[Skip to Content]
[Skip to Content Landing]
Citations 0
December 20, 2000

Non-English Reports of Medical Research

Author Affiliations

Stephen J.LurieMD, PhD, Senior EditorIndividualAuthorPhil B.FontanarosaMD, Executive Deputy EditorIndividualAuthor

JAMA. 2000;284(23):2996-2997. doi:10.1001/jama.284.23.2993

To the Editor: In their systematic review and meta-analysis, Drs Wu and Colford1 examined the role of chorioamnionitis as a risk factor for cerebral palsy and cystic periventricular leukomalacia (cPVL). They concluded that chorioamnionitis is a risk factor for both conditions, a finding with potentially important clinical and economic implications. The authors carefully examined potential sources of heterogeneity, which is of crucial importance in meta-analysis of observational studies.2 It is ironic, however, that they used sophisticated statistical methods to examine the possible presence of publication bias while completely ignoring an obvious source of reporting bias: a priori exclusion of studies published in languages other than English. An unknown number of relevant studies are thus missing from their review. This likely reduced the precision of their estimates and, more worrisome, may have introduced bias. For example, it has been shown that investigators based in German-speaking countries tend to publish clinical trials in English-language journals if the results are statistically significant but choose German-language journals if results are negative.3

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