Meta‐analyses on passive smoking and lung cancer effects of study selection and misclassification of exposure

Publisher Website
Google Scholar
Abstract
Meta-analyses of studies on lung cancer and passive smoking both by Wald and the authors have shown greater differences in the pooled relative risk. In our paper three scenarios with conservative rates of misclassfication are assumed and their effect on individual and pooled relative risks are examined. All ther esults of 27 meta-analyses are not statistically different from unity with the exception of 4 results, 3 of which are occurring if one selects only 4 case control studies with minor quality. Considering 10 case control studies along 24 out of 1,023 possible meta-analyses are technically significant. These are dominated by 3 studies of lower quality. Our results lead to the conclusion that based on meta-analyses with proper correction for misclassification the null hypothesis has to be further accepted.