of the limitations of randomization in small trials. However, the problem he describes is captured in the wider confidence limits that come with small trials—they are wide precisely because of the possibility of unequally distributed covariates. Furthermore, while stratification and matching can help protect against unequal distribution of covariates that are known, randomization is the only way to achieve equal distribution of all the covariates that are not known.I agree with the first point raised by Dr Goldschmidt and Ms Monaco—that new treatments should be paid for when they are offered as a part of well-designed trials. Whether the payment should be provided by private third-party payers as opposed to the government is a complex issue that I will sidestep for now, except to say that for reasons of logistics, marketplace function, and fairness, I prefer that the government do the paying. On
Eddy DM. Eradicating Legionella From Hospital Water-Reply. JAMA. 1997;278(17):1404. doi:10.1001/jama.1997.03550170034023