December 1983

Randomized v Historical Control Trials-Reply

Arch Intern Med. 1983;143(12):2342. doi:10.1001/archinte.1983.00350120144042

This article is only available in the PDF format. Download the PDF to view the article, as well as its associated figures and tables.


In Reply.  —Our study was intended not to resolve the HCT v RCT controversy but to provide a basis for assessing the credibility of published reports. Dr Manu proposes the discrimination index to express our finding that, as currently practiced, both RCTs and HCTs have serious flaws. His analysis leads to the pessimistic conclusion that little, if any, useful information can be gleaned from either type of study. We believe that the use of a single index obliterates useful distinctions that can aid in interpretation. The sensitivity and specificity calculations suggest that a therapy found effective by a well-designed RCT is likely to be truly effective, while a positive result from an HCT should be viewed with skepticism. Conversely, a negative finding in an HCT is more likely to be correct than a negative finding from an RCT. The RCTs could be improved by attention to sample size requirements, but

First Page Preview View Large
First page PDF preview
First page PDF preview