[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Citations 0
Books, Journals, New Media
June NaN, 1999


Author Affiliations

Margaret A.WinkerMD, Deputy EditorIndividualAuthorPhil B.FontanarosaMD, Interim CoeditorIndividualAuthor

JAMA. 1999;281(24):2287-2288. doi:10-1001/pubs.JAMA-ISSN-0098-7484-281-24-jbk0623

In Reply: The criticism that we did not investigate the potential effect of corporate-sponsored studies on the meta-analysis and that most of the authors had received financial support that could have led to bias is misinformed. It is not surprising that trials "without corporate sponsorship found no clinical effect" as they all had small sample sizes. Contrary to the assertions of Mr Koepp and Dr Miles, the most favorable trial was excluded because of deficiencies in design and analysis, and this study did not have corporate funding.1 Nevertheless, a glance at the point estimates of the treatment effects of the various studies reveals no apparent difference between corporate and noncorporate studies. As the test of heterogeneity was not significant, we had no reason to evaluate sponsorship as a stratification variable. Unplanned subgroup analysis often leads to spurious results,2 especially when there are so few studies. Most (large) drug trials are sponsored by industry and those used for registration are scrutinized by regulatory authorities, often after involvement in their design.3 The methods of the meta-analysis were unbiased. Also, in this favorable meta-analysis—as we discussed clearly in our article—the effect sizes generally were smaller than those reported in the individual publications.

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