I wrote here about Sequenom, a company which claimed to have developed an in vitro test for Downs Syndrome. The whole story dissolved into a heap of lawsuits, allegations of fraud, questions about whether Sequenom ever had such a test at all, and other craziness.
Now the SEC has charged a former executive of the company with lying to investors. Elizabeth Dragon publicly touted the accuracy of the company's test several times, in front of large groups, before the roof came in. Here's the official complaint (if that doesn't work, try here) and it's most interesting reading:
Dragon knew that all but six of the 51 samples had been tested on an unblinded basis, and also knew that, of the six blinded samples, her scientists incorrectly called one of the three T21 samples. In fact, in an e-mail to her lead scientist, Dragon asked how close the call was when she learned that the third T21 sample had not been called correctly on a blinded basis.
On June 3, 2008, Dragon presented data regarding the Down Syndrome Test at an analyst-and-investor briefing during the ISPD conference, which was simultaneously made available to the public via a webcast. . .The slides in Dragon's presentation indicated that the Test performed with 100% accuracy. . .
. . .Additionally, following her presentation, Dragon was asked a question regarding whether there was ambiguity in calling the samples. In response, she stated that T21 samples were very clear, and that "[t]he overall call is strongly positive . . ., you know it's definitely a Downs and you can read it as a Downs without any problem . . . "
She's also charged with stating that the unblinded tests were, in fact, blinded, and apparently never mentioned that unblinded tests were even being conducted. When blinded tests were run later, their accuracy was below that of the tests already on the market, actually. The SEC charges Dragon with pressuring the scientists involved to drop samples and re-evaluate their calls until the accuracy numbers came up. Her presentation of these runs also showed "100% accuracy", with no false positives and no false negatives. None of it was true.
I'm glad to see this prosecution happening. We live by accurate data, and seeing it sawed off and spray-painted like this makes me furious. If the SEC's allegations are true, I hope it can serve as an example to others who might be tempted to try the same sort of thing.