Scientific Evidence Portal
Meta-Analyses of The Epidemiological Evidence Relating ETS to Lung Cancer and Heart Disease | Peter Lee, Barbara Forey, Jan Hamling
Article Published: 2006
Published By: PN Lee Statistics and Computing Ltd
"Epidemiological evidence relating to two possible effects of ETS in never smokers is presented elsewhere, based on over 60 studies of lung cancer and over 30 studies of heart disease. In attempting to assess this overall evidence, the technique of meta-analysis is often used to combine the relative risk estimates from the individual studies. Results from a number of meta-analyses are presented in this document. However, before presenting these results, attention should be drawn to some features and limitations of the technique. "
"Meta-analysis was originally designed to combine evidence from randomised studies of similar design, to try to obtain a more accurate estimate of the possible effect of treatment. However, in recent years it has been increasingly used to combine evidence from epidemiological studies of quite widely varying design. It is important to realise that such a procedure may result in a combined relative risk estimate that has narrow confidence limits and therefore appears precise, but is in fact a quite inaccurate estimate of the true effect of treatment (if any)."