Leszczynski: Epi Studies Not Worth Much
GIGO. Garbage in, garbage out. That's what Dariusz Leszczynski thinks of Hardell's new reanalysis (see December 17 below). The tumor risk seen by the Hardell group may now be similar to the one in Interphone, but that doesn't mean much, says Leszczynski, because Interphone is really a bunch of garbage too. "Let us agree that both data sets are biased," he wrote in his blog over the weekend, and whatever you might do to them, the results will never be "scientifically reliable." Shouldn't we work with what we have?, we asked him. Sure, Leszczynski replied, but we shouldn't kid ourselves that we'll ever end up with anything better than what we already have. That is, more garbage. We don't agree. Hardell and Interphone are two independent data sets that point to two different types of tumor risks (glioma and acoustic neuroma) among long-term cell phone users. Plus we have the Interphone Israeli study that implicates a third type of tumor (of the parotid gland). A major branch of statistics (Bayesian) teaches us how to make the best of what we have in hand, imperfect though it may be. Since the Hardell and Interphone data represent the bulk of what we know about long-term cell phone risks, we disregard those disquieting findings at our peril. We would turn the tables on Leszczynski: To conclude that we have no evidence of a cell phone tumor problem would be garbage. We bring all this up because Leszczynski (a molecular biologist, not an epidemiologist) has been selected by IARC to be a member of its RF cancer assessment panel that will meet in May.